# btc oak — full site content > Concatenated markdown bodies of every published page on btcoak.com — globals (homepage, about, sources, methodology, glossary), every chart, every reference landing. Each section is self-contained and includes the page's URL and category. Generated from scripts/llms-content/. --- # btc oak — home URL: https://btcoak.com/ Category: global ## What btc oak is A free, no-login Bitcoin chart reference. 44 indicators across cycles, on-chain, statistical, flows, derivatives, mining, and macro categories — refit nightly from the complete daily Bitcoin history. Transparent methodology, no paywalls, no ads, no email gates. The site exists to replace the canonical free references that have steadily moved behind paywalls and login walls. Every chart is built to teach as well as plot: a TL;DR, the formula, a regime table, a per-cycle historical-readings table, persona framing, named failure modes, a visible FAQ, and primary-source citations. The full chart catalogue is grouped by category on this page; the methodology pillar at [/methodology](/methodology) carries the formula and reconstruction notes per slug. ## Categories - **Cycles** — long-run valuation models (Rainbow, Stock-to-Flow, Power-Law, Pi Cycle, 200W MA Heatmap, Golden Ratio Multiplier, Drawdown, Halving Cycles). - **On-chain** — realized-cap and HODL-cohort indicators (MVRV, MVRV-Z, NUPL, Puell Multiple, SOPR, Reserve Risk, RHODL, Realized Price, AHR999, HODL Waves, Active/New Addresses, Correlations, Fear & Greed). - **Mining** — Hash Rate, Hash Ribbon. - **Statistical** — pure-compute statistical lenses (Daily Issuance, Rolling CAGR, Volatility, Profitable Days, Time in Band, Halving-Cycle Overlay, Monthly Returns, Daily Transactions, Daily Fees, NVT Ratio). - **Flows** — ETF Flows, Exchange Reserves & Net Flow, Coinbase Premium. - **Derivatives** — Funding Rate, Open Interest, Long/Short Ratio, Liquidation Heatmap. - **Macro** — Bitcoin Dominance, Stablecoin Supply, Altcoin Season, BTC vs Global M2. ## Reference pages Four prose-first reference tools sit alongside the chart catalogue: - [Bitcoin halving countdown](/halving-countdown) — live block-anchored countdown to H5. - [Bitcoin halving dates](/halving-dates) — every halving past + projected with date, block, reward, halving-day price, forward-12m return. - [Bitcoin all-time high](/all-time-high) — current ATH, days since, distance, plus per-cycle peak table. - [Bitcoin total supply](/total-supply) — live circulating supply, the protocol-level 21M cap derivation, and the lost-coins range. ## Editorial bar Indicators are lenses, not forecasts. Every page carries a "when it fails" section for a reason. Best read as a portfolio of disagreeing signals; a single chart in isolation is the easiest way to talk yourself into the wrong cycle position. The full editorial standards live in [about](/about); the data provenance lives in [sources](/sources); the index of definitions lives in [glossary](/glossary). ## Refresh cadence All datasets refit nightly from full daily history. Spot price and spot-derived signals (band, regime, deviation) auto-refresh in the browser a few times a day. The bottom of every chart page carries a `Last refreshed` timestamp tied to the nightly pipeline run. --- # About btc oak URL: https://btcoak.com/about Category: global ## Who runs btc oak Josh is a pseudonymous builder who runs btcoak.com — a free, open reference for Bitcoin's most-cited on-chain, cycle, and macro charts. The interest in Bitcoin is narrower than the usual pitch: it is the only asset class where the entire ledger is public, the issuance schedule is fixed, and a careful chart can actually be reproduced from raw data rather than taken on faith. The last few years have been spent shipping public-data tools and design systems for non-technical readers — work that translates directly to the editorial problem here, which is making 44 quantitative charts legible without dumbing them down. ## Why the site exists btc oak exists because the obvious free references stopped being free. The canonical cycle charts moved behind paywalls in 2024; the cleaner derivatives and seasonality views sit behind paid dashboards; most replacements are SEO farms wrapped in affiliate links. This site publishes the same family of models — MVRV, Puell, Pi Cycle, Rainbow, Hash Ribbons, Power-Law and the rest — refit nightly from full daily history, with the methodology written down on every page. Every chart has to teach as well as plot, or it doesn't ship. There are no ads, no email gates, no affiliate links, and nothing on this site is investment advice — these are observations, not instructions. ## Editorial standards Every published page carries a TL;DR (4 labelled rows, no formulas), formula block in §02, regime table, per-cycle historical readings, persona framing, named failure modes, a visible FAQ, and primary-source citations in the commit message. Every numerical reading is computed from the underlying daily series, never hardcoded. Every external link with a factual claim attached is content-verified before merge — status-code 200 is not enough; the rendered page has to match the claim. ## Contact Corrections and disagreements are welcome and read carefully. The fastest channel is email at hello@btcoak.com. Useful: a methodology you'd compute differently, a dataset that looks stale, a chart that misleads at a particular zoom level. Less useful: partnership pitches, token launches, or anything that ends with the word 'synergies'. --- # Data sources URL: https://btcoak.com/sources Category: global ## What this page is The canonical disclosure surface for every upstream feed btc oak draws on. Vendor names appear here and only here; chart pages link to /sources for transparency without inlining provider names in body prose. The site's data philosophy: be transparent about chain of custody; let readers audit if they want to. ## Upstream feeds ### CoinGecko Pro The canonical USD price oracle. Daily Bitcoin closes from 28 April 2013 onward, merged with the pre-2013 CSV (below) into the price spine every chart references. Backs: every chart on the site reads the price spine for spot or model fits. Direct fits include Rainbow, Power-Law, Stock-to-Flow, Pi Cycle, 200-Week MA; the pure-compute statistical charts (drawdown, rolling CAGR, volatility, profitable days, halving-cycle overlay, time-in-band, monthly returns) derive from the same series. Coverage starts 28 April 2013; the pre-2013 CSV covers earlier history. The upstream occasionally emits two points for the same UTC day on refresh; we deduplicate by ISO day, last-write-wins. ### CoinGlass v4 The on-chain, sentiment, derivatives, ETF, dominance, stablecoin and macro feed. 28 charts are built from CoinGlass-only inputs; another three (MVRV, MVRV-Z, HODL Waves) combine its holder-supply series with locally-derived market cap. Substantive divergences from common references: - **Exchange Flows** — coverage starts 1 May 2024; pre-2024 context isn't available. Coinbase appears under two casings upstream; both normalised to one venue label so the per-day breakdown is internally consistent. - **Long/Short Ratio** — no aggregated cross-venue feed exists; the highest-volume venue is shipped as the public-market proxy. - **Funding Rate** and **Coinbase Premium** — rate values are in percent (`0.0584 = 0.0584%`), preserved end-to-end. The Coinbase Premium regime is keyed off the 7-day rolling mean to suppress raw daily outage spikes. - **BTC vs M2** — year-on-year M2 growth is derived locally from the supply series rather than the upstream-published growth field, so both baskets stay on one auditable definition. - **Correlations** — the upstream Pearson series uses aligned trading days, which runs more extreme than research-desk 90-day tables. - **MVRV / MVRV-Z** — realized cap is a two-bucket short-term-holder + long-term-holder approximation of per-UTXO ground truth; tracks within a few percent across the full history. - **NUPL** — pre-exchange zero-fill rows from 2009 are dropped so the Y-axis doesn't compress around zero. ### blockchain.info /charts/\* Hash rate, on-chain transaction count, transaction-fees-USD, and estimated-transaction-volume-USD — four free, fifteen-year-deep daily series that back Hash Rate, Hash Ribbon, Transactions, Fees, and NVT. Substantive divergences: - **NVT** — regime thresholds recalibrated to 100/220 (from Kalichkin's canonical 45/150) because the estimated-transaction-volume-USD denominator runs about twice as high as Kalichkin's adjusted-volume figure; 20/80 percentile split on post-warmup history sits at 100/220. - **Hash Ribbon** — the published 30/60 hash-rate cross without Edwards's price-momentum confirmation filter. The popular price-confirmed variant lags the canonical signal-date list by three to ten weeks; readers wanting the strict-Edwards signal can layer the 10/20 price-SMA cross on top. - **Hash Rate** — pre-mining-rig days where converted hash rate falls below 1 TH/s are dropped to keep the log axis legible. ### Pre-2013 daily closes A small CSV of 1,015 daily UTC closes from 2010-07-18 (Bitcoin's first week of Mt. Gox trading) through 2013-04-27, sourced from the CoinMetrics community data. The CoinGecko window starts 2013-04-28; without the early history the regression would lose Bitcoin's first three years. Newer data wins on overlap. ## Licensing The chart data published on btc oak is released under CC BY 4.0 — reuse is welcome with attribution to `btcoak.com`. Upstream feeds remain subject to their providers' terms; consult each provider's documentation before redistributing raw upstream data. --- # Methodology URL: https://btcoak.com/methodology Category: global ## What this page is The source-of-truth for every formula, window length, and reconstruction choice on the site. One section per chart slug, deep-linkable via `/methodology#` — every chart page links to its own methodology block; the SEO landing pages link similarly. The pillar holds the technical detail so the chart pages stay focused on interpretation. ## Structure The pillar is organised by the same seven chart categories surfaced in the navigation: Cycles, On-chain, Mining, Statistical, Flows, Derivatives, Macro. A separate "Reference" section covers the four prose-first SEO landing pages (Total Supply, Halving Dates, Halving Countdown, All-Time High). Each chart's section names the formula in ``, the window length, the inputs, the reconstruction notes (where btc oak deviates from a published canonical), and any known divergences against third-party references. ## Categories - **Cycles** — long-run valuation models. Rainbow, Stock-to-Flow, Power-Law, Pi Cycle, 200W MA Heatmap, Golden Ratio Multiplier, Drawdown. - **On-chain** — realized-cap and HODL-cohort indicators plus sentiment composites. MVRV, MVRV-Z, NUPL, Puell Multiple, SOPR, Reserve Risk, RHODL, Realized Price, AHR999, HODL Waves, Active/New Addresses, Correlations, Fear & Greed. - **Mining** — Hash Rate, Hash Ribbon. - **Statistical** — pure-compute statistical lenses. Daily Issuance, Rolling CAGR, Volatility, Profitable Days, Time in Band, Halving-Cycle Overlay, Monthly Returns, Daily Transactions, Daily Fees, NVT Ratio. - **Flows** — ETF Flows, Exchange Reserves & Net Flow, Coinbase Premium. - **Derivatives** — Funding Rate, Open Interest, Long/Short Ratio, Liquidation Heatmap. - **Macro** — Bitcoin Dominance, Stablecoin Supply, Altcoin Season, BTC vs Global M2. - **Reference** — Total Supply, Halving Dates, Halving Countdown, All-Time High. ## When pages drift from upstream When btc oak's published number diverges from a third-party reference, the cause is usually one of three: - A different upstream feed (we are transparent about which feeds we draw on — see [/sources](/sources)). - A different smoothing window (NVT Signal's 90-day SMA, our recalibrated NVT thresholds). - A different reconstruction (the two-bucket realized-cap split, where a per-UTXO feed isn't available to us). Where the divergence is material we document it in that chart's section. Otherwise it's well within rounding-error tolerance and reflects a published-canonical difference, not a btc oak error. ## Refresh cadence Every model on btc oak is reproducible from the same daily-close history, the same halving schedule, and the same nightly-refit regression. Numbers refresh nightly; the timestamp at the bottom of every chart page tracks the most recent run. --- # Glossary URL: https://btcoak.com/glossary Category: global ## What this page is A single-page index of every Bitcoin indicator term used across btc oak — definition, related terms, and a link out to the chart that visualises it. Emitted as a `DefinedTermSet` JSON-LD block so AI engines can cross-reference indicators by canonical name and acronym. Roughly 47 entries cover the indicator surface. ## What's in it Every chart on the site has at least one glossary entry. Acronyms (MVRV, NUPL, SOPR, RHODL, NVT) are cross-referenced to their full names; alternate names ("BTC.D" vs "Bitcoin Dominance") are folded under the canonical entry. Concepts that don't have their own chart but appear repeatedly in chart prose (Genesis Block, Realized Cap, STH/LTH cohorts, the 155-day boundary) get standalone entries. ## How it relates to the rest of the site - The methodology pillar at [/methodology](/methodology) carries the formula and reconstruction notes per indicator; the glossary carries the plain-English definition. - Each glossary term links to the chart page that visualises it. From a chart page, an unfamiliar term in body prose can be looked up in one click. - The `DefinedTermSet` JSON-LD block lets AI engines treat btc oak as a canonical entity reference for the indicators it covers — useful for "what is the Puell Multiple" style queries that route through Knowledge Panels. ## Refresh cadence The definitions themselves are reference content — they don't change unless an indicator's canonical definition is contested or revised. The page's `dateModified` reflects the most recent edit pass. --- # The Rainbow Chart URL: https://btcoak.com/rainbow Category: cycles ## What it is The Bitcoin Rainbow Chart is a logarithmic regression of Bitcoin's price against the natural log of days since genesis, overlaid with nine coloured valuation bands. The fit traces back to a 2014 BitcoinTalk thread by user Trolololo; the rainbow-band visualisation came separately from Reddit user azop, and Blockchain Center merged the two into the modern interactive form in 2019. Modern hosts refit the regression against the full daily-close history and treat it as a long-run valuation reference rather than a price forecast. ## How it is calculated btc oak fits `log₁₀(price) = a · ln(days since genesis) + b` by ordinary least squares on every daily close from 18 Jul 2010 to the most recent close. The centre line is that fit; the nine bands are drawn at parallel offsets of −0.4 to +0.4 in log₁₀ space, spaced every 0.1 (one 10× step bottom to top). Band assignment at any date is the offset between the actual log-price and the centre log-price on that day, bucketed at midpoints between successive band offsets. Days are counted from genesis (2009-01-03), not the first-traded date, so the `ln(days)` axis tracks Bitcoin's protocol history rather than its price history. ## How to read it Locate spot on the price axis, read across to today's centre line, and note the log-space offset. Negative offsets put Bitcoin below the regression median in the cool bands (1 to 4) — historically the territory of cycle bottoms and multi-year accumulation. Positive offsets push it above, into the warm bands (6 to 9), where past cycle tops and distribution phases have lived. Band 5 straddles the centre line: the model's "no strong view" regime. ## Historical readings Applying today's fit to every Bitcoin cycle extremum since 2013 surfaces a clear pattern. The 2013 and 2017 peaks saturated band 9 (Euphoric); the 2021 peaks sat in band 9 but at shallower log-space offsets; the 2024 pre-halving high registered only in the upper-mid bands. Each top has landed lower than the last. Cycle anchors btc oak computes against: - 2013-12-04 — 2013 cycle top - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2021-04-14 — 2021 April local top - 2021-11-10 — 2021 November cycle top - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## When it fails The regression slope flattens over time. Held static, the fit over-forecasts every future period — a 2018-vintage Rainbow would have placed the 2024 top in band 9, not in the band it actually printed. Nightly refitting fixes the forward bias and imposes a backward one: historical peaks look more extreme than they did in real time, because the current fit is shallower than the one that was live on the day. Both biases are real. Both are the model's shape, not its failure. The bands are descriptive, not predictive. Trolololo introduced the underlying regression in his 2014 BitcoinTalk thread with an explicit caveat — "just a model, not a crystal ball". Blockchain Center, who hosts the canonical version, frames it as a long-run reference rather than a forecasting tool. The bands are regions on a log-regression plot. They do not imply mean reversion, and they do not imply that a reading inside a band will produce the return a past reading in that band produced. It is one of many power-law-family models. The broader critique of log-regression fits on Bitcoin — laid out by Nico Cordeiro in his 2020 CoinDesk essay on stock-to-flow, and revisited around Giovanni Santostasi's Power-Law Theory — applies here. Bitcoin's long-run price may or may not follow a power law; the Rainbow implicitly assumes it does. ## Frequently asked **What is the Bitcoin Rainbow Chart?** It is a logarithmic regression of Bitcoin's price against the natural log of days since genesis, overlaid with nine coloured valuation bands. The regression came from BitcoinTalk user Trolololo in October 2014; the rainbow-band visualisation came separately from Reddit user azop; Blockchain Center merged the two in 2019. **How is the Rainbow Chart calculated?** btc oak fits `log₁₀(price) = a · ln(days since genesis) + b` by ordinary least squares on every daily close from 18 Jul 2010 to the most recent close. The centre line is that fit; the nine bands are drawn at parallel offsets of −0.4 to +0.4 in log₁₀ space, spaced every 0.1. **Who created the Bitcoin Rainbow Chart?** It is a merge of two 2014 contributions: BitcoinTalk user Trolololo's logarithmic-regression model — described by him as "just a model, not a crystal ball" — and Reddit user azop's rainbow-band visualisation. Blockchain Center stitched them together into the modern interactive form in 2019. **Is the Bitcoin Rainbow Chart accurate?** As a predictive model, no. Held static, the regression has over-forecast future price every time, and its slope has flattened cycle by cycle. Descriptively it earns its keep: more than fourteen years of price history compressed into nine bands, with spot's position relative to trend visible at a glance. Every peak since 2013 has landed in a lower band than the one before. --- # Stock-to-Flow URL: https://btcoak.com/stock-to-flow Category: cycles ## What it is Stock-to-flow is a scarcity-based valuation framework adapted from precious-metals literature. Stock is circulating supply; flow is annual new issuance; the ratio expresses how many years of issuance it would take to replace the existing stock. PlanB (@100trillionUSD), a pseudonymous Dutch institutional investor, applied the framework to Bitcoin in _Modeling Bitcoin's Value with Scarcity_ on Medium on 22 March 2019, with the regression form `ln(market value) = 3.3 · ln(SF) + 14.6` and a reported R² of 95%. PlanB followed in 2020 with an extended Stock-to-Flow Cross-Asset (S2FX) variant carrying higher post-2020 targets. ## How it is calculated Inputs are the daily Bitcoin USD closes paired with a deterministic supply schedule keyed off the four canonical halvings (block heights 210,000 / 420,000 / 630,000 / 840,000 with subsidies 25 / 12.5 / 6.25 / 3.125 BTC respectively). For every day _t_: ``` stock(t) = circulating supply at day t flow(t) = subsidy_t × 144 × 365 S2F(t) = stock(t) / flow(t) ``` The model is a log-log OLS regression of price on S2F: ``` ln(price_i) = a + b · ln(S2F_i) ``` with implied model price `price_model = exp(a) · S2F^b`. PlanB's original 2019 fit reported `b ≈ 3.3` and `a ≈ 14.6` on a market-cap regression; the nightly refit publishes daily coefficients with the latest cycle included. The flow term uses 144 blocks per day (the protocol target, not the realised ~147/day), which trims roughly 2% off the absolute flow figure but preserves the doubling step at every halving. The S2F ratio is daily, not annualised — a smoothed flow window would soften the step at each halving, which is the feature the model is built to highlight. ## How to read it The single read is the residual: the gap between spot and model price. PlanB's regime thresholds were not formally codified, but the convention used in most replications is "extremely undervalued" below −50%, "fair value" within ±25%, and "extremely overvalued" above +100%. | Reading | Regime | What it has meant | | ------------- | --------------------- | ------------------------------------------------------------------------------------------------ | | < −50% | Extremely undervalued | 2015 cycle low and the 2022 post-FTX trough printed deep into this band — −94% in November 2022. | | −50% to −25% | Undervalued | 2018 cycle low and the post-2024-halving lows lived here. | | −25% to +25% | Fair value | The model and realised price have agreed in this band; rare in cycle peaks or troughs. | | +25% to +100% | Overvalued | The 2017 cycle top printed here; the 2013 cycle had multiple visits. | | > +100% | Extremely overvalued | The 2013 Mt. Gox-era top exceeded the model by more than 2×; sustained only in early history. | ## Historical readings Sampling the canonical cycle anchors against the current fit shows the model's behaviour across cycles. The 2013 and 2017 tops printed above the model; the 2018 cycle low printed near it. The 2021 tops are the inflection point — both the April $64,863 peak and the November $68,789 ATH printed sharply below a model that, with S2F at the post-halving step of ≈ 56, was implying values well above $100,000. The 2022 trough at $15,587 represents the model's worst failure on the record, with a residual near −94%. Cycle anchors btc oak computes against: - 2013-04-10 — 2013 April peak - 2013-11-29 — 2013 November peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2021-04-14 — 2021 April peak - 2021-11-10 — 2021 November cycle top - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## The 2021 overshoot, in numbers For the two years following PlanB's March 2019 publication, S2F's model line and Bitcoin's realised price moved in close lockstep. PlanB explicitly forecast in the original piece that the predicted market value for Bitcoin after the May 2020 halving would be $1 trillion — translating to a Bitcoin price of $55,000. Bitcoin reached $55,000 in late February 2021, broadly on schedule. PlanB then followed with the Stock-to-Flow Cross-Asset (S2FX) variant in 2020, which set higher post-2020 average targets. What happened next is the model's defining failure. With S2F stepping to ≈ 56 after the May 2020 halving, the original regression implied a model price near $235,000. The realised April 2021 peak was $64,863. The realised November 2021 ATH was $68,789. The 2022 cycle trough was $15,587 — a residual of roughly −94% against the same model. PlanB's later S2FX variant fared even worse on that metric, with a stated "average price" for the 2020–2024 era of $288,000 against a realised average closer to $40,000. The 2024 halving stepped S2F from ≈ 56 to ≈ 119, which under the original regression implies a forward model price above $1.5 million — the chart's largest residual on record, and the reason most thoughtful operators no longer treat the model as a price target. ## When it fails **The 2021 overshoot is the model's defining failure.** Against a model price of roughly $235,000 at S2F ≈ 56, realised cycle peaks reached $64,863 (April 2021) and $68,789 (November 2021). The 2022 trough at $15,587 was a residual near −94%. The 2024 pre-halving high of $73,738 sat below the model's same pre-2024 implied value of around $100,000. Four consecutive cycle-anchor prints below the model is the empirical refutation; no narrative dressing covers it. **The regression is mathematically circular.** Stock and flow are both functions of time (cumulative issuance and current-block-subsidy respectively), so the regression of price on S2F is nearly a regression of price on time with extra steps. Harold Christopher Burger's March 2022 walkthrough at `hcburger.com/blog/whatsupwiths2f/` catalogues the Level39 framing of this circularity ("Stock is a function of Stock") and shows the adjusted R² collapses toward zero. Nico Cordeiro's June 2020 critique made the same point with a different methodology and arrived at the same conclusion. **Scarcity does not deterministically set price.** S2F assumes the market is a pure scarcity-pricing engine — that doubling the stock-to-flow ratio at every halving deterministically translates to a multi-fold price step. That mechanism plausibly applies to gold and silver, where industrial demand is large and supply is the swing variable. Bitcoin's demand curve is dominated by speculative and institutional flows; supply changes are well-known years in advance, so the market's price response to a halving is forward-discounted. ## The adversarial case Stock-to-Flow has attracted heavier econometric critique than any other mainstream Bitcoin valuation model. **Nico Cordeiro** (Strix Leviathan, 30 June 2020): "A Chameleon Model — Why Bitcoin's Stock-to-Flow Model is Fatally Flawed" argued the model's forecasting accuracy "will likely be about as successful at forecasting Bitcoin's future price as the astrological models of the past." The structural critique — that S2F's flow term is a step function of time, so any model regressing price against it is implicitly regressing price against time — has aged exceptionally well. **Level39 / Harold Christopher Burger** (27 March 2022): Burger's "What's up with S2F?" catalogues the strongest mathematical critique to date — the "Stock is a function of Stock" framing originally posted by the pseudonymous Level39 — that the regression is structurally circular and that adjusting for the look-ahead drops R² toward zero. Burger himself runs a separate diminishing-returns critique: S2F predicts roughly equal-magnitude rallies from each successive halving, while the empirical data shows the opposite (each cycle's peak multiple shrinks). **Eric Wall's catalogue** (1 July 2020): "A list of the greatest blows to the S2F model" catalogues the formal critiques: Sebastian Kripfganz's cointegration analysis (no statistical evidence for the relationship); Marcel Burger's spurious-regression walkthrough; and Nick Emblow's BTConometrics work showing the S2F correlation is "entirely spurious." Emblow's pseudonym is a play on "Bitcoin econometrics," and his S2F work was published under `btconometrics.com`. ## Frequently asked **What is the stock-to-flow model?** Stock-to-flow is a scarcity-based valuation framework adapted from precious-metals literature. Stock is circulating supply; flow is annual new issuance. The ratio expresses how many years of issuance it would take to replace the existing stock. PlanB applied the framework to Bitcoin in _Modeling Bitcoin's Value with Scarcity_ (Medium, 22 Mar 2019), with the regression form `ln(market value) = 3.3 · ln(SF) + 14.6`. **Who created Bitcoin Stock-to-Flow?** PlanB (@100trillionUSD), a pseudonymous Dutch institutional investor, published the original Bitcoin S2F adaptation on Medium on 22 March 2019. The piece reported R² of 95% and predicted a market value of approximately $1 trillion (≈ $55,000 per BTC) for the post-May-2020 halving era. PlanB also published an extended Stock-to-Flow Cross-Asset (S2FX) variant in 2020 with higher post-2020 price targets. **Is the stock-to-flow model accurate?** Descriptively over Bitcoin's first decade, yes — the regression had a high R² and the model price tracked realised price within roughly an order of magnitude through 2020. From 2021 forward the model has materially diverged: against a model price of around $235,000 at S2F ≈ 56, realised cycle peaks reached only $64,863 (April 2021) and $68,789 (November 2021), a residual near −71%. The 2022 trough of $15,587 was a residual near −94%. **When are Bitcoin halvings?** Bitcoin halves the block subsidy every 210,000 blocks — roughly every four years. The four halvings to date: 28 November 2012 (block 210,000, subsidy 25 BTC), 9 July 2016 (block 420,000, 12.5 BTC), 11 May 2020 (block 630,000, 6.25 BTC), 19/20 April 2024 (block 840,000, 3.125 BTC). The next halving is scheduled at block 1,050,000 in early 2028. **Is stock-to-flow debunked?** Several rigorous critiques agree the model has fundamental problems. Nico Cordeiro called it "a chameleon model" that fits historical data without predicting forward (June 2020). Harold Christopher Burger showed in 2022 that "Stock is a function of Stock" — i.e., the regression is mathematically circular — and that adjusting for the look-ahead structure drops R² to zero. Eric Wall, Sebastian Kripfganz, and Nick Emblow have catalogued additional cointegration and spurious-regression problems. --- # Power-Law Corridor URL: https://btcoak.com/power-law Category: cycles ## What it is The Bitcoin power law is a long-run regression of price against time on a log-log axis. The fit takes the form `price = A · days^n`. On the chart it appears as a near-straight line through Bitcoin's price history when both axes are logarithmic, with two parallel σ-bounds drawing the upper and lower edges of the corridor. The framework was first published in blog form by Harold Christopher Burger in September 2019, and given a separate theoretical argument by astrophysicist Giovanni Santostasi in March 2024. ## How it is calculated Inputs are a single series: every Bitcoin daily close, paired with a deterministic day-count from the Bitcoin genesis block on 2009-01-03. The regression is: ``` log₁₀(price_i) = n · log₁₀(days_i) + log₁₀(A) + ε_i ``` Fit by ordinary least squares on every observation since the first traded daily close. The corridor's upper and lower bounds are the fitted median shifted by ±σ-multiplier in log space, where σ is the standard deviation of the residual `log₁₀(price) − (n · log₁₀(days) + log₁₀(A))`. The default σ-multiplier of 1.5 corresponds to a corridor that bounds roughly 87% of the historical observations on a normal-residual approximation. Coefficients refit nightly against the full daily-close history. ## How to read it Locate spot on the price axis, read across to today's median, and note the log-residual. Negative residuals put Bitcoin below the corridor median — historical accumulation territory. Positive residuals push spot into the upper band, where past cycle tops have lived. The corridor's edges are not hard floors or ceilings — Bitcoin has briefly traded outside on both sides — but excursions outside have been brief in absolute days and small in absolute log-distance. | Reading | Regime | What it has meant | | ----------------------- | --------------- | ------------------------------------------------------------------------------------------ | | < 15% (near support) | Near support | The lower band. 2015, 2019, 2022 cycle troughs all touched here. | | 15% – 40% (lower band) | Lower band | Below-trend value. Bear-recovery and post-trough accumulation regimes. | | 40% – 60% (mid-range) | Mid-range | On the median line. No directional conviction from this chart alone. | | 60% – 85% (upper band) | Upper band | Above-trend extension. Historically bull-cycle territory; first distribution signals here. | | > 85% (near resistance) | Near resistance | The upper band. 2013, 2017, 2021 cycle peaks all touched here. 2024 pre-halving did not. | ## Historical readings Sampling the canonical cycle anchors against today's fit surfaces the corridor's defining property: every cycle peak has lived in the upper band and every cycle trough has lived near support, even as the absolute prices that delivered those bands span four orders of magnitude. The 2024 pre-halving high is the most muted on the table — it cleared the median but did not reach the upper band of the current fit. Cycle anchors btc oak computes against: - 2013-04-10 — 2013 April peak - 2013-11-29 — 2013 November peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2021-04-14 — 2021 April peak - 2021-11-10 — 2021 November peak - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## Burger vs Santostasi The two canonical primary sources offer different exponents — empirical and theoretical. Harold Christopher Burger published _Bitcoin's natural long-term power-law corridor of growth_ on 3 September 2019 with the empirical OLS constants `a = −17.01593313` and `b = 5.84509376` at `hcburger.com/blog/powerlaw/`. Giovanni Santostasi argues separately for a _theoretical_ exponent of `n = 6`, derived from a scaling-law argument (`users ∝ t³`, Metcalfe-style `price ∝ users²`, hence `price ∝ t⁶`) in _The Bitcoin Power Law Theory_ (Medium, 20 March 2024). The empirical fit and the theoretical derivation differ by roughly 3% on the exponent — a meaningful gap, not equivalence. Santostasi's piece motivates the empirical fit by analogy to allometric scaling in biology and physics — Geoffrey West's universal ¾-power scaling of metabolic rate with body mass, and the broader power-law families that show up in city-size distributions and river-network topologies. What makes the analogy useful is what it disciplines: a power law fit through any growth process will look good on a log-log axis; the question is whether the generating process actually has the scale-free properties biological allometry relies on. Neither author delivers the other: the empirical slope is what the regression spits out; the theoretical slope is what the scaling argument says it _should_ be. A fit dressed up as physics is harder to reject than a fit honestly framed as a fit. The nightly refit follows Burger's empirical method against the full daily-close history. ## The exponent has flattened over time An honest reading of the chart includes the slow drift in the fit itself. Burger's 2019 publication used `n = 5.84509376`; the nightly refit currently sits a touch lower. The exponent has flattened — not by much, but visibly. Each refit incorporates more low-volatility post-2018 time and proportionally less Mt. Gox-era doubling, which mechanically lowers the slope. The drift is the model adapting honestly. It is also the seed of the second-largest criticism of the framework: a model whose parameters drift cycle by cycle and whose historical residuals are recomputed under the latest fit will always look better retrospectively than it did in real time. ## When it fails **The high R² is partly an artefact of regressing two trending series.** Tim Stolte of Amdax laid out the formal critique in 2022 (Medium, 2 Sep 2022): "Logarithmically scaling time is possibly the weirdest thing I have ever seen in time series analysis" and "we are dealing with a so-called spurious regression… Making something look nice and familiar doesn't mean that it's useful or legit." When both axes share a deterministic growth component (time on the x-axis, log-price tracking time roughly on the y-axis), classical OLS R² inflates without telling you the relationship has predictive content. The corridor's R² is high; that fact alone is not the proof of quality it appears to be. **Early-history prints carry disproportionate weight.** Marty Kendall laid out the issue at mNAV Insights: "the earliest data has a massively more influential effect because each decade occupies the same width on a log axis." On a log-time axis, the years 2010–2012 occupy the same horizontal width as 2015–2018 or 2018–2024. The OLS slope is anchored heavily by the Mt. Gox-era prints — a regime that no longer characterises the asset and that any honest model would weight down. **Power laws don't survive forever.** Bitcoin's adoption and hash-rate growth have plausibly fit a power law for fifteen years. There is no reason to expect the fit to persist as market cap approaches a meaningful share of global financial assets, sovereign adoption changes the order-flow regime, or regulatory shifts re-anchor the demand curve. Lyn Alden has framed the same observation as a shift from issuance-cycle dynamics to liquidity-cycle dynamics — a structural change the corridor cannot accommodate without a major refit. ## Frequently asked **What is the Bitcoin power law?** A long-run regression of price against time on a log-log axis. The fit takes the form `price = A · days^n`; Burger's 2019 empirical fit gives _n_ ≈ 5.84, while Santostasi argues for a theoretical _n_ = 6 derived from scaling-law reasoning. On the chart it appears as a near-straight line through Bitcoin's price history when both axes are logarithmic. First published by Harold Christopher Burger in September 2019, and given a separate theoretical statement by Giovanni Santostasi in March 2024. **Who created the Bitcoin power law?** There are two legitimate primary sources. Harold Christopher Burger published _Bitcoin's natural long-term power-law corridor of growth_ on hcburger.com on 3 September 2019, with explicit log-log regression coefficients (a = −17.01593313, b = 5.84509376). Giovanni Santostasi's _The Bitcoin Power Law Theory_ on Medium, dated 20 March 2024, is the canonical theoretical statement, anchoring the empirical fit in scaling-law arguments. There is no peer-reviewed Santostasi-authored paper at this time despite occasional secondary citations to one. **How is the Power-Law Corridor drawn?** The median line is the OLS fit of `log₁₀(price) = n · log₁₀(days_since_genesis) + log₁₀(A)`. The corridor's upper and lower edges are the median shifted by ±σ-multiplier of the residuals — bounding roughly 87% of historical observations on a normal-residual approximation. Coefficients refit nightly against the full daily-close history. **Is the Bitcoin power law accurate?** Descriptively, the fit explains roughly 96% of the variance in log-price across fifteen years of daily data. As a forecast, it is much more contested: Tim Stolte called the methodology "a so-called spurious regression" and warned that "logarithmically scaling time is possibly the weirdest thing I have ever seen in time series analysis." Critics also note that the regression is heavily weighted by early-history prints because each decade occupies the same width on a log axis. Read it as a long-run shape, not a precise level forecast. **When does the power law predict a million-dollar Bitcoin?** Direct extrapolation of today's fit places `$1M Bitcoin` in the late 2030s — but the band on either side is wide, and the fit's slope itself is a rolling target (it has flattened slightly cycle by cycle as new data arrives). A $1M arrival date pulled off the corridor is a model output, not a market forecast. --- # Pi Cycle Top URL: https://btcoak.com/pi-cycle Category: cycles ## What it is The Pi Cycle Top Indicator is a moving-average-cross signal designed to identify Bitcoin cycle peaks. It plots the 111-day simple moving average against twice the 350-day simple moving average; when the shorter line crosses above the longer, history says the cycle is at or near a top. Created by Philip Swift in April 2019, the indicator has fired four times — April 2013, December 2013, December 2017, and April 2021 — and missed twice (November 2021 and the entire 2024–2025 cycle so far). The ratio 350 ÷ 111 ≈ 3.153 — the closest two-integer approximation of π using small whole numbers — gave the indicator its name. ## How it is calculated Inputs are the daily Bitcoin USD closes. For every day _t_: ``` 111DMA(t) = mean(price[t−110..t]) 350DMA(t) = mean(price[t−349..t]) trigger(t) = 2 × 350DMA(t) ``` A signal is recorded on day _t_ when `111DMA(t−1) ≤ trigger(t−1)` and `111DMA(t) > trigger(t)` — i.e. the faster line was at or below the slower-times-two line on the previous day, and is above it today. Crossings in the reverse direction are not signals; this is a one-directional indicator. Some implementations use the daily-close ratio `111DMA ÷ 350DMA` against a fixed threshold of 2.0 instead of the crossover. Mathematically equivalent at the trigger; the crossover form is the author-canonical one. ## How to read it The single read is the gap between the two lines. When the 111-day MA is far below twice the 350-day MA, the cycle is in expansion; when it tightens to within a few percent, the indicator is on the watch list; when it crosses above, the historical pattern says the cycle is topping. The four regime labels map to gap thresholds that capture the typical run-up to a cross. | Reading | Regime | What it has meant | | ----------------------- | ---------- | ----------------------------------------------------------------------------------------------------- | | gap ≤ −30% | Far | Mid-cycle. The 111-DMA is well below the trigger; the next signal is at least months away. | | −30% to −5%, widening | Diverging | The gap is opening as price reverses or chops. No imminent signal. | | −30% to −5%, tightening | Converging | The gap is closing as price accelerates. Watch list — every prior signal traversed this regime first. | | \|gap\| < 5% | Signal | Within striking distance. Historical signals have all fired inside this band. | ## Historical readings Emitting one row per published crossover surfaces the indicator's full history at a glance. For each cross the daily close on the cross date is read, the highest price in the 365 days that follow is located, and the drawdown from the cross over the next 200 days is computed. The gap from cross to cycle peak is what makes Pi Cycle famous; the post-cross drawdown is what made it actionable. Crossover anchors: - 2013-04-09 region — first historical cross, days from cross to peak in the single digits - 2013-12-08 region — clean call ahead of the November 2013 close-print top - 2017-12-17 — within three days of the cycle close-print top - 2021-04-14 — caught the April 2021 local peak of $64,863 ## The 2024–2025 misses The headline question for any current reader of Pi Cycle is the one the indicator itself does not answer cleanly: did the 2024 cycle peak come and go without a signal? Inside the run-up to the 14 March 2024 pre-halving high — defined as the maximum value of `(111DMA − 2·350DMA) / (2·350DMA)` observed in the window — the closest the 111-day MA came to its 350-day-×-2 trigger was several percentage points below crossing. Bitcoin printed a new all-time high without the indicator firing. The October 2025 ATH window tells the same story. Closest approach again fell short of the trigger; the realised peak printed without a Pi Cycle signal. Two ATH windows in a row without a Pi Cycle signal — the first time that has happened in the data. Matt Crosby formalised the broader observation in late 2025: "the Pi Cycle Top Indicator failed to provide precise timing or price signals despite being closely watched by many traders" (Crosby, Dec 2025). The canonical reference page now carries the qualifier directly: the indicator "may cease to be relevant in this new market structure." ## The 2021 cycle: only half a signal The 2021 cycle is the only one Pi Cycle has caught in part. The first peak window printed at $64,863 on 14 April 2021; the indicator crossed within days, calling the top of that first run-up cleanly. Bitcoin then fell roughly 54% to a $30k summer trough before climbing back to a fresh ATH at $68,789 on 10 November 2021. The 111-day MA never re-crossed the 350-day-×-2 trigger on the second push higher. The structural explanation is that 2021 was the first cycle with a bifurcated peak structure rather than a single blow-off. After the April top reset the moving averages, the 111-day MA could not re-accelerate fast enough to break above 2 · 350-day MA on the second leg — the second leg was higher in price but lower in momentum. ## When it fails **The 2024 and 2025 misses are the chart's signature failure mode.** Two consecutive ATH windows have come and gone without a cross. Whether that means the indicator's calibration is broken, or the cycle has not topped yet, is the live question. The canonical reference page now hosts a direct walk-back: the indicator "has worked during Bitcoin's adoption growth phase, the first 15 years or so of Bitcoin's life. With the launch of Bitcoin ETF's and Bitcoin's increased integration into the global financial system, this indicator may cease to be relevant in this new market structure." **The 111-day and 350-day windows are post-hoc fits.** Philip Swift selected them because they aligned with the prior cycle peaks — not because of any structural derivation. The π-trivia is rounding: 350 ÷ 111 ≈ 3.153. The same caveat applies to every band-and-window indicator on the site, but Pi Cycle's reliance on two narrow window sizes makes the recalibration risk especially visible when the signal misses. **It is silent on cycle bottoms.** Pi Cycle does not have a paired bottoming signal — the down-cross direction is not a published trigger. A reader using it for accumulation timing has to pair it with something else. Pi Cycle answers the cycle-top question; bottom timing belongs to a different chart. ## Frequently asked **What is the Pi Cycle Top indicator?** A moving-average-cross signal designed to identify Bitcoin cycle peaks. It plots the 111-day simple moving average against twice the 350-day simple moving average; when the shorter line crosses above the longer, history says the cycle is at or near a top. Created by Philip Swift in April 2019. The indicator has fired four times — April 2013, December 2013, December 2017, and April 2021 — and missed twice (November 2021 and the entire 2024 cycle so far). **Why is it called "Pi Cycle"?** The ratio 350 ÷ 111 ≈ 3.153 — the closest two-integer approximation of π using small whole numbers. Swift selected the 111-day and 350-day windows because they fit the prior cycle tops; the π trivia is a coincidence of arithmetic, not a derivation, but it gave the indicator its name. **Did the Pi Cycle predict the 2021 top?** It fired cleanly at the April 2021 local peak of $64,863 — within days of the close-print top of that first peak window. It did **not** fire at the actual cycle high of $68,789 in November 2021. The 111-day MA never re-crossed the 350-day-×-2 line on the second push higher; the 2021 cycle is the first one with a divided peak the indicator caught only in part. **Did Pi Cycle fire at the 2024 Bitcoin top?** No. The 14 March 2024 pre-halving high of $73,738 came and went without a Pi Cycle cross. The closest approach during the run-up got to within a few percent of crossing but never quite met the trigger. Matt Crosby summarised the broader cycle pattern in late 2025: "the Pi Cycle Top Indicator failed to provide precise timing or price signals despite being closely watched by many traders." **Is the Pi Cycle still useful?** It is useful as a confirmation indicator with a known and disclosed failure mode. The cross has identified the exact peak day in three of Bitcoin's prior cycles and the first peak of the bifurcated 2021 cycle. It missed the November 2021 top and has not fired through the 2024–2025 cycle so far. Read it as part of a panel of cycle indicators, not on its own. --- # 200-Week Moving Average Heatmap URL: https://btcoak.com/200w-ma Category: cycles ## What it is The 200-Week Moving Average Heatmap plots Bitcoin's price coloured by the four-week percentage change of its 200-week simple moving average. Cold colours (purple, blue) mark days where the long-run trend is flat or barely rising — historically cycle bottoms. Hot colours (orange, red) mark days where the long-run trend is accelerating — historically late-cycle blow-off territory. The original was published by PlanB (@100trillionUSD), the same pseudonymous analyst behind the Stock-to-Flow model, in January 2019. Philip Swift built a structurally similar but distinct tool — the 2-Year MA Multiplier — at around the same time; the two are often confused in secondary sources. ## How it is calculated Inputs are the daily Bitcoin USD closes. For every day _t_: ``` 200WMA(t) = mean(price[t−1399..t]) delta_4w(t) = (200WMA(t) − 200WMA(t−28)) / 200WMA(t−28) × 100 ``` The first row with a fully-formed 200WMA lands on day 1,400 of the price history (18 May 2014); rows before that are excluded from the chart. The four-week rate of change is unitless (percentage per four weeks) and is the value the heatmap maps to colour. Thresholds at 2%, 4%, 6%, 8%, 10% bracket the historical bottoming and topping windows but are _not_ author-canonical — PlanB's original 2019 post described the colour mapping as month-on-month percentage change without publishing exact breakpoints. The window is exactly 1,400 daily closes, not 200 × (5- or 7-trading-day weeks): Bitcoin trades every day, so calendar weeks and trading weeks coincide. The rate-of-change window is four calendar weeks (28 days), again to keep the metric calendar-aligned rather than market-day-aligned. ## How to read it Two readings carry the chart. The first is spot's distance from the 200-week MA — the price-to-MA ratio. Below 1.0× has historically tagged the start of an accumulation window; staying within 1.0–2.0× is the typical mid-cycle envelope; past 4× is late-cycle territory. The second is the colour band: cold means the MA itself is flat, which is structurally consistent with a bottoming phase; hot means the MA is being repriced upward fast, which is structurally consistent with late-cycle acceleration. | Reading | Regime | What it has meant | | ------------ | ------------- | -------------------------------------------------------------------------------------------------- | | ≤ 2% / 4w | Cold (purple) | The 200-week MA is flat or near-flat. Every prior cycle bottom (2015, 2019, 2023) printed here. | | 2 – 4% / 4w | Cool (blue) | Early-bull recovery. The MA is starting to lift, often a year before the next ATH. | | 4 – 6% / 4w | Mid (green) | Mid-cycle expansion. Trend continuation regime; little informational signal either way. | | 6 – 8% / 4w | Warm (yellow) | Strong-trend phase. The MA is rising fast enough that even a year-long horizon is repriced upward. | | 8 – 10% / 4w | Hot (orange) | Late-cycle acceleration. 2017 and 2021 both touched this band before topping; 2024 did not. | | > 10% / 4w | Red | Cycle-top blow-off. Historically a sell zone; reached only in late 2013 and late 2017. | ## Historical readings Sampling the seven canonical cycle anchors against the rate-of-change pane shows the pattern directly. Cycle bottoms cluster in the cold-purple band, cycle tops cluster in the warm bands; the 2024 pre-halving high in March stayed below the late-cycle red that 2017 and 2021 each visited, an early signal of the muted cycle the data has gone on to print. Cycle anchors btc oak computes against: - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2021-04-14 — 2021 April local top (first ATH window) - 2021-11-10 — 2021 November cycle top - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## The shape of the moving average The 200-week MA grows monotonically in log space across nearly the entire price history. The few exceptions — brief flatlines in mid-2015, late-2018, and mid-2022 — are themselves cycle-bottom signals. A 1,400-day window is long enough to absorb any individual quarter; the MA can only flatline if the rolling window has dropped a higher-priced run from the back end while the front end has been treading water. That co-occurrence has only happened at cycle troughs. The MA can only accelerate to the "hot" bands when the front-end of the window is meaningfully higher than the equivalent prior-cycle prints rolling off. Late 2013 and late 2017 both saw four-week MA growth above 12%; April 2021 came close (just under 9%); November 2021 again did not quite reach the historical "red" threshold, and March 2024's pre-halving high passed without the MA's growth even reaching 6%. ## The post-ETF blind spot One observation worth surfacing on its own: the 2024 cycle has yet to print a "hot" band. The 14 March 2024 pre-halving high registered around 4% / 4w — squarely inside the cool-mid band. The October 2025 ATH that followed printed at a rate of change in the same range. By the standard the prior three cycles set, that is a quiet cycle; an indicator that historically "flashed sell" at every cycle top has not yet flashed at this one. Whether that means the cycle has not topped or that the heatmap's calibration no longer fits the post-ETF era is the live question. Both readings — the price line and the underlying rate of change — surface so a reader can form their own view. ## When it fails **The colour thresholds are not author-canonical.** PlanB's January 2019 framing — "depending on the month-by-month % increase of the 200 week moving average, a colour is assigned to the price chart" — describes the colour mapping qualitatively. It does not publish exact percentage breakpoints. Different implementations therefore disagree about where "orange" ends and "red" begins. The btc oak implementation exposes the underlying rate-of-change pane so a reader can compare the same date across implementations. **Slow by construction.** A 1,400-day window cannot react to a cycle change in fewer than weeks. The heatmap is a confirmation indicator, not a timing one — bottom prints sit on the cold band for months, top prints often spend multiple weeks on the warm bands before reversing. Matt Crosby summed up the broader 2024 case in late 2025: long-cycle indicators "remained untested" this cycle (Crosby, Dec 2025). The 200WMA's silence is consistent with the category-wide silence of long-window cycle indicators. **The Bitcoin cycle may no longer be a halving cycle.** Lyn Alden's reframing of the Bitcoin cycle as a global-liquidity cycle rather than a halving cycle (via Crypto Briefing, 2025) implies that the long-window cycle indicators built around the four-year cadence — this one, Pi Cycle, the Golden-Ratio Multiplier — may need recalibration as ETFs and regulation reshape the demand cycle. The 200WMA's silence in 2024–2025 is a data point in that direction. ## Frequently asked **What is the 200-week moving average heatmap?** The 200-Week Moving Average Heatmap plots Bitcoin's price coloured by the four-week percentage change of its 200-week simple moving average. Cold colours (purple, blue) mark days where the long-run trend is flat or barely rising — historically cycle bottoms. Hot colours (orange, red) mark days where the long-run trend is accelerating — historically late-cycle blow-off territory. The original was published by PlanB in January 2019. **Who created the 200-week moving average heatmap?** PlanB (@100trillionUSD), the same pseudonymous analyst behind the Stock-to-Flow model, published the 200-Week MA Heatmap in January 2019. Philip Swift built a structurally similar but distinct tool — the 2-Year MA Multiplier — at around the same time; the two are often confused in secondary sources. **Why 200 weeks?** 1,400 calendar days of daily closes is roughly four years — close to one full Bitcoin halving cycle. A moving average over that window absorbs short-term volatility and captures the long-run growth slope without being so long that it dilutes cycle structure. Shorter MAs (50d, 200d) react too quickly to be a cycle lens; the 100-week MA loses the bottoming texture that makes the heatmap legible. **Has Bitcoin ever traded below its 200-week MA?** Yes, briefly, on three occasions. The early-2015 low pierced the 200-week MA at roughly $200; the December 2018 low touched the MA at $3,200 (price-to-MA ratio of about 0.97×); the November 2022 post-FTX low traded well below the MA, with a ratio of about 0.7× — the deepest below-200WMA print on record. Each instance has so far marked a multi-month accumulation window before the next leg up. **Are the heatmap colour bands canonical?** PlanB's original post described the heatmap as month-on-month percentage change of the 200-week moving average mapped to a colour scale, but did not publish exact percentage thresholds. The breakpoints used here (2%, 4%, 6%, 8%, 10%) are conventional and bracket the historical bottoming and topping windows; they are not author-set. The underlying 4-week rate of change is exposed directly so a reader can audit any colour assignment against the data. --- # Fear & Greed Index URL: https://btcoak.com/fear-greed Category: on-chain ## What it is A 0–100 daily sentiment composite for the crypto market, modelled on the older equity-market gauge and active since the first daily print on 1 February 2018. Zero is maximum fear; one hundred is maximum greed. Six components combine with fixed weights into the headline scalar, with four named regimes split at 25, 50, and 75. ## How it is calculated | Component | Weight | What it scores | | ------------------------ | ------------ | ------------------------------------------------------------------------------------ | | Volatility | 25% | Current 30-day realised volatility and max drawdown vs trailing 30 / 90-day averages | | Market momentum / volume | 25% | Volume and momentum vs the same trailing windows | | Social media | 15% | Post and interaction rates on bitcoin hashtags | | Surveys | 15% (paused) | Weekly market-sentiment survey, currently inactive | | Dominance | 10% | Bitcoin's share of total cryptoasset market cap | | Trends | 10% | Search-volume change on bitcoin-related queries | Five of the six components are price- or attention-derivative; only the survey component would have provided an independent input, and that component is currently paused. The structural consequence is that the index moves with price more often than it moves before price. ## Regimes | Reading | Regime | What it has meant | | ------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 0–24 | Extreme fear | High volatility, sharp drawdowns, fearful search activity. Has bracketed (not exactly marked) the November 2018, March 2020, June 2022 (Three-Arrows / Luna), and November 2022 (FTX) stress windows. | | 25–49 | Fear | Cautious tape, no panic. Most cycle drawdowns spend at least some weeks here. | | 50–74 | Greed | Constructive tape. Default range during sustained uptrends. | | 75–100 | Extreme greed | Crowded long. Has fired around April and November 2021, March 2024 pre-halving, and shorter post-ETF windows. | ## The forward-return cluster For every daily Extreme Fear (≤25) and Extreme Greed (≥75) print since the series began, median 30 / 90 / 180-day forward BTC returns are computed against the daily-close history. The cluster table refreshes every night. The distribution: median forward return after Extreme Fear has been positive on 90-day and 180-day horizons — "buy fear" has held in the data, with caveats. Median forward return after Extreme Greed has been negative on the 90-day horizon — "sell greed" carries less force than the slogan implies, because Extreme Greed regimes can persist for weeks before a topping inflection. The spring 2021 run held above 75 for nearly two months. ## When it fails **The index is coincident, not predictive.** Five of six components are price-derivative; the survey component that would provide an independent input is currently paused. Sustained Extreme regimes can persist for months, and the index will follow rather than lead the next inflection. The November 2022 post-FTX stress window is the canonical case — sustained single-digit prints for weeks before the cycle low actually printed. **The single-day print whips on event noise.** A liquidity flush, a regulatory headline, or a macro print can produce one-day swings of 20 points that retrace the next session. The 30-day average is the slower regime read. **The composite hides component disagreement.** Many useful regimes — rising volatility with falling volume, rising dominance with falling momentum — collapse into a single index value that obscures the disagreement. Reading the index without checking the components is reading the summary statistic without checking the distribution. **The methodology is opaque in its weights but firm in its 0–100 output.** The index publishes daily values without daily component breakouts; the weights are public, but per-component daily contributions are not. A reader auditing why a particular day printed 47 rather than 53 cannot reproduce the calculation from the public methodology alone. ## Frequently asked The lowest reading on the daily series is **6**, printed inside the June 2022 Three-Arrows / Terra stress window. The November 2018 cycle-low region produced single-digit prints (a **9** on 25 November 2018 marked the multi-year low for that cycle). The Extreme Fear regime has historically clustered near, not exactly at, cycle bottoms — useful as a regime confirmation, weak as a buy trigger. Extreme Greed regimes have persisted for weeks at a time before topping inflection, especially in spring 2021 and around the March 2024 pre-halving high. --- # Puell Multiple URL: https://btcoak.com/puell-multiple Category: on-chain ## What it is The Puell Multiple divides daily Bitcoin coin-issuance value (in USD) by its 365-day moving average. It captures whether miners are earning significantly more or less than their recent yearly trend, providing a miner-revenue lens on Bitcoin cycle regimes. The metric was introduced by David Puell on 30 March 2019 in his Medium essay _The Puell Multiple_. It was originally framed as a way to characterise miner-supply-side pressure across cycles, and is part of a broader miner-revenue framework Puell developed at Adaptive Capital. ## How it is calculated The formula is short: ``` Puell(t) = Issuance_USD(t) / SMA365(Issuance_USD) ``` Daily issuance USD is the deterministic product of three numbers: today's block subsidy in BTC (from the canonical halving schedule), the number of blocks mined in the trailing 24 hours, and the day's spot price. The denominator is the 365-day simple moving average of that same daily issuance USD. The result is a unitless multiple where 1.0 means "today's issuance value matches the trailing yearly average" — that is, miners are earning at trend. Why a 365-day window. Bitcoin's halving schedule produces a mechanical four-year-period cycle in absolute issuance, with subsidy halving every 210,000 blocks (roughly four years). A 365-day moving-average smooths most of the day-to-day noise while still letting the ratio respond meaningfully to price moves that outpace the slow halving-driven baseline. A shorter window would track price too closely; a longer window would miss the cycle structure. What the multiple captures. When daily issuance value runs far above its yearly trend, the inference is that price has appreciated faster than the miner-cost baseline can adjust — a classic blow-off shape. When it falls far below, the inference is that price has collapsed faster than the baseline can adjust, often to the point that some marginal miners go bankrupt or pause operations. Both regimes are mechanically self-correcting: blow-offs end with miner profit-taking pressure, capitulations end with hashrate-floor consolidation forcing a supply-side shift. ## How to read it Puell is most informative at the extremes. Sustained sub-0.5 readings have only fired at cycle bottoms — historically only true during the aftermath of 2011, mid-2015, late 2018, and late 2022. Each of these windows preceded a multi-year bull cycle. Sustained readings above 4.0 have only fired at the 2013 and 2017 cycle peaks; later peaks have arrived lower. | Reading | Regime | What it has meant | | ----------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ | | Puell ≤ 0.5 | Capitulation | Daily issuance value below half its yearly trend. Has fired at every cycle bottom: 0.31 (Jan 2015), 0.30 (Dec 2018), 0.43 (Mar 2020), 0.36 (Dec 2022). | | 0.5 < Puell ≤ 1.2 | Accumulation | Below long-run average but above capitulation. The band where mid-cycle bottoms fade into the next expansion. | | 1.2 < Puell ≤ 2.5 | Mid-cycle | Bitcoin spends more days here than in any other band. Useful as a low-conviction baseline; weak signal on its own. | | 2.5 < Puell < 4.0 | Elevated | Late-cycle expansion. Both modern peaks lived here: April 2021 at 3.46 and March 2024 at 2.44. | | Puell ≥ 4.0 | Blow-off zone | Puell's original cycle-top reference. Last fired December 2017 at 6.62. The 2013 peaks topped at 10.49 and 9.29. | ## Historical readings Reading every canonical cycle anchor against the live series surfaces the same regime-decay pattern that drives the rest of the realized-cap-family extremes. Cycle peaks at 10.49, 9.29, 6.62, 3.46, 1.95, 2.44. Cycle troughs have held a remarkably tight band: 0.31, 0.30, 0.43, 0.36 — all four bottoms within ±0.07 of the 0.36 mean. Cycle anchors btc oak computes against: - 2013-04-09 — 2013 Apr peak - 2013-11-29 — 2013 Nov peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-07 — 2018 cycle low - 2020-03-18 — 2020 Covid low - 2021-02-19 — 2021 Apr peak - 2021-10-25 — 2021 Nov peak - 2022-12-24 — 2022 cycle low (post-FTX) - 2024-03-11 — 2024 pre-halving high ## The 4× ceiling decay The cleanest way to see the regime shift on this chart is the per-cycle Puell peak, in order: | Cycle | Peak | Hit 4.0? | | ------------- | ----- | -------- | | Apr 2013 peak | 10.49 | Yes | | Nov 2013 peak | 9.29 | Yes | | Dec 2017 peak | 6.62 | Yes | | Apr 2021 peak | 3.46 | No | | Nov 2021 peak | 1.95 | No | | Mar 2024 peak | 2.44 | No | Three of six cycle peaks on record cleared 4.0 — all of them in the pre-2018 era. April 2021 at 3.46 came reasonably close but missed. November 2021 at 1.95 fell short by more than half. March 2024 at 2.44 fell short by nearly half. The trend is monotonic since 2017. Two structural causes drive the decay. First, the halving compresses the multiple mechanically. The supply-side denominator halves every four years; if price appreciates the same multiple cycle-over-cycle, Puell reads roughly the same. But because price has appreciated by a smaller multiple in 2021 (10×) and 2024 (~8×) compared to 2017 (~20×) and 2013 (~80×), Puell has compressed in lockstep with the diminishing-returns price multiple. Second, the ETF-era miner economy is structurally more efficient. Modern industrial miners hedge production on derivative markets, custody-borrow against their stack to fund operations, and treasury-finance through equity rather than spot sales. Daily issuance hitting the open market at any given moment is therefore a smaller share of total daily volume than in 2013–2017 — Puell still measures issuance value, but issuance value as a market-pressure signal has weakened. The 2024 cycle's structural inflows from US spot Bitcoin ETFs (launched January 2024) further diluted miner sells as a fraction of net market demand. Lost coins do not affect Puell directly — the metric measures issuance, not circulation. But Patoshi-era hashrate dynamics did. Sergio Demian Lerner's research on the Patoshi mining pattern documents how a single dominant miner produced roughly 1.1 million BTC in the first ~14 months of the chain, and how that miner stopped voluntarily in May 2010. Cross-cycle comparisons before 2012 should be treated with caution. ## When it fails The 4.0 ceiling has stopped firing. Three of six cycle peaks cleared it cleanly (2013-Apr at 10.49, 2013-Nov at 9.29, 2017-Dec at 6.62). Three did not (2021-Apr at 3.46, 2021-Nov at 1.95, 2024-Mar at 2.44). The threshold was anchored on the 2010s blow-off shape and is now well outside the modern operating range. A "wait for Puell > 4" rule would have missed two consecutive cycles. Halving-window distortion. The 365-day denominator smooths the halving transitions but does not eliminate them. The index mechanically deflates for roughly six months after each halving as the new lower subsidy works its way through the moving average. Treat the half-year windows after November 2012, July 2016, May 2020, and April 2024 with caution — Puell prints in those windows are biased lower independent of any underlying price action. Subsidy-only construction misses fee revenue. Puell traditionally tracks block subsidy only, not subsidy + fees. In periods of extreme fee pressure — late 2023 Ordinals/inscriptions, the May 2024 Runes launch, the December 2017 mempool congestion — the index under-reports total miner revenue by anywhere from 10% to 40%. If you are modelling miner economics directly, cross-reference with fee-inclusive miner-revenue data; if you are using Puell as a cycle-extreme signal, the subsidy-only construction is fine because subsidy dominates multi-month averages. ETF-era flow dynamics blunt the signal. Modern industrial miners hedge production via derivatives and finance operations through equity issuance rather than direct spot sales. The fraction of daily issuance that hits the open market as immediate sell pressure has shrunk over time; Puell continues to measure issuance value, but the link between issuance value and selling pressure has weakened. The November 2021 peak at 1.95 is the clearest example of this decoupling — price set an all-time high while Puell stayed in the mid-cycle band. ## Frequently asked **What is the Bitcoin Puell Multiple?** The Puell Multiple divides daily Bitcoin coin-issuance value (in USD) by its 365-day moving average. It captures whether miners are earning significantly more or less than their recent yearly trend, providing a miner-revenue lens on Bitcoin cycle regimes. The metric was introduced by David Puell on 30 March 2019 in his Medium essay _The Puell Multiple_. **What does a high Puell Multiple mean?** In the original framing, Puell > 4 marked the "blow-off" zone — daily miner revenue running more than 4× its yearly average. The 2013 and 2017 cycles cleared that threshold cleanly (Apr 2013 at 10.49, Nov 2013 at 9.29, Dec 2017 at 6.62). The threshold has not fired since: April 2021 topped at 3.46, November 2021 at 1.95, and March 2024 at 2.44. Like the rest of the cycle-extreme indicators on btc oak, the Puell ceiling has compressed substantially in the modern cycles. **How is the Puell Multiple calculated?** Take the daily block-subsidy issuance in BTC, multiply by the day's USD spot price to get daily issuance value in dollars, then divide by the 365-day simple moving average of that same daily issuance value. The 365-day window smooths the four-year halving steps so the ratio remains comparable across cycles, but does not fully eliminate them — Puell mechanically deflates for roughly six months after each halving as the lower post-halving subsidy works its way through the moving average. **What does a low Puell Multiple mean?** Sustained readings at or below 0.5 mean daily issuance revenue has collapsed to less than half its yearly average — a regime that has only fired four times in Bitcoin history: the aftermath of the 2011 Mt. Gox era, mid-2015, late 2018, and late 2022. Each of those windows preceded a multi-year bull cycle. The deepest cycle troughs on the daily-close record sit at 0.30 (Dec 2018), 0.31 (Jan 2015), 0.36 (Dec 2022), and 0.43 (Mar 2020 Covid). The bottom-side compression has been mild — sub-0.5 has continued to fire every cycle. **Who created the Puell Multiple?** David Puell published _The Puell Multiple_ on Medium on 30 March 2019. The metric was originally introduced as a way to characterise miner-supply-side pressure across cycles, and was subsequently adopted by most major on-chain analysis platforms. Murad Mahmudov (Puell's co-author on the MVRV paper) credits the metric explicitly in their MVRV essay published earlier in 2018; the Puell Multiple in turn is part of a broader miner-revenue framework Puell developed at Adaptive Capital. --- # Net Unrealized Profit / Loss URL: https://btcoak.com/nupl Category: on-chain ## What it is NUPL — Net Unrealized Profit/Loss — measures the share of Bitcoin's market capitalisation that exists as paper profit. The formula is `(Market Cap − Realized Cap) / Market Cap`, identical in shape to `1 − 1/MVRV`. The framing dates to a February 2019 essay by Adamant Capital authors Tuur Demeester, Tamás Blummer and Michiel Lescrauwaet titled _A primer on bitcoin investor sentiment and changes in saving behaviour_, where the metric was called "Relative Unrealized P&L". The five-band naming on this chart is downstream cycle-watcher convention rather than authorial. ## How it is calculated The formula is short: ``` NUPL = (Market Cap − Realized Cap) / Market Cap ``` Market cap is the deterministic product of spot price and circulating supply; realized cap values each coin at the price it last moved on-chain. The Adamant Capital primer that introduced this construction in February 2019 published it as "Relative Unrealized P&L"; the "NUPL" abbreviation and the five-band naming crystallised in cycle-watcher publications afterward. The original authors used three qualitative phases — greed, fear, capitulation/apathy — rather than the modern five-band ladder. Algebraically NUPL is a monotonic transform of MVRV: ``` NUPL = 1 − 1/MVRV ``` The two indicators carry the same information; NUPL's 0-centred 0–1 frame is more legible than MVRV's ratio scale, particularly in capitulation regimes where the values sign-flip clearly. btc oak reconstructs realized cap as a two-bucket weighted sum of short-term-holder and long-term-holder cohort series rather than from a per-UTXO last-spent-price feed; the gap is documented on the methodology page. ## How to read it The band names map onto a sentiment ladder, but the underlying mathematics is just paper-profit share. Negative values say the average coin is underwater; positive values say it is in profit. The thresholds at −0.25, 0, 0.25, 0.5, and 0.75 are cycle-watcher convention rather than authorial — the original Adamant primer named the regimes qualitatively without numeric boundaries. | Reading | Regime | What it has meant | | ----------------- | ------------------ | --------------------------------------------------------------------------------------------------------------------------- | | NUPL < −0.25 | Capitulation | Average coin deeply underwater. Has bracketed only the most stressed cycle floors — early 2015 and late 2014 in our data. | | −0.25 ≤ NUPL < 0 | Hope / Fear | Network at aggregate loss but past the deepest stress. Dec 2018, Mar 2020 Covid, and Nov 2022 post-FTX prints all sat here. | | 0 ≤ NUPL < 0.25 | Optimism / Anxiety | Bitcoin spends more days here than in any other band. Wide mid-cycle range with no strong directional conviction. | | 0.25 ≤ NUPL < 0.5 | Belief / Denial | Late-cycle expansion. The 2024 pre-halving high topped at 0.640; the 2021 Nov peak at 0.659. | | 0.5 ≤ NUPL < 0.75 | (transitional) | A narrower upper-mid band. The 2021 April peak at 0.749 stalled here; this is now the de-facto cycle-top band. | | NUPL ≥ 0.75 | Euphoria / Greed | Three quarters of market cap as paper profit. Last fired Dec 2017 at 0.793. Eight years and counting since the band fired. | ## Historical readings Reading every canonical cycle anchor against the live NUPL series surfaces the cycle decay cleanly. Cycle peaks: +0.83 (April 2013), +0.84 (November 2013), +0.79 (December 2017), +0.75 (April 2021, just touching), +0.66 (November 2021), +0.64 (March 2024). Cycle bottoms: −0.71 (January 2015), −0.43 (December 2018), −0.28 (November 2022). Both top and bottom extremes have been shrinking each cycle. Cycle anchors btc oak computes against: - 2013-04-10 — 2013 Apr peak - 2013-12-04 — 2013 Nov peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2021-04-14 — 2021 Apr peak - 2021-11-10 — 2021 Nov peak - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## Euphoria-band compression The headline observation on this chart is not the current reading — it is the way the bands have compressed cycle by cycle. Recompute the share of trading days each cycle spent in the Euphoria band (NUPL ≥ 0.75) and the pattern is unmistakable. | Cycle | Euphoria-band share | | ---------- | ------------------- | | 2010–2014 | 14% | | 2015–2018 | 0.6% | | 2019–2022 | 0% | | 2023–today | 0% | In the first cycle on record, Bitcoin spent fourteen percent of all trading days above the 0.75 threshold — roughly one in seven days. In 2015–2018 that fell to under one percent. Since 2019 the band has not fired at all. The Capitulation column has compressed similarly. The 2015–2018 cycle spent over a fifth of its days below zero; the current cycle has spent essentially none. The middle bands (Optimism/Anxiety and Belief/Denial) have absorbed both extremes — the 2023-onward cycle has spent over four-fifths of its days inside those two bands alone. Two readings of this pattern are available. The first is structural: as Bitcoin has matured and ETF / institutional flows have damped peak-cycle euphoria and trough-cycle capitulation, the indicator's extreme bands fire less often. The second is mechanical: lost coins growing as a share of issued supply slowly compress realized cap relative to circulating supply, biasing NUPL's denominator. Both are real; both apply. Either way, the canonical cycle-watcher framing — "NUPL above 0.75 marks every cycle top" — no longer describes recent data. ## When it fails The Euphoria band has stopped firing. 2017 was the last time NUPL closed above 0.75 (peak +0.793 on 7 December 2017). Two cycle peaks since (April 2021 at 0.749, November 2021 at 0.659) have stayed below the band. The 2024 high topped at 0.640 — well inside the Belief/Denial range. A "wait for Euphoria" rule would have missed the entire 2021 cycle and is on track to miss this one. NUPL is mathematically MVRV in costume. The relationship `NUPL = 1 − 1/MVRV` means the two indicators carry the same information. Pages and dashboards that display them as independent confirmations of one another are double-counting the same signal. Treat NUPL and MVRV as two views of one lens, not two lenses. Realized cap inherits lost-coin distortion. Chainalysis research has estimated 2.78 to 3.79 million BTC permanently lost; Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have never moved since 2010. Both categories sit in realized cap with near-zero per-coin valuations, biasing the denominator down and NUPL up. The effect is structural and one-directional, and accumulates slowly as a fraction of circulating supply. The bands are convention, not theory. The 0, 0.25, 0.5, 0.75 thresholds are not in the original Adamant Capital primer. They emerged from cycle-watcher publications and were anchored on the 2013 and 2017 peak readings. As those readings decay, the thresholds become harder to defend on first principles. Either Bitcoin is structurally different in the post-ETF era and the bands need recalibration, or the bands were always over-fit to two cycle samples. ## Frequently asked **What is Bitcoin NUPL?** NUPL — Net Unrealized Profit/Loss — measures the share of Bitcoin's market capitalisation that exists as paper profit. The formula is `(Market Cap − Realized Cap) / Market Cap`, identical in shape to `1 − 1/MVRV`. The framing dates to a February 2019 Adamant Capital essay by Tuur Demeester, Tamás Blummer and Michiel Lescrauwaet, where the metric was called "Relative Unrealized P&L"; the five-band naming is downstream cycle-watcher convention. **What does negative NUPL mean?** NUPL below zero means the network's market cap has fallen below its realized cap — the average coin is held at a paper loss. On btc oak's daily-close series this regime has fired at every cycle bottom on the record: −0.71 in January 2015, −0.43 in December 2018, and −0.28 in November 2022. Capitulation depth has shrunk cycle by cycle, mirroring the maturation pattern visible across most on-chain top/bottom indicators. **How is NUPL calculated?** NUPL = (Market Cap − Realized Cap) / Market Cap, where market cap is spot price × circulating supply and realized cap values each coin at the price it last moved on-chain. btc oak reconstructs realized cap from short-term-holder and long-term-holder cohort series; the gap to a per-UTXO ground-truth realized cap is documented on the methodology page. **What does NUPL above 0.75 mean?** Above 0.75 the network is in "Euphoria/Greed" by the conventional band naming — three quarters of market cap exists as paper profit. Historically this regime has fired only at cycle peaks: 0.83 in April 2013, 0.84 in November 2013, and 0.79 in December 2017. It has not fired since: April 2021 topped at 0.75 (just touching the band), November 2021 at 0.66, and March 2024 at 0.64. Whether the band will fire in this cycle is an open question. **Is NUPL the same as MVRV?** NUPL is mathematically `1 − 1/MVRV` — a monotonic transform that maps MVRV onto a 0-centred 0–1 visual frame. The two indicators carry the same information, but the five-band naming on NUPL (Capitulation / Hope-Fear / Optimism-Anxiety / Belief-Denial / Euphoria-Greed) and the symmetry around zero make NUPL more legible at a glance. Both share the same realized-cap reconstruction limitations; if one is correct, the other is. --- # Bitcoin ETF Flows URL: https://btcoak.com/etf-flows Category: flows ## What it is The Bitcoin spot-ETF flows chart plots a daily bar for net USD flow across every tracked US spot Bitcoin ETF, paired with a continuous line for cumulative net inflow on the left scale. Each bar sums the day's authorised-participant creations minus redemptions across the spot complex; the cumulative line carries the running total since launch. The chart starts on 11 January 2024, the SEC-approved first-trading-day for ten of the eleven approved spot products. ## How it is calculated ``` NetFlow_USD(t) = Σᵢ NetFlow_USDᵢ(t) Cumulative(t) = Σ_{k≤t} NetFlow_USD(k) ``` Per-issuer daily flow is published end-of-day by each spot ETF as its net creation-redemption activity in shares, converted to a USD-denominated delta at the day's NAV. The page sums those deltas across every tracked spot product, retains the per-ticker breakdown for tooltip detail, and accumulates the running cumulative inflow on the secondary line. Three conventions matter. First, flow is _net_ — creations minus redemptions, not gross activity. Second, the spot universe excludes ProShares BITO, a CME-futures product launched in October 2021; BITO appears in some upstream lists but is not a spot product, and the per-issuer ranking on this page filters it out. Third, the [10 January 2024 SEC omnibus order](https://www.sec.gov/files/rules/sro/nysearca/2024/34-99306.pdf) approved eleven products, but only ten traded on day one; Hashdex's DEFI was approved on the same day but [did not convert to spot until 27 March 2024](https://www.coindesk.com/business/2024/03/27/us-has-its-eleventh-spot-bitcoin-etf-after-hashdex-fund-conversion). The Grayscale Bitcoin Mini Trust (BTC) added a twelfth spot product on 31 July 2024 as a 10%-of-GBTC spin-off with a 0.15% expense ratio. ## How to read it The trailing five-day net flow resolves on three bands. | Reading | Regime | What it has historically meant | | ------------ | -------------- | ----------------------------------------------------------------------------------------------------------------------------- | | > +$2B / 5d | Strong inflows | Sustained AP creations across the spot complex. Has clustered around extension legs and around quarter-end NAV concentration. | | −$1B to +$2B | Neutral | Creations and redemptions cancelling at the AP layer. Daily prints inside this band carry no fresh contrarian signal. | | < −$1B / 5d | Outflows | Net redemption regime. Pre-2024 a clean directional bear cell; on this dataset, mostly GBTC product-rotation, not bear bet. | ## Historical readings Six anchors trace the cumulative-inflow arc: - 2024-01-11 — Spot ETF launch day (first daily close) - 2024-03-29 — 2024 Q1 close (GBTC unwind dominant) - 2024-09-30 — 2024 Q3 close - 2024-12-31 — Year 1 close - 2025-12-31 — Year 2 close - 2026-04-20 — Most recent close ## Flow is not the same as buying The page's most useful disagreement with the consensus reading is that ETF flow is a structural component of the broader demand picture, not its leading edge. Stated plainly: a net-flow day is the sum of authorised-participant creations and redemptions cleared at the day's NAV fixing, not a count of net new spot demand. Authorised participants create new ETF shares by delivering the underlying — bitcoin, in the spot complex, sourced through OTC desks at the daily benchmark fixing — and redeem existing shares by reversing the trade. A creation absorbed in-kind shifts custody from one wallet to another but does not, on its own, add new market demand: the AP already had the bitcoin and is simply moving it into the trust's book. A creation absorbed in cash, in contrast, requires the issuer to go buy the bitcoin at the fixing window. The two route differently through the lit orderbook, but a net-flow report aggregates them identically. Flow and spot decoupled most cleanly in the early-2026 drawdown. Bitcoin fell roughly forty percent peak-to-trough, and the spot-ETF complex [retained roughly 93% of its assets](https://www.coindesk.com/markets/2026/02/05/bitcoin-etfs-barely-flinch-as-btc-slides-40-bloomberg-s-eric-balchunas-says) across the same window — only about 6.6% of AUM exited via redemption. As Bloomberg's senior ETF analyst Eric Balchunas framed it, ETF holders had behaved as "1–2% hot sauce" allocators rather than tactical traders. Flow was sticky; spot moved anyway. The two series correlate over months and diverge meaningfully at the week. The structural counterpoint, equally important, is that AP custody redirection now routes through Coinbase Custody Trust Company, LLC for the bulk of the spot complex (per the [IBIT prospectus](https://www.ishares.com/us/literature/prospectus/p-ishares-bitcoin-trust-12-31.pdf) and equivalent designations on most Day-1 issuers; FBTC is the significant exception, self-custodying through Fidelity Digital Asset Services). A coin moving from a Coinbase trading wallet to a Coinbase Custody wallet shows up as an [exchange outflow](https://btcoak.com/exchange-flows) with a corresponding ETF-creation print here. The post-2024 cycle's structural story resolves only when the two charts are read together. ## When it fails The GBTC outflow window was product rotation, not directional flow. The first quarter of 2024 produced a record-setting $14.7 billion in GBTC outflows against $13.9B into IBIT, $7.5B into FBTC, $2.2B into ARKB and $1.8B into BITB. Bloomberg ETF analysts Eric Balchunas and James Seyffart [attributed roughly a third of the bleed](https://fortune.com/crypto/2024/01/31/grayscale-bitcoin-etfs-gbtc-billions/) to fee-arbitrage rotation: GBTC carried a 1.5% management fee against sub-0.30% on the new entrants, a $125,000-per-year savings on $10M of bitcoin exposure. The Genesis bankruptcy estate added the rest: court-approved liquidation of roughly $1.3 billion in GBTC shares on 14 February 2024, with total disposals approaching $1.4 billion across the quarter. Anyone reading those Q1 outflow days as a bear bet misread the mechanism. Quarter-end NAV concentration biases the flow signal. Authorised participants create and redeem at the daily NAV fixing — the 4pm New York window — and quarter-end days carry additional rebalancing weight from advisor-driven model portfolios adjusting allocations to a target. The [independent market-structure analysis](https://research.kaiko.com/insights/btc-etfs-impact-on-spot-market-structure) from June 2024 documented benchmark-fixing concentration rising to over 6.7% of all volume from 4.5% in the fourth quarter of 2023. Read quarter-end and month-end flow prints as concentration windows, not fresh demand signals. The flow line is not the buying line. AP creations can be cleared in-kind (the AP delivers bitcoin already on its book in exchange for newly-issued shares) or in cash (the AP delivers USD; the issuer buys bitcoin at the fixing). The flow report aggregates the two. The on-chain custody side at [/exchange-flows](https://btcoak.com/exchange-flows) shows where the redirected coins sit. ## Frequently asked **What is a spot Bitcoin ETF?** An exchange-traded fund whose shares are backed by actual bitcoin held by a regulated custodian, rather than by bitcoin futures contracts. The first ten US spot Bitcoin ETFs began trading on 11 January 2024 after the SEC's January 10 omnibus approval order. **What is the biggest Bitcoin ETF?** BlackRock's iShares Bitcoin Trust (IBIT) is the largest US spot Bitcoin ETF by assets under management, and the top three issuers (IBIT, FBTC, GBTC) account for the bulk of total spot-ETF AUM. ProShares BITO is sometimes counted alongside but holds CME futures rather than spot bitcoin. **Why did GBTC bleed in early 2024?** Three drivers: a 1.50% management fee against sub-0.30% fees on every other Day-1 product, the closing of the discount-collapse trade GBTC's pre-conversion holders had been using, and the Genesis bankruptcy estate's court-approved $1.3 billion liquidation of GBTC shares on 14 February 2024. **Can ETF flows decouple from spot?** Yes — routinely, and at multi-week horizons. The cleanest recent demonstration came across the early-2026 drawdown, when only ~6.6% of ETF assets exited via redemption despite a 40% spot drawdown. Flow stayed shallow while spot cratered. --- # Exchange Reserves & Net-Flow URL: https://btcoak.com/exchange-flows Category: flows ## What it is The Bitcoin exchange-flows chart pairs total reserves held by centralised exchanges with the trailing 30-day net flow into those wallets. The reserves line is the on-chain custody-side complement to the lit-orderbook reads the rest of the site surfaces; the net-flow line tells you which direction those custody balances have been moving. The two together let a reader see the sign and the magnitude of supply rotation between exchange and non-exchange custody at a glance. ## How it is calculated ``` Reserves(t) = Σᵥ Reservesᵥ(t) NetFlow_30d(t) = Reserves(t) − Reserves(t − 30d) ``` Per-venue reserves are the on-chain balance of every wallet cluster identified as belonging to a tracked centralised exchange at the daily close. The 30-day net flow is the simple difference of total reserves at t versus 30 days prior, denominated in BTC. Positive readings mean coins moved onto exchanges over the window; negative readings mean coins moved off. Coverage starts 2024-05-01 — the upstream feed's first publication date for the daily-close per-venue cluster series. The chart and methodology are honest about the gap: the 2022 FTX cliff is documented in prose with [contemporaneous Week 46](https://insights.glassnode.com/the-week-on-chain-week-46-2022/) and [Week 47](https://insights.glassnode.com/the-week-on-chain-week-47-2022/) on-chain reports rather than as a row in the historical table. ## How to read it Five regimes resolve on the trailing 30-day net flow. | Reading | Regime | What it has historically meant | | ---------------------- | -------------- | ------------------------------------------------------------------------------------------------------- | | ≤ −50,000 BTC / 30d | Heavy outflows | Aggressive supply withdrawal. Post-2024, much of this routes to ETF custodians, not self-custody. | | −50,000 to −15,000 BTC | Outflows | Routine drawdown of exchange-held supply. | | −15,000 to +5,000 BTC | Balanced | Net rotation roughly neutral; reserves hold steady within range. | | +5,000 to +20,000 BTC | Inflows | Coins moving onto venues. Historically a soft distribution tell. | | ≥ +20,000 BTC / 30d | Heavy inflows | Sustained loading of exchange wallets. Brackets the deepest distribution legs on the historical record. | ## Historical readings The chart's daily series begins 2024-05-01. Six anchors from across the in-coverage window: - 2024-05-01 — Series start - 2024-09-30 — 2024 Q3 close - 2024-12-31 — Year-1 close (post-launch) - 2025-04-20 — 2025 Q2 anchor - 2025-12-31 — Year-2 close - 2026-04-20 — Most recent close ## The meaning of "outflow" changed in 2024 The chart's most important interpretive note: the classic narrative — that an exchange outflow is a presumptive self-custody move and therefore structurally bullish — stopped being clean in January 2024. The largest single sink for outflows post-launch became Coinbase Custody Trust Company, LLC, which the [IBIT prospectus](https://www.ishares.com/us/literature/prospectus/p-ishares-bitcoin-trust-12-31.pdf) names as the bitcoin custodian for the BlackRock spot ETF; equivalent designations cover most Day-1 issuers, with Fidelity (FBTC) the significant exception (self-custodied through Fidelity Digital Asset Services). A coin moving from a Coinbase trading wallet to a Coinbase Custody wallet shows up here as a centralised-exchange outflow with a corresponding ETF-creation print on [/etf-flows](https://btcoak.com/etf-flows). The same coin is in two charts; the on-chain custody designation changed, the underlying ownership did not necessarily change. The "same coins, different wrapper" framing is the right mental model for post-2024 outflow weeks. The 2022 FTX cliff sits before the chart's coverage window but lives in the methodology because the size matters. Contemporaneous on-chain reports documented roughly 72,900 BTC leaving centralised venues in the week following the FTX collapse, with the cumulative flush across November 2022 approaching 172,700 BTC per month at peak — a real self-custody-driven drawdown of exchange reserves, before the post-ETF era reshaped what an outflow means. ## When it fails Cluster-identification disagreement. Different on-chain providers identify exchange wallet clusters slightly differently — total reserves can disagree across publishers by 100k+ BTC on the same day. The chart's labels reflect one provider's clustering; absolute levels should be read with that caveat. The trend and direction are robust across providers; the precise total is not. The 2024-05-01 start. The pre-coverage record (2018-2024) lives in primary sources, not in the chart's daily series. A user looking at the chart alone will see only the post-ETF era; the methodology page documents the pre-2024 history with the canonical Glassnode citations. The two-line read. Reserves and net flow are not redundant — reserves carry the level, net flow carries the recent direction. A reader looking only at the reserves line will miss the flow inflection that precedes it; a reader looking only at the flow line will miss the structural starting point. Read them together. ## Frequently asked **What does Bitcoin exchange flow mean?** The net direction of bitcoin moving between centralised-exchange wallets and the rest of the on-chain network, summed over a 30-day window. Negative readings (outflows) mean coins left exchange custody on net; positive readings (inflows) mean the reverse. **Are exchange outflows bullish?** Pre-2024, generally yes — outflows tagged self-custody moves and structurally bullish positioning. Post-2024, the framing requires the ETF lens: much of the largest-magnitude outflow weeks now route to ETF custodians rather than self-custody, so the "outflow = removed from market" inference no longer holds cleanly. **What is the Bitcoin exchange reserve?** The total BTC held in wallet clusters identified as belonging to centralised exchanges. The number reads the on-chain custody side of the supply picture — how much sits on venues versus elsewhere. **Why does the chart start in May 2024?** Coverage of the daily-close per-venue cluster series begins there in the upstream feed. The 2022 FTX cliff and earlier pre-2024 history is documented in the methodology and in § 05 of the chart page, with primary on-chain citations. --- # Coinbase Premium Index URL: https://btcoak.com/coinbase-premium Category: flows ## What it is The Coinbase Premium Index plots the percent spread between Bitcoin's USD-quoted price on Coinbase and its USDT-quoted price on the highest-volume offshore venue, daily. Positive prints mean the US-quoted leg is bid above the offshore leg; negative prints mean the reverse. The chart surfaces the raw daily print and a 7-day rolling mean, with the regime classifier keyed off the smoothed series rather than the noisy spot reading. ## How it is calculated ``` Premium(t) = (Coinbase_BTC-USD(t) − Reference_BTC-USDT(t)) / Reference_BTC-USDT(t) ``` The convention is fractional upstream — `0.0010` means a 0.10% premium — and the percent multiplier is applied only at the rendering layer to keep the math clean. The reference leg is the highest-volume USDT-quoted Bitcoin pair at the same UTC daily close as the Coinbase USD print. The metric was introduced by [Ki Young Ju in January 2021](https://x.com/ki_young_ju/status/1352506737488617473) as a US-bid tell during a window when offshore liquidity dominated mark-prints; the [canonical platform definition](https://userguide.cryptoquant.com/data-and-indicators/exchange-flows/coinbase-premium-index) documents the construction in full. The structural logic is straightforward: when US institutional buyers lift Coinbase faster than offshore arbitrage can compress the spread, the index turns positive; when offshore venues lead the print and US flow lags, it turns negative. ## How to read it Five regimes resolve. The seven-day mean sorts into bands at ±0.10% and ±0.25%. Above +0.25%, the US bid is materially ahead of the offshore mark — the regime that historically tagged institutional accumulation legs. Below −0.25%, the offshore mark is leading and Coinbase is dragging — the regime that has bracketed distribution windows. The middle band, ±0.10%, is the post-ETF era's modal reading. | Reading | Regime | What it has historically meant | | ------------------ | --------------- | --------------------------------------------------------------------------------------------------------------------------- | | ≥ +0.25% / 7d mean | Strong premium | US bid materially leading offshore. Pre-2024 framed institutional accumulation legs cleanly; post-2024 carries less weight. | | +0.10% to +0.25% | Mild premium | Modest US lead. Confirmation, not signal. | | ±0.10% | Balanced | The post-ETF modal band. Daily prints inside this range carry no fresh contrarian signal. | | −0.10% to −0.25% | Mild discount | Offshore leg leading. Pre-2024 framed distribution windows; today reads as routine cross-venue noise. | | < −0.25% / 7d mean | Strong discount | Sustained offshore lead. The regime that bracketed the 2018-2020 distribution legs. | ## Historical readings Six anchors trace the regime rotation across the chart's life. Cycle anchors: - 2018-12-15 — 2018 cycle bottom (deep negative regime) - 2020-03-12 — Covid flush - 2021-04-14 — 2021 first-leg high - 2024-03-14 — 2024 pre-halving high - 2025-12-01 — 2025 late-cycle leg - 2026-04-20 — Most recent close ## The post-ETF amplitude compression The shape of the indicator's distribution shifted materially after January 2024. Pre-ETF, daily prints regularly cleared ±0.50% and the 7-day mean held above ±0.25% for weeks at a time during directional regimes. Post-ETF, with [authorised-participant flow routing through Coinbase Custody Trust Company, LLC for most Day-1 issuers](https://www.ishares.com/us/literature/prospectus/p-ishares-bitcoin-trust-12-31.pdf) and continuous arbitrage compressing cross-venue spreads, readings cluster within ±0.10% even on directional days. The [post-ETF spot-market-structure analysis](https://research.kaiko.com/insights/btc-etfs-impact-on-spot-market-structure) framed the structural shift directly: the benchmark fixing window's share of total spot volume rose from 4.5% in Q4 2023 to over 6.7% by mid-2024, and the residual cross-venue spread that the premium reads narrowed in lockstep. The practical consequence: the print pre-2024 carried meaningful predictive weight; post-2024, it reads as confirmation rather than leading-edge. ## When it fails Single-day artefacts. The 19 May 2021 Coinbase outage produced a multi-hour spike where the Coinbase mark stayed flat while the offshore reference moved — the regime classifier filters this by reading off the 7-day mean. The November 2023 Binance settlement window similarly produced a temporary distortion as offshore liquidity rebalanced. The naming collision. "Coinbase Premium" in this chart is the cross-venue spread; "Coinbase One Premium" is a $30/month subscription product. Different topics entirely. The regime carries the signal; the daily print rarely does. A reader treating any single-day spike as a signal will misread the chart most of the time. The 7-day mean is the load-bearing surface. ## Frequently asked **What is the Coinbase Premium Index?** The percent spread between Bitcoin's USD price on Coinbase and its USDT price on the highest-volume offshore venue, daily. Positive means Coinbase is bid above the offshore reference. **Is the Coinbase Premium bullish?** Sometimes — and the era matters. Pre-2024, sustained positive readings tagged US-institutional accumulation legs cleanly. Post-2024, ETF-flow arbitrage compressed the amplitude such that even materially bullish flow days print inside the ±0.10% band. **What is "Coinbase One Premium"?** A separate product entirely — Coinbase's $30/month subscription tier with reduced trading fees. Not what this chart measures. **Why did the indicator compress after 2024?** Because authorised-participant ETF flow routes through Coinbase Custody, and continuous cross-venue arbitrage by basis-trade desks now compresses the residual spread the index reads. The signal pre-2024 lived in a wider band; today's band is materially narrower. --- # AHR999 Index URL: https://btcoak.com/ahr999 Category: on-chain ## What it is AHR999 is a dollar-cost-averaging positioning indicator devised by the Chinese pseudonymous analyst Ah Hui (Weibo handle 999) in mid-2018. The construction multiplies two ratios: spot price divided by the 200-day geometric mean, and spot price divided by a long-run log-fit of price against days since the genesis block. Values below 0.45 historically marked "bottom-fishing" periods where systematic accumulation outperformed; values above 1.2 marked late-cycle overbought windows. The indicator was developed in the wake of the 2018 bear market and explicitly designed for retail dollar-cost-averaging. ## How it is calculated The AHR999 formula is the product of two ratios: ``` AHR999 = (Price / GeoMean_200) × (Price / LogFit(t)) ``` The 200-day geometric mean factor measures how stretched price is from its medium-term baseline. A geometric (rather than arithmetic) mean is used because Bitcoin moves multiplicatively — the geometric mean of $30k, $60k, $30k is $42k, which is a more honest representation of the "typical price level" of that window than the arithmetic mean of $40k. The long-run power-law factor measures how stretched price is from a multi-year secular line. The fit is `price ≈ a × t^b` where _t_ is days since the Bitcoin genesis block (2009-01-03). The power-law line grows sublinearly as the data accumulates — a price that "feels" the same relative to trend in 2024 as in 2014 produces a similar AHR999 contribution from this factor. Multiplying the two captures both timeframes simultaneously. The geometric-mean ratio swings widely with price; the power-law ratio swings less. Their product yields a compact indicator that goes deeply sub-1 only when price is depressed on both timeframes (cycle lows), and goes well above 1 only when price is elevated on both (cycle peaks). The middle range — between 0.45 and 1.2 — is the index's normal habitat. Coverage starts 1 February 2011, after the log-fit has stabilised against early data. Ah Hui's 2018 essay framed AHR999 as a tool for systematic dollar-cost-averaging investors deciding whether to lean into a regular schedule (sub-0.45 territory), maintain it (0.45–1.2), or pause it (> 1.2). The indicator is explicitly not a trade signal — it is a regime tag for portfolio sizing. ## How to read it AHR999 is most informative at extremes. Sub-0.45 readings have only fired during deep cycle bottoms, with troughs clustered tightly between 0.244 (March 2020 Covid) and 0.285 (January 2015). The overbought side has compressed dramatically — the same reasoning that drives MVRV, MVRV-Z, RHODL, Reserve Risk, and Puell ceiling decay applies here. | Reading | Regime | What it has meant | | ------------------- | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------------- | | AHR999 < 0.45 | DCA zone | Bottom-fishing band. Has fired every cycle bottom: 0.285 (Jan 2015), 0.282 (Dec 2018), 0.244 (Mar 2020 — record low), 0.273 (Nov 2022). | | 0.45 ≤ AHR999 ≤ 1.2 | Accumulation range | The index's typical habitat. Bitcoin spends roughly 60% of trading days in this band. | | AHR999 > 1.2 | Overbought | Late-cycle expansion. Modern peaks: 9.25 (Apr 2021), 3.80 (Nov 2021), 1.95 (Mar 2024). Pre-2018 peaks were an order of magnitude higher. | ## Historical readings Reading every canonical cycle anchor against the live series surfaces both the stability of the bottom band and the dramatic compression of the top band. Cycle peaks at 45.98, 83.60, 25.82, 9.25, 3.80, 1.95. Cycle troughs at 0.285, 0.282, 0.244, 0.273 — all four bottoms within ±0.04 of the 0.27 mean. Cycle anchors btc oak computes against: - 2013-04-09 — 2013 April peak - 2013-11-30 — 2013 November peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2020-03-16 — 2020 Covid low (lowest reading ever) - 2021-02-21 — 2021 April peak - 2021-11-09 — 2021 November peak - 2022-11-22 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## The DCA-zone consistency The cleanest way to see what the indicator does well is the bottom column. Cycle troughs cluster between 0.244 and 0.285 across four full bear cycles spanning a decade — a tighter band-fit than any other cycle-extreme indicator on the site. The corresponding top-side compression is dramatic: 45.98 (Apr 2013) → 83.60 (Nov 2013) → 25.82 (Dec 2017) → 9.25 (Apr 2021) → 3.80 (Nov 2021) → 1.95 (Mar 2024). The highest print on the entire history (83.60 in November 2013) is more than 40× the most recent peak. The 200-day geometric mean grows with price. Modern cycle expansions are smaller in magnitude than the early ones (the 2017 cycle ran 100×, 2021 ran 6×, 2024 ran ~3× from the prior bottom), so the geometric-mean factor rarely stretches past 2× even at peak euphoria. The early 2013 cycle ran > 100× in months, producing geometric-mean factors near 8× on its own. The long-run power-law fit grows with the data. Each additional year of price data pulls the secular trendline further from the early-2013 absolute levels, so the power-law-ratio factor compresses naturally. By 2024 the power-law fit had accumulated 13+ years of observations against a denominator that was effectively empty in 2013. The bottom-side threshold holds because both factors compress symmetrically at cycle lows. When price drops to 30% of the 200-day baseline AND 30% of the long-run trend, the product is roughly 0.27 — which is almost exactly where the four canonical cycle troughs printed. The construction of the index is structurally favourable to a stable bottom signal in a way that the top-side reading is not. Lost coins do not affect AHR999 directly — the metric is purely price-based. ## When it fails The 1.2 ceiling has compressed by an order of magnitude. The 2013 and 2017 peaks printed extreme readings (45.98, 83.60, 25.82) that have not been matched since. The 2021 (×2) and 2024 peaks all cleared 1.2 but topped well below the historical extremes. The original Ah Hui "pause DCA above 1.2" rule would have pulled investors out at the start of every modern cycle — a costly false signal. The power-law fit drifts. AHR999's long-run factor depends on a log-fit of price against days since genesis. As the data accumulates, the fit shifts — modern AHR999 is not directly comparable to 2014 readings in absolute terms. The regime thresholds of 0.45 and 1.2 remain the cited anchors, but the expected time spent at each end of the range has narrowed as Bitcoin's volatility has compressed. No cohort or behavioural component. AHR999 is a price-only indicator. It does not see realized cap, holder cohorts, miner revenue, or derivative positioning. Its bottom-side reliability comes from the symmetric compression of both factors at deep lows; its top-side weakness comes from the same mechanism in reverse. Treat AHR999 as a positioning lens, not a market-microstructure tool. ## Frequently asked **What is the Bitcoin AHR999 Index?** AHR999 is a dollar-cost-averaging positioning indicator devised by the Chinese pseudonymous analyst Ah Hui (pseudonym 999) in mid-2018. The construction multiplies two ratios: spot price divided by the 200-day geometric mean, and spot price divided by a long-run log-fit of price against days since the genesis block. Values below 0.45 historically marked "bottom-fishing" periods where systematic accumulation outperformed; values above 1.2 marked late-cycle overbought windows. **What does an AHR999 reading below 0.45 mean?** Sustained sub-0.45 readings have only fired during deep cycle bottoms. The four canonical lows in the series sit between 0.244 (March 2020 Covid — the lowest reading on the entire history) and 0.285 (January 2015). December 2018 hit 0.282 and November 2022 post-FTX touched 0.273. Sub-0.45 days are rare — roughly 12% of all trading days on the record — and almost all of them cluster within a few months of cycle lows. The threshold has continued to fire every cycle without compression. **How is AHR999 calculated?** AHR999 = (price ÷ 200-day geometric mean) × (price ÷ long-run power-law fit). The geometric-mean factor captures how stretched price is from its medium-term baseline; the power-law factor captures how stretched price is from its multi-year secular trend. Multiplying the two yields a compact summary that tracks both timeframes at once. Ah Hui's original formulation appeared in a Chinese-language essay in mid-2018; the metric was popularised on Chinese-language Bitcoin forums and subsequently adopted by major on-chain platforms. **What does a high AHR999 reading mean?** Above 1.2 is the canonical overbought zone. The 2013 (×2) and 2017 cycle peaks all printed extreme readings (45.98, 83.60, 25.82) — driven by the early-cycle log-fit growing slowly while spot exploded. The 2021 peaks compressed to 9.25 (April) and 3.80 (November), and the March 2024 pre-halving high reached only 1.95. Like the rest of the cycle-extreme indicators on btc oak, AHR999's upper extremes have flattened dramatically, but the indicator still cleanly separates "overbought" from "mid-range" from "DCA zone." **Who created the AHR999 index?** The metric was published by the Chinese pseudonymous analyst _Ah Hui_ (阿瓜) — also known as _999_ after his Weibo handle — in mid-2018. The indicator was developed in the wake of the 2018 bear market and explicitly designed for retail dollar-cost-averaging. The English-language Bitcoin community picked up the metric around 2020 as Bitcoin DCA strategies became more widely discussed. AHR999 has no relation to the 999 numerical anchor in the formula — the name is simply the author's pseudonym. --- # Reserve Risk URL: https://btcoak.com/reserve-risk Category: on-chain ## What it is Reserve Risk is an on-chain valuation indicator developed by Hans Hauge in 2019. It divides the current spot price by the accumulated HODL bank — a path-integral of value-days-destroyed weighted by how cheaply coins have been held over their lifetime. Low values mean price is depressed relative to the conviction long-term holders have already shown by sitting through pain; high values mean those same holders have started selling into elevated prices. The indicator is the deepest LTH-conviction lens in on-chain analysis. ## How it is calculated The headline formula is short: ``` Reserve Risk = Price / HODL Bank ``` The HODL bank is the interesting part. It accumulates "conviction" from Coin Days Destroyed — the chain-level primitive that captures how many coin-days a spend extinguishes (a 1 BTC coin held for 100 days then spent destroys 100 BTC-days). Hauge weights each spend by the spot price at the time of the spend versus a reference price drawn from the running historical median, building up an integral of conviction shown by long-term holders sitting through depressed prices. When old coins move at low prices, they add little to the HODL bank. When they sit through long stretches of pain, the bank grows. When they finally move at elevated prices, the bank only barely shrinks — the conviction-bank is a near-monotonic accumulator. Spot price divided by that accumulator is therefore a direct expression of "how much are we paying for this stack of conviction right now." Coin Days Destroyed itself is a pre-2011 primitive — the metric was introduced as "Bitcoin Days Destroyed" on the bitcointalk forum by user ByteCoin in 2011. Reserve Risk is one of several on-chain valuation indicators Hauge built on top of that primitive. Coverage starts 17 August 2010, but the chart's display window starts in 2013 once the HODL bank had accumulated enough mass to stabilise the denominator. ## How to read it Reserve Risk is most informative at the extremes. Sustained readings at or below 0.002 have only fired at cycle bottoms; sustained readings above 0.008 have only fired at the 2013, 2017, and April 2021 cycle peaks. The middle range — between 0.002 and 0.008 — covers most days and carries weak signal on its own. | Reading | Regime | What it has meant | | --------------------- | -------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Value ≤ 0.002 | Strong conviction | Deep-bottom band. Has fired every cycle: 0.00104 (Jan 2015), 0.00159 (Dec 2018), 0.00120 (Mar 2020), 0.00079 (Nov 2022 — record low). | | 0.002 < Value ≤ 0.008 | Neutral | Mid-cycle band. Bitcoin spends more days here than in either extreme. Both the November 2021 peak (0.0055) and the March 2024 high (0.0025) topped inside this band. | | Value > 0.008 | Conviction exhausted | Late-cycle distribution band. Fired at 2013 (×2), 2017, and the April 2021 peak. The Nov 2021 / Mar 2024 cycles never re-entered. | ## Historical readings Reading every canonical cycle anchor against the live series surfaces the same regime-decay pattern that drives the rest of the realized-cap family — but in Reserve Risk the compression is more dramatic than in any other indicator on the site. Cycle peaks at 0.037, 0.039, 0.032, 0.0094, 0.0055, 0.0025. Cycle troughs at 0.00104, 0.00159, 0.00120, 0.00079. Cycle anchors btc oak computes against: - 2013-04-09 — 2013 April peak - 2013-11-29 — 2013 November peak - 2015-01-14 — 2015 cycle low - 2017-12-16 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2020-03-12 — 2020 Covid low - 2021-03-13 — 2021 April peak - 2021-10-20 — 2021 November peak - 2022-11-21 — 2022 cycle low (post-FTX, lowest on record) - 2024-03-13 — 2024 pre-halving high ## The 0.008 ceiling compression The cleanest way to see the regime shift on this chart is the per-cycle Reserve Risk peak, in order: 0.0373 / 0.0387 / 0.0323 / 0.0094 / 0.0055 / 0.0025. Four of six cycle peaks on record cleared 0.008 — and three of those four were pre-2018. The November 2013 peak printed the highest Reserve Risk in the series at 0.0387; April 2013 hit 0.0373; December 2017 hit 0.0323. The April 2021 peak just cleared the threshold at 0.0094, but the November 2021 peak fell back to 0.0055 and the March 2024 pre-halving high topped at 0.0025 — well inside Neutral territory. The 2024 peak is roughly an order of magnitude lower than the 2013 and 2017 peaks. The denominator is doing the work. The HODL bank is a near-monotonic accumulator — every additional bear-market base coin adds conviction, and that conviction never fully evaporates even when older coins eventually move. By 2024 the bank had been compounding for ~13 years of Bitcoin history, against a relatively young denominator-stock at the 2013 peak. Even very large dollar-denominated coin movements at the modern cycle tops barely dent the HODL bank, so the ratio compresses naturally. Lost coins amplify the effect. Every Patoshi-era coin that has not moved since 2010 contributes maximum value-days-destroyed to the HODL bank but never spends, so it adds to the denominator forever. Chainalysis research has estimated 2.78 to 3.79 million BTC permanently lost; Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have never moved since 2010. These coins act as a permanent floor under the HODL bank, structurally lowering every Reserve Risk print. A practical corollary: a "wait for Reserve Risk > 0.008" rule would have missed the November 2021 and March 2024 peaks consecutively. The bottom-side threshold (≤ 0.002) has held up far better — every cycle bottom on the record has fired sub-0.002, with troughs clustered tightly between 0.00079 and 0.00159. ## When it fails The 0.008 ceiling has compressed by an order of magnitude. Three of six cycle peaks cleared it cleanly (2013×2 at ~0.038, 2017 at 0.032). One barely cleared (2021-Apr at 0.0094). Two did not (2021-Nov at 0.0055, 2024-Mar at 0.0025). The threshold was anchored on the 2010s blow-off-top shape and is now well outside the modern operating range. A "wait for 0.008" rule would have missed the entire November 2021 cycle and would currently be missing the 2024 cycle as well. The HODL bank is invisible to off-chain custody. Coins held inside ETF wrappers, lending platforms, or exchange custody appear as on-chain HODL only when the underlying coin sits unmoved at a custodian address. ETF creation and redemption can churn the same underlying position multiple times per week without any change in the "real" conviction of the end holder. The 2024 reading is biased downward by post-Jan-2024 spot-ETF flows that age-reset coins via custodial reshuffling. 2020 Covid was a regime exception in both directions. The Mar 2020 trough at 0.00120 was clean (Strong-conviction band). But spot price recovered fast enough that there was no clean "peak" window in the 2020 cycle — the run from $5,000 to $69,000 happened in 21 months without a clear pause. Reserve Risk printed 0.0094 on 13 March 2021, then 0.0055 on 20 October 2021, with a not-very-elevated mid-cycle in between. The signal was directionally correct but the cycle's shape was unusual. ## Frequently asked **What is Bitcoin Reserve Risk?** Reserve Risk is an on-chain valuation indicator developed by Hans Hauge. It divides the current spot price by the accumulated HODL bank — a path-integral of value-days-destroyed weighted by how cheaply coins have been held over their lifetime. Low values mean price is depressed relative to the conviction long-term holders have already shown by sitting through pain; high values mean those same holders have started selling into elevated prices. **What does a low Reserve Risk reading mean?** Sustained readings at or below 0.002 have bracketed every cycle bottom on the daily-close record: 0.00104 in January 2015, 0.00159 in December 2018, 0.00120 in March 2020 (Covid), and 0.00079 in November 2022 (post-FTX) — the lowest reading on the entire series. Low Reserve Risk means long-term holders are demonstrating conviction by not selling at depressed prices; historically every cycle bottom has fired in this regime, with bottoms compressing slightly tighter each cycle. **How is Reserve Risk calculated?** Reserve Risk = Price ÷ HODL Bank. The HODL bank is computed by integrating Coin Days Destroyed weighted by the price at which coins are spent versus held. Each day a coin sits unspent at a price below realised-price-history, it adds "conviction" to the bank. Each spend at elevated prices subtracts conviction. The result accumulates years of patient long-term holding into a single denominator the spot price is then divided by. Hauge's 2019 paper documents the full mechanics. **What does a high Reserve Risk reading mean?** Readings above 0.008 historically marked late-cycle distribution windows where old coins were moving aggressively into euphoria. The threshold cleared cleanly at the 2013 (×2), 2017, and Apr 2021 peaks — but the November 2021 peak at 0.0055 fell short, and the March 2024 pre-halving high at 0.0025 fell short by a factor of three. The high-side ceiling has compressed dramatically alongside the rest of the realized-cap family. Treat 0.008 as a 2010s-era reference rather than a modern signal. **Who created Reserve Risk?** Hans Hauge published Reserve Risk in 2019, building on Bitcoin's Coin Days Destroyed primitive (introduced as "Bitcoin Days Destroyed" on bitcointalk in 2011). Hauge's methodology paper is hosted on his Medium archive. The metric has since been adopted across most major on-chain analysis platforms; btc oak recomputes it nightly from independent on-chain data so the series remains free and open. --- # RHODL Ratio URL: https://btcoak.com/rhodl Category: on-chain ## What it is RHODL — the Realized HODL Ratio — divides the realized cap of 1-week-old coins by the realized cap of 1-to-2-year-old coins, then multiplies by a supply-adjustment factor that accounts for the growth in circulating supply over time. Philip Swift introduced the metric in February 2020. It is built from the HODL Waves age-band framework and surfaces when speculative young-coin activity dominates the realized-cap distribution — a structural late-cycle pattern. The series is one of the longest on the site, covering daily resolution from 17 August 2010. ## How it is calculated The mechanic is a ratio of two HODL-wave realized-cap slices, scaled by total supply: ``` RHODL = (RealizedCap_1w / RealizedCap_1y2y) × supply_factor ``` The numerator is the dollar value of all coins that last moved in the past seven days, valued at the price they last moved at. The denominator is the same calculation for coins last moved between one and two years ago. The supply factor rescales the ratio against total circulating supply at time of calculation — without it, the same realised-value distribution would yield mechanically larger RHODL prints in 2024 than in 2014 just because supply grew. The intuition: when speculative new buyers churn coins through fresh on-chain hands, the 1-week bucket fills with fresh cost basis at high prices, while the 1-to-2-year bucket reflects what holders bought during the previous bear. The ratio rises sharply in distribution; it falls in accumulation when new activity is muted and the 1–2y band slowly absorbs more of the network's realized cap. The series is built on the same HODL-Waves age-band feed used elsewhere on btc oak. The first observation is 17 August 2010 — among the earliest in any on-chain dataset, because HODL Waves is one of the few metrics computable from the very first chain epoch. ## How to read it RHODL is most informative at extremes. Sustained sub-1,000 readings have only fired at cycle bottoms (102 in February 2015, 192 in January 2019, 198 in December 2022). Sustained readings above 50,000 have only fired at the 2013 and 2017 blow-off tops. The middle range — between roughly 1,000 and 10,000 — is the mid-cycle baseline that covers most days. | Reading | Regime | What it has meant | | ----------------------- | -------------- | ------------------------------------------------------------------------------------------------------------------------- | | RHODL < 1,000 | Accumulation | Old coins dominate realized cap. Has bracketed every cycle bottom: 102 (Feb 2015), 192 (Jan 2019), 198 (Dec 2022). | | 1,000 ≤ RHODL < 10,000 | Mid-cycle | The bulk of trading days. Bitcoin spends more time here than in any other band. | | 10,000 ≤ RHODL < 50,000 | Elevated | Late-cycle distribution. Both 2021 peaks lived here (15.3k Apr, 14.7k Nov); the 2024 pre-halving high topped at 7.8k. | | RHODL ≥ 50,000 | Cycle-top zone | Swift's original cycle-top reference. Last fired December 2017 at 105.7k. The 2013 November peak printed an outlier 212k. | ## Historical readings Reading every canonical cycle anchor against the live series surfaces the same regime-decay pattern that drives the rest of the realized-cap family. Cycle peaks at 15.8k, 212k, 106k, 15.3k, 14.7k, 7.8k. Cycle troughs at 102, 192, 198 — three modern cycle bottoms within a factor of two of each other, suggesting the bottom-side threshold has held up better than the top-side ceiling. Cycle anchors btc oak computes against: - 2013-04-10 — 2013 April peak - 2013-11-29 — 2013 November peak - 2015-02-10 — 2015 cycle low - 2017-12-13 — 2017 cycle top - 2019-01-27 — 2018–19 cycle low - 2020-03-31 — 2020 Covid low - 2021-02-23 — 2021 April peak - 2021-10-25 — 2021 November peak - 2022-12-27 — 2022 cycle low (post-FTX) - 2024-03-13 — 2024 pre-halving high ## Cycle-top decay The cleanest way to see the regime shift on this chart is the per-cycle RHODL peak, in order: 15.8k (Apr 2013) → 212k (Nov 2013) → 106k (Dec 2017) → 15.3k (Apr 2021) → 14.7k (Nov 2021) → 7.8k (Mar 2024). Two of six cycle peaks on record cleared 50,000 — both in the pre-2018 era. The April 2013 peak topped at 15.8k (still in elevated territory but not at the 50k ceiling). The November 2013 peak hit 212k, the highest RHODL print on record. December 2017 hit 105.7k. Then nothing: the April 2021 peak topped at 15.3k, November 2021 at 14.7k, and March 2024 at 7.8k. The trend is monotonic since 2017. The denominator shifted. RHODL is sensitive to the maturity of the 1–2y cohort — coins acquired during the prior bear-market accumulation phase. As Bitcoin matured, that cohort grew larger and more diversified each cycle. By 2021, the 1–2y bucket held a much greater share of realized cap than in 2017, so the same level of young-coin churn produced a smaller RHODL print. By 2024, the addition of spot-ETF flows distorted the picture further: ETF creation and redemption can age-reset coins via custodial reshuffling, biasing the 1-week bucket up without implying real distribution. Lost coins compound the effect. Every Patoshi-era coin that has not moved since 2010 ages permanently into the "over 10y" bucket and never returns to the 1–2y denominator. Chainalysis research has estimated 2.78 to 3.79 million BTC permanently lost; Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have never moved since 2010. The 1y2y denominator is mechanically smaller than "all coins ever active in a 1–2y window since 2010" would imply, but it is also more meaningful — a slowly-decaying baseline of recently-aged conviction holders. ## When it fails The 50k ceiling has stopped firing. Two of six cycle peaks cleared it (2013-Nov at 212k, 2017-Dec at 106k). Four did not (2013-Apr at 15.8k, 2021-Apr at 15.3k, 2021-Nov at 14.7k, 2024-Mar at 7.8k). The threshold was anchored on the blow-off shape of the early cycles, and is now well outside the modern operating range. A "wait for 50k" rule would have missed two cycles consecutively. Custodial reshuffling distorts the young-coin numerator. RHODL treats every on-chain transfer as a coin-age reset. When an exchange or ETF custodian consolidates UTXOs internally, those coins move into the 1-week bucket without any "real" distribution by an investor. The effect has grown since the January 2024 launch of US spot Bitcoin ETFs, which churn substantial daily on-chain volume at the custodial layer. The 1y2y bucket can be unevenly populated. A bear market that lasted 11 months (rather than 13–24) leaves the 1y2y bucket light at the start of the next cycle, mechanically inflating RHODL prints during the early bull phase. The 2020 Covid flush bottomed at an anomalously high 1,118 partly for this reason — the 1y2y bucket was thin because the 2018-19 bear had been short. Spot price recovered faster than the HODL-wave structure could age into the deep-bottom regime. ## Frequently asked **What is the Bitcoin RHODL ratio?** RHODL — Realized HODL — divides the realized cap of 1-week-old coins by the realized cap of 1-to-2-year-old coins, then multiplies by a supply-adjustment factor that accounts for the growth in circulating supply over time. The metric was introduced by Philip Swift in February 2020. It is built from the HODL Waves age-band framework and surfaces when speculative young-coin activity dominates the realized-cap distribution — a structural late-cycle pattern. **What does a high RHODL ratio mean?** A high RHODL means recently-moved coins make up a much larger share of realized cap than long-held coins. In the 2013 and 2017 blow-off cycles, RHODL cleared 50,000 — the canonical Swift top-zone reference. The 2021 cycle topped at 15,270 (April) and 14,695 (November); the March 2024 high reached only 7,755. Like the 3.7 ceiling on raw MVRV, the 50k Swift threshold has not fired in eight years; the new working "extreme" range looks more like 10,000 to 30,000 on a modern cycle. **How is the RHODL ratio calculated?** The numerator is the realized cap of the 1-week HODL Waves bucket — coins that last moved in the past seven days, valued at the price they last moved. The denominator is the same realized-cap calculation for the 1-year-to-2-year bucket. The ratio is then multiplied by total circulating supply at time of calculation (a market-age scaling factor) so cross-era comparisons remain consistent. The result is a unitless indicator that historically spans from below 200 at deep bear-market lows to over 100,000 at blow-off tops. **What does a low RHODL ratio mean?** Sustained sub-1,000 readings have bracketed every cycle bottom on the record: 102 in February 2015, 192 in January 2019, 198 in December 2022 post-FTX. The Mar 2020 Covid flush bottomed near 1,118 — anomalously high, because spot recovered before the HODL waves had time to age into the deep-bottom band. Sub-1,000 prints mean old coins overwhelmingly dominate realized value: classic cold-storage behaviour, typical of bear-market basing. **Who created the RHODL ratio?** Philip Swift published RHODL in a February 2020 essay, building on the HODL Waves age-band framework that Unchained Capital had introduced in 2018. Swift's public dashboards popularised the metric, and it has since been adopted across most major on-chain analysis platforms. --- # STH / LTH SOPR URL: https://btcoak.com/sopr Category: on-chain ## What it is SOPR — the Spent-Output Profit Ratio — values every coin spent on a given day at the price it last moved on-chain (its "cost basis") versus the price it spent at, then averages across all spends for the day. Above 1.0 the network is realising profit on average; below 1.0 it is realising loss; at 1.0 every coin moved at exact breakeven. Renato Shirakashi introduced the metric in his April 2019 Medium essay _Introducing SOPR_. btc oak ships two cohort splits — short-term-holder (coins last moved < 155 days ago) and long-term-holder (≥ 155 days) — because the aggregate network SOPR hides as much as it reveals. ## How it is calculated The SOPR formula is short: ``` SOPR(t) = mean over spends(t) of ( price_spent / price_acquired ) ``` For every spent output on day _t_, divide the price at which it was spent by the price at which it was last moved on-chain. Average over every spend on the day. The result is a unitless ratio. The cohort split applies the same formula to two separate sub-populations. STH SOPR restricts the average to outputs younger than 155 days at the time of spend; LTH SOPR restricts to outputs aged 155 days or more. The 155-day boundary follows Schultze-Kraft and Heeg's 2020 cohort paper, which identified that threshold as the empirical point at which coin spending probability stabilises — short-term reactive behaviour gives way to long-term structural holding. Coverage starts in mid-2010 alongside meaningful daily spend volumes; reads from the first ~3 years of the chain should be treated with caution due to thin sample size. ## How to read it Watch STH SOPR for trend-initiation signals. Crossing 1.0 from below after extended sub-unit weakness has historically marked cycle restarts. Crossing 1.0 from above signals early-cycle stress or the start of distribution. Watch LTH SOPR for the long-range regime: sustained readings above 3 indicate old coins entering the market (late-cycle distribution); sustained sub-1 readings during prolonged drawdowns indicate capitulation. Direction and the relationship between the two lines matter more than absolute levels. | Reading | Regime | What it has meant | | --------------------- | ----------------- | ----------------------------------------------------------------------------------------------------- | | STH < 1 · LTH < 1 | Mutual loss | Both cohorts realising loss. Deep-capitulation regime — fired at every cycle bottom on record. | | STH < 1 · LTH ≥ 1 | Late-cycle stress | STH cohort underwater while LTH still selling profitably. Often appears mid-correction or pre-bottom. | | STH ≥ 1 · LTH < 1 | Bear-bounce | Reactive STH profit-taking inside a wider bear regime. Rare; bear-market relief rallies. | | STH ≥ 1 · 1 ≤ LTH < 3 | Balanced profit | Both cohorts profitable but neither distributing aggressively. The bulk of mid-cycle days. | | STH ≥ 1 · LTH ≥ 3 | Distribution | LTH selling into a profitable retail bid. Fired at every cycle peak in the realized-cap record. | ## Historical readings Reading every canonical cycle anchor against both SOPR series surfaces both the gentle decay of STH peaks and the dramatic decay of LTH peaks. STH cycle peaks have spanned roughly 1.11 to 1.40 — a 30% range over six tops. LTH cycle peaks have collapsed by an order of magnitude across the same six. Cycle bottoms have held a tighter band: STH lows of 0.85 / 0.89 / 0.84 / 0.94, LTH lows of 0.35 / 0.27 / 0.57 / 0.35. Cycle anchors btc oak computes against: - 2013-04-09 — 2013 April peak - 2013-11-22 — 2013 November peak - 2015-01-21 — 2015 cycle low - 2017-12-07 — 2017 cycle top - 2018-11-20 — 2018 cycle low - 2020-03-12 — 2020 Covid low - 2021-01-07 — 2021 January peak (STH) - 2021-04-14 — 2021 April peak (LTH) - 2021-11-12 — 2021 November peak - 2022-11-09 — 2022 cycle low (post-FTX) - 2024-03-13 — 2024 pre-halving high ## STH vs LTH cycle decay The cleanest way to see the regime shift on this chart is the per-cycle peak for both lines, in order. STH SOPR has been remarkably stable across six cycle peaks: 1.29 (Apr 2013) → 1.40 (Nov 2013) → 1.26 (2017) → 1.19 (Apr 2021) → 1.11 (Nov 2021) → 1.21 (Mar 2024). Even the lowest STH peak (Nov 2021 at 1.11) sat clearly above breakeven; the most recent print at 1.21 is squarely in the long-run norm. LTH SOPR has compressed by an order of magnitude over the same window: 35 → 298 → 56 → 13 → 5.1 → 5.6. The 2013 peaks printed extreme readings because the 155-day-old cohort in 2013 included Patoshi-era coins last moved at fractions of a dollar — every spend at $1,000+ generated a five-figure SOPR. By 2017 the cohort had aged forward but still included substantial sub-$100 cost-basis coins. By 2021–2024, most LTH spends have cost bases in the thousands, not cents. Lost coins compound the bias: Chainalysis has estimated 2.78 to 3.79 million BTC permanently lost, and Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have not moved since 2010 — they sit in the cost-basis baseline at near-zero but almost never spend. A practical corollary: the "LTH SOPR > 3 = top" rule has held up better than most realized-cap-family ceilings, because it is a cohort-relative threshold rather than an absolute one. The 2021 (×2) and 2024 cycles all cleared LTH-SOPR > 3 at their peaks — only the absolute heights have compressed. ## When it fails Custodial reshuffles register as "spends." SOPR treats every output spend equally, regardless of whether it represents a real change of ownership or an internal custodian movement. ETF creation and redemption, exchange cold-to-hot wallet transfers, and OTC settlement all count as spends. The structural noise floor has grown since US spot Bitcoin ETFs launched in January 2024. The April 2018 Coinbase Multibit migration is a known historical artefact. On 9–10 April 2018, Coinbase migrated coins from legacy Multibit wallets to its modern custody infrastructure. The migration produced a multi-day SOPR spike — coins last moved at fractions of a cent in 2010–2012 suddenly appeared as spends at 2018 prices. Treat that window as custody-distorted rather than market-driven. The 155-day boundary is empirical, not mechanical. Coins on either side of the line behave on a continuum, not a step. Schultze-Kraft and Heeg identified 155 days as the point where coin spending probability stabilises and the choice has become standard, but a 120-day or 180-day cutoff would tell a similar story with marginally different numerics. Daily noise is also high — most analysts apply a 7-day moving average for visualisation; the underlying daily series is preserved here for transparency. ## Frequently asked **What is Bitcoin SOPR?** SOPR — the Spent-Output Profit Ratio — values every coin spent on a given day at the price it last moved on-chain versus the price it spent at, then averages across all spends. SOPR > 1 means the network is realising profit on average; SOPR < 1 means realising loss; SOPR = 1 is exact breakeven. Renato Shirakashi introduced the metric in his April 2019 essay _Introducing SOPR_. **What is the difference between STH-SOPR and LTH-SOPR?** Short-term-holder SOPR restricts the average to coins that last moved less than 155 days ago — the reactive cohort whose behaviour drives most cycle momentum. Long-term-holder SOPR restricts to coins last moved 155+ days ago — old hands whose realisations matter most at cycle extremes. The 155-day boundary follows Schultze-Kraft and Heeg's 2020 cohort paper, which identifies that threshold empirically as the point where coin spending probability stabilises. **How do you read SOPR signals?** STH-SOPR crossing 1.0 from below after extended sub-unit weakness has historically marked cycle restarts. STH-SOPR crossing 1.0 from above signals early-cycle stress or the start of distribution. LTH-SOPR sustained well above 1 (especially > 3) is structural distribution. LTH-SOPR sustained below 1 is bear-market capitulation. Direction and the relationship between the two lines matter more than absolute levels. **What does SOPR below 1 mean?** SOPR below 1 means coins are being spent at a loss on average. STH-SOPR below 1 fires routinely during bull-market drawdowns and bear markets — the 2015 trough hit 0.846, the 2020 Covid flush hit 0.840, the 2022 post-FTX low at 0.940 was the shallowest cycle bottom on the STH series. LTH-SOPR below 1 is rarer and signals capitulation by old-hand holders — typical only at cycle bottoms (0.35 in Jan 2015, 0.27 in Jan 2019, 0.35 in Nov 2022). **Who created the SOPR metric?** Renato Shirakashi published _Introducing SOPR_ on Medium on 24 April 2019. The original definition uses a single network-wide SOPR; the STH/LTH split was a downstream extension built on the 155-day cohort framework Schultze-Kraft and Heeg published in 2020. btc oak serves both lines together because the cohort split is where the regime distinctions become legible — the network-wide aggregate hides as much as it reveals. --- # STH / LTH Realized Price URL: https://btcoak.com/realized-price Category: on-chain ## What it is Realized price is the average cost basis of all coins in circulation, valued at the price each coin last moved on-chain rather than at today's market price. It is the per-coin form of realized capitalisation, the framework Antoine Le Calvez and Nic Carter introduced at the Baltic Honeybadger 2018 conference on 23 September 2018. Splitting the per-coin average by holder age gives two distinct levels: short-term-holder realized price (coins moved within the last 155 days) and long-term-holder realized price (coins older than 155 days). On btc oak the two lines are plotted on a shared log price axis with spot Bitcoin for direct comparison. ## How it is calculated Both cost-basis lines are slices of realized capitalisation. Realized cap values each unspent transaction output (UTXO) at the price it last moved on-chain rather than at today's market price; per coin, that yields realized price. The cohort split traces to Rafael Schultze-Kraft and Kilian Heeg's 2020 paper _Quantifying Short-Term and Long-Term Holder Bitcoin Supply_, which identified 155 days as the inflection point at which a coin's probability of being spent flattens out. Their methodology applies a logistic transition (midpoint 155 days, 10-day width) rather than a hard cutoff so the cohort boundary breathes smoothly through time. The two series compute as: ``` STH realized price = Σ (last_spent_price · value) over UTXOs aged < 155d ÷ STH supply LTH realized price = Σ (last_spent_price · value) over UTXOs aged ≥ 155d ÷ LTH supply ``` Both series are sourced upstream as already-aggregated cohort lines; per-UTXO last-spent prices are not reconstructed from the raw blockchain. The methodology page records the exact upstream feed, the smoothing convention, and known divergences. Coverage starts 17 August 2010. ## How to read it Three regimes carry distinct historical meaning. Above both cost bases is the bull-regime default — every cohort in aggregate profit, with the spread between spot and LTH a coarse measure of cycle maturity. Between the cost bases means recent buyers are underwater while long-term holders remain in profit, the transitional shape typical of late corrections and early bear markets. Below both means even long-term holders are at a paper loss — a regime that has bracketed the 2015, 2018, and 2022 cycle bottoms on the daily-close record. The 2020 Covid flush is the only modern bottom that did not reach this regime, by a margin of five percent. | Reading | Regime | What it has meant | | ---------------- | -------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------- | | spot < LTH < STH | Below LTH cost basis | 2015 trough printed −43% below LTH; 2018 trough −27%; 2022 post-FTX trough −25%. Statistically rare regime occupied at three of four cycle bottoms. | | LTH < spot < STH | Between cost bases | Recent buyers underwater; long-term holders still in profit. Typical of late-bear or late-correction phases. | | spot > STH > LTH | Above all cost bases | Both cohorts in aggregate profit. The bull-regime default — every prior cycle peak fired with spot well above STH. | ## Historical readings Reading every canonical cycle anchor against today's cohort series surfaces the regime pattern. The 2015 cycle low printed deepest at roughly −43% below LTH; the 2018 trough at −27%; the 2022 post-FTX trough at −25%. The 2020 Covid sell-off is the sole exception — the closest daily close to an LTH breach (17 March 2020) bottomed +5.34% above LTH cost basis and then recovered. Three of four bottoms confirm the regime; the fourth illustrates that even a cleanly defined signal misses occasionally. Cycle anchors btc oak computes against: - 2013-04-10 — 2013 April peak - 2013-12-04 — 2013 November peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2020-03-17 — 2020 Covid flush (closest LTH approach) - 2021-04-14 — 2021 April peak - 2021-11-10 — 2021 November peak - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## Loss-residency across cycles The single most informative summary statistic on this chart is not either line by itself — it is the share of all-time daily closes where spot has closed below the LTH cost basis. The figure recomputes nightly against the current series. The figure is the bear-market anatomy of the asset in one number: long-term holders are, on average, in unrealised profit on the great majority of all trading days. The residency in below-LTH territory is concentrated almost entirely in three windows — mid-2014 to mid-2015, late 2018, and mid-to-late 2022 — with only a single brief modern occupation outside those bear cycles. The STH/LTH spread carries the cycle-phase read alongside the loss-residency lens. Spread strongly positive (STH well above LTH) is the bull regime: new capital is paying premiums over the older cohort's cost basis, and every prior cycle peak has fired with the spread maximally extended. Spread compressing toward zero is the late-bull or early-bear shape: new capital has slowed and long-term holders' average cost is catching up to recent buyers'. Spread negative (STH below LTH) is the bear-market signature: long-term holders' average cost basis exceeds recent buyers' because LTHs include coins picked up near the prior cycle top. Every Bitcoin bear since 2015 has ended only after the spread closed and crossed back, with the cross dates landing in early 2015, late 2018, and mid-2022. ## When it fails The 2020 Covid bottom did not breach LTH. The closest daily close on btc oak's series was 17 March 2020, with spot at $5,032.50 against an LTH realized price of $4,777.30 — a margin of +5.34%. A mechanical "below LTH = bottom" rule would have missed the entry. The signal is high-conviction when it fires; the absence of the signal does not preclude a cycle-trough print. Lost coins inflate LTH supply at near-zero cost basis. Chainalysis research has estimated 2.78 to 3.79 million BTC permanently lost; Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have never moved since 2010. Both categories sit in the LTH bucket with near-zero realized prices, biasing LTH cost basis down and LTH supply up; in particularly stressed bottoms this can make the LTH line look more "held" than the active float justifies. Cohort migration noise. The 155-day boundary is statistical, not economic. Coins purchased near a cycle peak age past the boundary into LTH continuously, lifting LTH realized price even when no holder has bought. In strong bull runs this mechanical drift accounts for a meaningful fraction of LTH cost-basis growth — the indicator is not capturing fresh LTH demand, it is capturing the calendar. Single-entity reset events compound this: when a custodian consolidates UTXOs, every affected coin is repriced to the consolidation-day value even though no economic ownership has changed (the Coinbase cold-storage migration in December 2018, deep-dived by Felipe at Paradigma Capital, is one well-documented example). ## Frequently asked **What is Bitcoin realized price?** Realized price is the average cost basis of all coins in circulation, valued at the price each coin last moved on-chain rather than at today's market price. It is the per-coin form of realized capitalisation, the framework Antoine Le Calvez and Nic Carter introduced at the Baltic Honeybadger 2018 conference on 23 September 2018. Splitting the per-coin average by holder age gives two distinct levels: short-term-holder realized price (coins moved within the last 155 days) and long-term-holder realized price (coins older than 155 days). **What is the difference between STH and LTH realized price?** Short-term-holder realized price is the average cost basis of coins last moved within 155 days; long-term-holder realized price is the average for coins held longer. The 155-day boundary comes from Rafael Schultze-Kraft and Kilian Heeg's 2020 paper _Quantifying Short-Term and Long-Term Holder Bitcoin Supply_, where they identified 155 days as the empirical inflection point at which a coin's probability of being spent flattens out. **What does it mean when Bitcoin trades below LTH realized price?** Spot below LTH realized price means the average long-term holder is at a paper loss — a statistically rare condition that has bracketed the deepest cycle bottoms on record. On btc oak's daily-close series, three of four cycle troughs have closed below the LTH line: 2015 (spot roughly 43% below), 2018 (27% below), and 2022 post-FTX (25% below). The 2020 Covid flush is the exception — it bottomed 5.34% above LTH. Below-LTH is a strong, but not perfect, generational-bottom signal. **Why do STH and LTH cost bases sometimes invert?** STH cost basis falls below LTH cost basis when long-term holders' average price exceeds recent buyers' — the signature of a late bear market. Coins purchased near a cycle top age past the 155-day boundary into the LTH cohort and lift its average; meanwhile, recent buyers picked up coins at depressed prices. Every Bitcoin bear since 2015 has ended only after this spread closed and inverted, with STH crossing back above LTH in early 2015, late 2018, and mid-2022. **How accurate is realized price?** Realized price is a definitional aggregate, not a forecast — the formula is unambiguous given a complete UTXO history. Two structural caveats: lost coins (Chainalysis estimates 2.78 to 3.79 million BTC permanently lost; Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have never moved since 2010) inflate the LTH cohort with near-zero cost basis, and cohort migration means coins ageing past 155 days dilute LTH realized price toward STH cost basis even when no holder has bought. --- # Golden-Ratio Multiplier URL: https://btcoak.com/golden-ratio Category: cycles ## What it is The Golden-Ratio Multiplier is a cycle-regime band chart developed by Philip Swift in April 2019, published in long form on Medium that June, as part of the Bitcoin Investor Tool. It plots the 350-day simple moving average of Bitcoin price and seven Fibonacci multiples of it — 1.6 (the golden ratio), 2, 3, 5, 8, 13, and 21 — as parallel resistance bands. Historical cycle tops have landed inside specific bands, and a clean diminishing pattern emerges across cycles. ## How it is calculated The inputs are Bitcoin's daily USD closes. For every day _t_: ``` 350DMA(t) = mean(price[t−349..t]) band_k(t) = m_k × 350DMA(t) for m_k ∈ {1.6, 2, 3, 5, 8, 13, 21} ``` Band assignment at any date is the largest multiple `m_k` the live price is at or above. The chart's underlying data file ships only the 350DMA series; the band lines are reconstructed client-side from the seven multipliers, so a reader can swap the multiplier set without re-fetching. The 350-day window is Swift-selected, not derived — he picked it because it produced the cleanest band alignment for the 2013 and 2017 cycle tops at the time of publication. The multiplier set 1.6/2/3/5/8/13/21 is a Fibonacci sequence with the golden ratio (≈ 1.618) rounded to 1.6 at the lower end. It is a heuristic, not a model. ## How to read it Read by band, not by dollar value. A move from one band to the next is a regime shift in the long-run cycle. Crossing above ×1.6 after a sustained accumulation stretch has historically marked the start of a bull phase; crossing below ×1.6 after extended time in the upper bands has marked the transition into bear-market capitulation. The upper bands — ×5, ×8, ×13, ×21 — are escalation zones, not mandatory targets; later cycles have broken back down well before reaching the band the prior cycle did. | Reading | Regime | What it has meant | | ---------- | -------------------- | ------------------------------------------------------------------------------- | | Below ×1.6 | Accumulation | 2014–15, 2018–19, and 2022–23 each spent extended time here. | | ×1.6 – ×2 | Early-cycle recovery | Above the golden-ratio band. Historically the start of a sustained bull phase. | | ×2 – ×3 | Mid-cycle expansion | The 2024 pre-halving high lived here. No prior cycle has topped in this band. | | ×3 – ×5 | Late-cycle window | The 2021 April and November peaks topped inside this band — both at roughly ×3. | | ×5 – ×8 | Cycle-peak window | The 2017 December top finished here at roughly ×5–6. Swift mapped 2017 to ×5. | | ×8 – ×13 | Speculative extreme | The 2013 November top finished here. Swift mapped 2013 to ×13. | | > ×13 | Mt. Gox-era extreme | Reached only by the 2011 cycle peak (×21 in Swift's mapping). | ## Historical readings Computing each cycle peak's price ÷ 350DMA inline shows the diminishing-returns pattern Swift first noted in 2019, extended through the two cycles he had not yet seen. The 2011 and 2013 April anchors fall before the 350DMA's full warm-up window and are flagged as such — never faked in the table. Cycle peaks the page tracks against: - 2011-06-08 — 2011 cycle top (Mt. Gox-era) - 2013-04-10 — 2013 April peak - 2013-11-29 — 2013 November peak - 2017-12-17 — 2017 cycle top - 2021-04-14 — 2021 April peak - 2021-11-10 — 2021 November peak - 2024-03-14 — 2024 pre-halving high ## Peak-multiplier decay By peak multiplier, cycle by cycle: 2011 hit ×21; 2013 November hit ×13; 2017 December hit ×5–8; both 2021 peaks topped near ×3; the 2024 pre-halving high cleared only ×2. The trend is mechanical — each cycle has reached a smaller multiple of its 350-day MA than the cycle before. Swift's June 2019 mapping framed the same observation forward-looking and the two cycles since publication have continued the pattern. By that pattern, the next cycle peak would land below ×3, not above ×5. The ×1.6 line is the chart's most durable single threshold. It marks both the upper edge of accumulation territory (when spot is below it) and the lower edge of the bull-cycle envelope (when spot is above it). Crossings have been sticky: 2015 upcross preceded the 2017 peak; 2018 downcross marked the cycle bottom; 2019 upcross preceded the 2021 peak; 2022 downcross marked the post-FTX bottom. ## When it fails The 350-day window and Fibonacci multipliers are post-hoc fits, not derivations. Swift's June 2019 piece (https://positivecrypto.medium.com/the-golden-ratio-multiplier-c2567401e12a) presents the multipliers as "the mathematically organic nature of Bitcoin adoption", mapping past cycle tops to specific bands. The mapping fits well enough to be useful — but it is curve-fitting on three cycles. Tim Stolte of Amdax framed the broader category critique in a 2022 post on Bitcoin power-law fits: "there's no logic or wisdom there, just pure guesswork and picking whatever looks nice" (https://medium.com/amdax-asset-management/bitcoins-power-law-corridor-debunked-1b40783657bf). The upper bands have gone silent since 2017. Swift's mapping was historical — ×21 / ×13 / ×8 / ×5 — but no cycle since publication has reached even ×5. The ×8, ×13, and ×21 bands are now scenery, not operational thresholds. Matt Crosby summarised the broader 2024 situation: long-cycle indicators "remained untested" this cycle (https://bitcoinmagazine.com/markets/why-bitcoin-price-top-indicators-failed). The four-year cycle scaffolding may be ending. If Lyn Alden and Matt Hougan are right that the structural drivers are shifting from issuance halvings to ETF flows and global liquidity (https://experts.bitwiseinvestments.com/cio-memos/the-four-year-cycle-is-dead-welcome-to-the-ten-year-grind), the band-decay pattern may continue all the way down — to the point where the multiplier framework collapses into the ×1.6 line as the regime-marker and the higher bands are decorative. ## Frequently asked **What is the Golden-Ratio Multiplier?** A cycle-regime band chart developed by Philip Swift in April 2019 (published in long form on Medium that June) as part of the Bitcoin Investor Tool. It plots the 350-day simple moving average and seven multiples of it — 1.6, 2, 3, 5, 8, 13, and 21 — as parallel resistance bands. Historical cycle tops have landed inside specific bands, and a clean diminishing pattern emerges across cycles. **Who created the Golden-Ratio Multiplier?** Philip Swift (@PositiveCrypto) first published the indicator in April 2019 and laid out the framework in long form in _The Golden Ratio Multiplier: Unlocking the mathematically organic nature of Bitcoin adoption_ on Medium on 17 June 2019. The piece introduced the multiplier framework alongside the structurally similar Pi Cycle Top indicator. Swift mapped each prior cycle peak to a multiplier band: ×21 for 2011, ×13 for 2013, ×8 for the 2014 echo, ×5 for the 2018 high. **Why the 350-day window?** 350 days is just under one calendar year — long enough to absorb intra-year volatility but not so long that it dilutes cycle structure. Swift's original tool used 350 because it produced the cleanest band alignment with the cycle tops available at the time (2013 and 2017). Subsequent cycles have largely held the alignment. Shorter windows (200d) overreact to mid-cycle dips; longer windows (500d) lag the signal. **Which Bitcoin cycle peak hit the highest multiplier band?** The 2011 cycle, by a wide margin. By Swift's mapping the 2011 top reached ×21 of the 350DMA, the 2013 November high ×13, the 2017 December high ×5–8, and the 2021 cycle peak windows topped out around ×3. The 2024 pre-halving high cleared only ×2. Each successive cycle has reached a lower maximum multiplier. **Is the ×1.6 band a buy signal?** Empirically the ×1.6 line has marked the regime boundary between accumulation and bull-trend territory more often than not — a sustained close above ×1.6 has historically preceded major rallies, and a sustained close back below has marked the transition from bull to bear. It is not a buy signal in isolation; the sample size is four full cycles, and the ×1.6 line itself is a heuristic Swift selected because the golden ratio (≈ 1.618) provided the cleanest fit. --- # Long/Short Ratio URL: https://btcoak.com/long-short Category: derivatives ## What it is The Bitcoin long/short ratio counts how many trader accounts on the highest-volume Bitcoin perpetual-futures venue currently hold a net-long position versus a net-short position, then divides one by the other. A ratio above 1 means more accounts are net long than net short; below 1, the reverse. It is a head count, not a money count — a thousand small longs against a dozen large shorts can still print long-heavy. ## How it is calculated The canonical construction takes the top 20% of accounts on the venue ranked by margin balance, classifies each as net long or net short by current open positions, and computes: - `Long Account % = Accounts of top traders with net long positions / Total accounts of top traders with open positions` - `L/S = Long Account % / Short Account %` The verbatim spec is published in the [venue's API documentation](https://developers.binance.com/docs/derivatives/usds-margined-futures/market-data/rest-api/Top-Long-Short-Account-Ratio). The same venue also publishes a position-weighted version that often disagrees with the account-weighted one. The chart shows one venue. Aggregating long/short across venues to a single "global" reading is not possible upstream — each venue defines its own ratio surface (top-X-percent by margin on one, all open-position holders on another, top traders by recent volume on a third), snapshots on its own cadence, and exposes only the precomputed result. There is no shared account-classification spec to sum across. ## How to read it Three regimes resolve cleanly. Below `0.7`, short-account share has pushed past 60% and the book is shorts-crowded — historically a setup for a squeeze, since the under-capitalised side covers fast on any rally. Above `1.3`, long-account share has pushed past 56% and the book is longs-crowded — historically a setup for a leverage flush, since stops cascade when the trend stalls. The middle band is descriptive: it tells you who is positioned, not where price is going next. | Reading | Regime | What it has historically meant | | ----------------- | -------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | L/S < 0.70 | Shorts crowded | Contrarian setup. The smaller, often less-capitalised short side covers quickly on any spot rally, producing a sharper bounce than the order book alone would predict. | | 0.70 ≤ L/S ≤ 1.30 | Balanced | Both sides within the historical norm. The chart carries no fresh contrarian signal — treat it as background. | | L/S > 1.30 | Longs crowded | Contrarian setup. Stops cluster on the long side and any down-move cascades through them faster than spot supply would imply. | ## Historical readings Six anchor windows since the start of our daily series make the indicator's character visible. Spot prices are nightly closes from our pipeline; ratios are pulled from the same daily snapshot that powers the chart. Cycle anchors: - 2024-03-14 — 2024 pre-halving high - 2024-08-05 — 2024 yen-carry unwind (global cross-asset flush) - 2024-11-21 — 2024 Trump-rally cycle leg - 2025-01-31 — January 2025 ATH - 2025-12-01 — 2025 late-cycle leg - 2026-04-20 — Most recent close ## The single-venue confession The chart shows one venue. That has been the indicator's honest framing since the first Bitcoin perpetual swap launched in May 2016: every venue computes its own ratio on its own user subset, and there is no upstream feed that reconciles them. So the ratio reflects whichever venue currently carries the largest perp-volume share, and the cycle-to-cycle survivor shifts. The original perpetual's home venue dominated through 2018; a different one rose through 2020–2022 and then collapsed; a handful of offshore books split the post-2022 book with one venue consistently on top. A long/short series spliced across that history is not a like-for-like comparison even within the same dashboard. Two specific consequences. The survivorship effect means the indicator "works" differently in each cycle; a 2021-vintage 1.3 reading captured a different slice of trader behaviour than a 2025 reading. The account-vs-position fork means the same venue's reading can flip sign depending on which surface a dashboard pulls. We plot the account-weighted version because retail positioning — the contrarian story — sits in account counts; the position-weighted divergence is a caveat, not a competing chart. ## When FTX exposed the metric The clearest failure window is November 2022. On 2 November 2022, reporting on the Alameda balance sheet exposed an FTT-collateralised hole at Bitcoin's second-largest perp venue; by 11 November FTX had filed Chapter 11. Long/short feeds on FTX simply stopped — the venue went from live to dark over a weekend. The dominant survivor's reading then became, by default, the new "global" reading without any change in the underlying methodology, even though the user base feeding it had been swelled by a venue migration that was nothing like organic positioning. The post-FTX liquidity gap framework — the [Alameda Gap](https://research.kaiko.com/insights/looking-back-on-ftxs-impact) — documented that in the week after the collapse, "global crypto liquidity had halved," and that depth had still recovered to only half its pre-FTX level eleven months in. FTX's peak derivatives market share was around 15%; that share didn't move smoothly — it disappeared into one weekend, and the ratios on the remaining venues drifted to extremes that reflected the migration as much as any real positioning shift. ## When it fails Account counts ignore size. A hundred small retail longs stacked against five large institutional shorts will print long-crowded even though the dollar imbalance runs the other way. Always cross-read against the same venue's position-weighted ratio if a directional decision rests on the call. The two regularly disagree on direction. The dominant venue rotates. November 2022 is the canonical example, but the broader pattern is that perp market share migrates between successive dominant venues on regulatory, geographic, and product-design timescales. A 2021 long/short reading and a 2026 long/short reading are sampled from different user bases, under different rule sets, in different volume regimes. The line on the chart hides that. It is the most gameable metric on the site. Funding rates reflect a real cash payment and open interest reflects a settled book, so both are hard to manipulate without leaving evidence. Long/short ratios are a derived classification of accounts; even small population changes — new sub-accounts opening, dormant accounts closing, regional bans culling part of the user base — can shift the readout without any trader changing their view. Treat extremes as suggestive and absolute levels with suspicion. ## Frequently asked **What is the Bitcoin long/short ratio?** The number of trader accounts on the highest-volume Bitcoin perpetual-futures venue currently net-long, divided by the number net-short. Above 1 means more accounts are long; below 1 means more are short. Account-weighted, not position-weighted. **How is it calculated?** `Long Account % / Short Account %`, computed on the top 20% of accounts by margin balance, sampled at each daily close. Same venue also publishes a position-weighted version that often disagrees. **Is it a contrarian indicator?** Best at the extremes. Below 0.7 has historically preceded short squeezes; above 1.3 has historically preceded long-side leverage flushes. The middle band — 0.7 to 1.3 — is descriptive, not predictive. **Why only one exchange?** No upstream feed aggregates long/short across venues to a single comparable number. Each venue defines its own subset and snapshots on its own cadence. Summing them is meaningless. The honest construction picks the highest-volume venue and discloses the caveat. --- # Funding Rate URL: https://btcoak.com/funding-rate Category: derivatives ## What it is The Bitcoin funding rate plots the volume-weighted 8-hour funding payment across the major perpetual-futures venues, as a daily close. Y-axis is symmetric around zero; rust fill above the baseline marks settlements where longs paid shorts, slate fill below marks the reverse. Funding is the price of carrying a leveraged Bitcoin position — not a sentiment vote. ## How it is calculated The canonical perpetual-funding formula is identical in shape across the major venues, with only minor variations in clamp magnitude and settlement cadence: ``` F = P + clamp(I − P, −0.05%, +0.05%) ``` Where `F` is the funding rate per 8-hour interval, `P` is the time-weighted average premium of the perpetual price over the index price across the window, and `I` is a fixed interest baseline of `+0.01%` per 8 hours on USD-collateralised contracts. The clamp constrains the (interest minus premium) differential to plus or minus five basis points; an outer cap then bounds `F` itself. The most useful decomposition for reading the chart is to remember that _in flat markets, with the perpetual tracking spot tightly, the premium is zero and funding settles to the interest baseline alone_ — a structural `+0.01%` positive bias the chart cannot escape. The [2025 Q3 derivatives report](https://www.bitmex.com/blog/2025q3-derivatives-report) framed it cleanly: "The perpetual swap formula has a built-in interest component, forcing rates to cluster around 0.01% (positive bias)." Two consequences of the structural bias are worth flagging. A "neutral" reading is not zero — it is centred on the interest baseline. Anything within a few basis points of `+0.01%` per 8h is the formula clearing its throat. Sustained negative readings are mechanically harder to produce than sustained positive ones because the perpetual has to trade persistently below spot for the (interest minus premium) term to drag the rate negative through the clamp. Sustained negative funding stretches concentrate around cycle-bottom panic windows for that reason. Three settlements per day at 00:00, 08:00 and 16:00 UTC; we aggregate the three into one daily close so the series is comparable to the rest of the site. For intraday funding spikes during a flush, the upstream venue's live feed is the right surface; this chart is the nightly summary. ## How to read it Three regimes resolve. Above `+0.03%` per 8h, longs are paying meaningful yield to stay in — an annualised carry north of `+33%`. That is expensive, and any reversal can snowball as leveraged longs unwind to avoid the cost. Below `−0.02%`, shorts are paying longs — the regime that historically brackets squeeze setups, when bearish positioning crowds into a shrinking pool. The middle band, around the `+0.01%` structural baseline, is the formula at rest. | Reading | Regime | What it has historically meant | | ---------------- | ---------------- | ----------------------------------------------------------------------------------------------------------------------------- | | > +0.03% / 8h | Overheated longs | Crowded long-side positioning. Historically precedes leverage flushes when momentum stalls. Cross-read against open interest. | | −0.02% to +0.03% | Neutral | The structural-bias band, including the `+0.01%`-per-8h interest baseline. The chart carries no fresh contrarian signal. | | < −0.02% / 8h | Shorts pay | Shorts crediting longs — the rare regime where carry has flipped. Brackets cycle-bottom panic windows historically. | ## Historical readings Six anchors since the start of our daily series make the regime rotation visible. Cycle anchors: - 2024-03-14 — 2024 pre-halving high (extreme positive funding) - 2024-08-05 — 2024 yen-carry unwind - 2024-11-21 — 2024 Trump-rally cycle leg - 2025-01-31 — January 2025 ATH window - 2025-12-01 — 2025 late-cycle leg - 2026-04-20 — Most recent close ## Funding is carry, not sentiment The framing matters more than any single reading. Arthur Hayes, who co-founded the venue that launched the original Bitcoin perpetual swap, framed the mechanism plainly in his [2021 essay on perp carry](https://cryptohayes.medium.com/all-aboard-4d50435190d6): the perpetual swap exchanges a funding rate — an interest income — between longs and shorts every 8 hours. The unit of analysis is the payment, not the sentiment. A single basis-trade desk that shorts the perp against spot can move funding without expressing any directional view; the chart records the carry, not the conviction. Hayes returned to the framing in his [2025 retrospective](https://cryptohayes.medium.com/adapt-or-die-6d14649bdd90) with a worked example: "if the perp traded at an average 1% premium to spot over the last eight hours, if you held a position at the funding period timestamp, longs pay 1% to shorts." The perp's premium drives the payment, full stop. Reading funding as "people are bullish" or "people are bearish" mistakes the formula for a poll. Read it instead as: at this rate, this is what it costs to be on the crowded side of the book. ## The post-FTX rate-floor regime The shape of the funding distribution shifted after November 2022. Pre-FTX, extreme positive funding readings — multiple settlements per cycle north of `+0.10%` per 8h — were a regular feature; the basis-trade ceiling, where cash-and-carry desks short the perp into spot until the premium compresses, was thin and often disappeared in panic windows. Post-FTX, that ceiling has thickened. The [2025 Q3 derivatives report](https://www.bitmex.com/blog/2025q3-derivatives-report) framed the structure: "the large pool of undeployed capital acts as a ceiling for funding rates – preventing it from staying high for long." Extreme positive funding still happens, but it persists for hours, not days. ## When it fails The regulated leg is invisible. Quarterly cash-settled Bitcoin futures — the regulated venue's leg of the book — do not have a funding mechanism; they settle on contract expiry instead. During cycles where regulated capital drives marginal price action (the post-spot-ETF era is the clearest example), perp funding can decouple from spot behaviour because the leg moving the tape is not in this chart. Venue dispersion during dislocations. The [post-FTX market-structure analysis](https://research.kaiko.com/insights/looking-back-on-ftxs-impact) documented that "a week after the collapse, we noticed that global crypto liquidity had halved" — a window in which one venue's funding feed simply stopped publishing while panic shorts crowded into surviving books. A volume-weighted aggregate across that window averaged a stale cell with live feeds; the daily print misread the dispersion as neutrality. The structural floor anchors readings positive. The formula's `+0.01%`-per-8h interest baseline biases the chart toward positive readings even in flat markets. A casual reading of "funding is slightly positive, longs must be euphoric" misreads the formula clearing its throat. Always interpret the print against the `+0.01%` floor, not against zero. ## Frequently asked **What is the Bitcoin funding rate?** The periodic payment exchanged between long and short holders of a perpetual swap, designed to keep the perpetual price anchored to spot. Every major venue settles a funding payment every 8 hours; positive means longs pay shorts. **How is it calculated?** `F = P + clamp(I − P, −0.05%, +0.05%)`, where `P` is the time-weighted premium of the perpetual over the index across the funding window and `I` is a `+0.01%` per 8h interest baseline. An outer cap bounds `F` itself. **What happens if funding is negative?** Short-position holders pay long-position holders at the next settlement. Sustained negative funding requires the perpetual to trade persistently below spot, which historically coincides with cycle-bottom panic windows. **Is negative funding bullish?** Mildly — as a contrarian setup, not a directional forecast. The position pays you to hold and the short side bears the cost of waiting. In practice it has bracketed cycle bottoms more often than tops. --- # Open Interest URL: https://btcoak.com/open-interest Category: derivatives ## What it is Bitcoin open interest plots the total USD notional of outstanding Bitcoin futures contracts each day, stacked by venue. Each band in the area is one venue's contribution; the thin line at the top is the aggregate. Y-axis on a logarithmic ruler so the pre-2022 era at roughly $3B stays legible against the modern scale. Open interest is a stock measured at instant _t_, not a flow over an interval — different from volume, which most explainers conflate. ## How it is calculated Two inputs flow in: each venue's reported USD open interest at the daily close, and the deterministic Bitcoin circulating supply on that day. Three computations: - `OI_total(t) = Σ_v OI_v(t)` across covered venues, with the residual outside the top five rolled into a single `Other` bucket. - `market_cap(t) = price(t) × circulating_supply(t)`, where circulating supply is the closed-form value at block-height _t_ — no pricing-feed dependency. - `OI/mcap(t) = OI_total(t) / market_cap(t)` — the leverage-vs-spot proxy that drives the regime classifier. OI is reported in three equivalent forms across venues — contract count, USD notional, coin notional — and a coin-margined contract's USD value moves with spot even when no new contracts are opened. We pull USD notional everywhere we can, and accept that on coin-margined exposure the dollar reading carries a small mechanical price-sensitivity. Regime thresholds are `OI/mcap > 2.5%` for "Leveraged" and `< 1.5%` for "Deleveraged" — descriptive cuts on the post-2020 distribution, not bright lines. The distinction worth holding onto: open interest is a stock measured at instant _t_; volume is a flow measured over interval [t−1, t]. They answer different questions and behave differently in a flush. A leveraged blow-off can show falling OI (positions closing) and exploding volume (closures producing trades) on the same day. The two are not interchangeable. ## How to read it The instructive read pairs OI with price simultaneously, in a four-quadrant frame. The cleanest published version sits in a [July 2025 leveraged-positioning research note](https://insights.glassnode.com/leverage-position-openings-and-closures/), which describes the four cells: | Reading | Regime | What it has historically meant | | -------------- | -------------- | ------------------------------------------------------------------------------------------------------------------------------------------ | | ↑ price · ↑ OI | Long openings | New leveraged longs entering on top of the trend. Healthy in a young rally; fragile when the regime is already "Leveraged" on the OI/mcap. | | ↓ price · ↑ OI | Short openings | New shorts opening into a falling tape. Squeeze setup — under-water shorts cover at a loss if spot bounces. | | ↑ price · ↓ OI | Short closures | Shorts unwinding into strength — the mechanical leg of a squeeze. Often follows "short openings" within days. | | ↓ price · ↓ OI | Long closures | Leveraged longs closing into weakness. Healthy mid-cycle; outright capitulation when paired with deeply negative funding. | ## Historical readings Seven anchors since the start of our daily series sketch the cycle. Aggregate OI has compounded from the COVID flush at roughly $3B to the post-ETF era above $50B; the path between them is the cycle. Cycle anchors: - 2020-03-12 — COVID liquidity flush - 2021-04-14 — 2021 Apr cycle peak (Coinbase listing day) - 2021-11-09 — 2021 Nov cycle top - 2022-11-08 — Pre-FTX collapse (one day before withdrawals halted) - 2022-11-30 — Post-FTX redistribution (three weeks after Chapter 11) - 2024-03-14 — 2024 pre-halving high - 2025-01-31 — January 2025 ATH window ## The FTX redistribution exhibit The cleanest case study sits across two rows. On 2022-11-08, the day before FTX paused withdrawals, aggregate OI sat at roughly $13.1B. By 2022-11-30 — three weeks after the [Chapter 11 filing](https://restructuring.ra.kroll.com/FTX/) — total OI had fallen to $10.6B. The instinctive read is "market deleveraged twenty percent." That is the wrong frame. The missing notional did not unwind — the venue carrying it disappeared. Positions on the FTX perpetual book were socialised into the bankruptcy estate; aggregator totals stepped down because one of the cells went to zero, not because every other venue's holders closed out. The post-FTX [Alameda Gap analysis](https://research.kaiko.com/insights/looking-back-on-ftxs-impact) put FTX's peak derivatives market share at around 15%; the share didn't drift, it disappeared in one weekend. Surviving venues absorbed the flow over the following months — the aggregate climbed back through the 2023 banking-crisis chop and into the spot-ETF approvals — but the November 2022 step is the cleanest reminder that aggregate open interest is not a neutral leverage gauge across venue regimes. ## The regulated-leg flip The composition of the stack changed permanently after the [January 2024 spot-ETF approvals](https://www.coindesk.com/markets/2024/02/14/bitcoin-futures-open-interest-tops-21b-highest-since-november-2021). Through the 2021 cycle the largest band in the stack was an offshore perpetual venue; the regulated futures leg sat third or fourth. By [November 2024](https://www.coindesk.com/markets/2024/11/21/futures-open-interest-on-cme-surpasses-215k-bitcoin-for-the-first-time-as-btc-eyes-100k) the regulated leg had crossed 218,000 BTC ($21.3B) of OI — first time at that scale — and the cycle's top of the stack runs there consistently. That flip matters for how the chart reads. A 2021-era $24B total OI was retail-led, dominated by unregulated leverage with weekly liquidation cascades. The post-ETF era is structurally different — a third of it sits inside cash-settled, compliant exposure that does not produce the same weekend wicks. The OI/mcap ratio reads broadly the same numbers cycle to cycle, but the volatility a given reading produces has been falling as the regulated band has thickened. ## When it fails Aggregator coverage drift. The list of venues we report is not static. New venues become material; older ones fade. The post-2020 stack adds coverage faster than any cyclical signal moves, so the absolute total carries a slow upward bias purely from coverage expansion. The OI/mcap ratio absorbs most of that bias because both numerator and denominator grow with the book; the per-venue stack does not. Venue discontinuity. November 2022 is the clean example — aggregate OI fell because a venue disappeared, not because positions closed. A read of November 8 against November 30 records that as a 19% deleverage; the [Alameda Gap research](https://research.kaiko.com/insights/looking-back-on-ftxs-impact) shows that figure conflates two unrelated things. Whenever the stack has a step-change in venue composition, treat the total as discontinuous. Coin-margined contracts move with spot. A non-trivial slice of perpetual OI is coin-margined — quoted in BTC notional, displayed in USD after multiplying by current spot. On those contracts, USD-denominated OI moves mechanically with price even if no new contracts opened or closed. In a sharp drawdown, coin-margined OI will appear to fall faster than the actual position count drops, exaggerating the "deleveraging" visible on the chart. ## Frequently asked **What is Bitcoin open interest?** The total number of outstanding Bitcoin derivative contracts that have not yet been closed or settled, multiplied by their USD notional value. Measures held positions, not turnover. **How does OI differ from volume?** OI is a stock at instant _t_; volume is a flow over an interval. A market with $30B OI and $50B daily volume is differently leveraged than one with $30B OI and $200B daily volume — different turnover ratios, different positioning. **Is rising OI bullish?** Not on its own. Rising OI says new contracts opened; the direction of the new positioning depends on the simultaneous price move. The four-quadrant frame holds: rising-OI-plus-rising-price marks long-side accumulation; rising-OI-plus-falling-price marks short-side positioning. **How is the OI/market-cap ratio interpreted?** A leverage proxy — what fraction of free-float economic value is being expressed via leverage rather than spot. Above 2.5% prints "Leveraged"; below 1.5% prints "Deleveraged". Descriptive cuts, not absolute laws. --- # Liquidation Heatmap URL: https://btcoak.com/liquidations Category: derivatives ## What it is The Bitcoin liquidation heatmap renders a two-dimensional density grid. Horizontal axis is calendar date across the rolling year, vertical axis is BTC price on a logarithmic scale, and cell colour intensity reflects the modelled USD size of expected forced-unwind flow if price traded at that cell during that day. Bright clusters above current spot are upside leverage; bright clusters below are downside leverage. The grid is a topographical map of leverage, not a forecast. ## How it is calculated No exchange publishes individual position liquidation prices. Every liquidation heatmap on the public web is therefore a _modelled_ estimate — built from the open-interest snapshot, leverage-tier assumptions (typically 100×, 50×, and 25× longs and shorts), and the standard maintenance-margin formulas each venue uses. The clearest published primary description sits in a [cell-construction methodology note](https://academy.hyblockcapital.com/tools/liquidation-levels-1): the heatmap "calculates the liquidation levels based on market data and different leverage amounts. The calculated levels are then added to a price bucket on the chart." The grid is resampled from a fine-grained upstream feed of 283 price levels across 360 daily columns into 20 log-spaced price buckets × daily columns so the shipped JSON stays under 100 KB. The compression keeps zoom and pan responsive without losing the cell-level shape. The ratio in the reading row is computed as: ``` cluster_up = Σ_{p ∈ (spot, spot × 1.05]} grid(today, p) cluster_down = Σ_{p ∈ [spot × 0.95, spot)} grid(today, p) ratio = cluster_up / cluster_down ``` Above 2.0, the upside cluster dominates — a small spot rally would cascade through the stacked stops into forced buy-backs (a short-squeeze setup). Below 0.5, the downside cluster dominates and a small spot drop can cascade into forced sells (a long-liquidation setup). The single most important caveat is published in the same methodology note: "the Liquidation Heatmap predicts where liquidation levels are opening but not closing. Thus, the actual number of liquidations will be lower." Realised liquidations are bounded above by the modelled clusters; the chart never under-counts what would unwind, but it routinely over-counts, because positions close before they liquidate. Treat dense cells as upper bounds on potential flow, not as forecasts of realised flow. ## How to read it Three regimes resolve from the ±5% cluster ratio. The numerical thresholds have weakened over time as the rolling window's composition shifts; treat them as anchors on the recent distribution, not absolute laws. | Reading | Regime | What it has historically meant | | ----------------- | ---------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- | | ratio > 2.0 | Short-squeeze setup | Upside cluster more than twice the downside. A small rally has historically cascaded through stops into forced buy-backs. | | 0.5 ≤ ratio ≤ 2.0 | Balanced | Both sides within the historical mid-band. The chart carries no clean directional read; rely on funding or open interest for a tie-break. | | ratio < 0.5 | Long-liquidation setup | Downside cluster more than twice the upside. A small drop has historically cascaded through stacked long stops into forced sells. | ## Historical readings Seven monthly snapshots through the rolling year sketch the regime rotation. Each row reports the ±5% cluster ratio on the snapshot day, the spot price the cluster centred on, and the regime that ratio resolved into. Cycle anchors: - 2025-04-30 — Spring 2025 cycle leg (start of rolling window) - 2025-06-15 — Mid-2025 grind - 2025-08-15 — Late summer 2025 chop - 2025-10-15 — Autumn 2025 leg - 2025-12-15 — 2025 late-cycle window - 2026-02-15 — February 2026 distribution - 2026-04-15 — Most recent snapshot ## When the magnet held — and when it didn't The folkloric framing is that price "seeks" dense liquidation clusters because market makers know where forced flow will hit and probe those zones to trigger it. The cleanest recent positive case is spring 2024, when sub-$50k upside clusters magnetised wicks during the early-cycle consolidation before BTC extended higher. The cleanest counter-example is November 2022. In the days before [FTX's Chapter 11 filing](https://restructuring.ra.kroll.com/FTX/), the dominant cluster on every published heatmap sat on FTX-perp positions below the prevailing spot. Price did not seek that magnet on the way down — the magnet evaporated. Positions were socialised into the bankruptcy estate; the cluster zeroed out by mid-November not because price obeyed it but because the venue that carried it disappeared. The [post-FTX market-structure analysis](https://research.kaiko.com/insights/looking-back-on-ftxs-impact) put FTX's peak derivatives market share at roughly 15%; that share didn't drift, it disappeared in one weekend, and the heatmap of November 7 was a topographical map of a country about to be erased from the atlas. The honest framing: the heatmap is a topographical map of leverage as it currently sits, conditional on the venues currently in the model staying live. When that conditional breaks — venue failure, regulatory shutdown, mass migration — the magnet narrative breaks with it. ## When it fails Modelled, not measured. The canonical methodology note flags it explicitly: the heatmap "predicts where liquidation levels are opening but not closing. Thus, the actual number of liquidations will be lower." Realised liquidations depend on real-time balances, leverage choices that change continuously, and venue-specific margin rules. Dense cells are upper bounds on potential flow, not forecasts of realised flow. Venue discontinuity. November 2022 is the canonical failure mode — the dominant cluster vanished overnight when the venue carrying it failed. Any heatmap is conditional on the model's venue list staying live; a regulatory shutdown, a delisting, or a venue-failure event can erase a cluster without any trader closing a position. Tactical horizon. The heatmap is a day-to-week tool, not a cycle-level lens. It says nothing about whether Bitcoin is rich or cheap on a multi-year horizon — that work belongs to the realised-price, RHODL, and NUPL charts. ## Frequently asked **How do you read a Bitcoin liquidation heatmap?** As a topographical map of leverage, not as a price prediction. Bright cells are clusters of expected forced-unwind flow if price reaches that bucket; magnet behaviour is folkloric, not mechanical. **What does each cell represent?** The model's estimate of the USD of leveraged positions that would be force-closed if BTC traded at that price during that day. Built from open-interest snapshots, leverage-tier assumptions, and standard margin formulas. **Is the heatmap accurate?** It is a model with explicit limits. Realised liquidations are bounded above by modelled clusters but routinely lower because positions close before they liquidate. Dense cells are upper bounds on potential flow. **What is a magnet zone?** A price bucket with a dense cluster of expected liquidation flow that price has tended to probe before reversing. The narrative held in many windows and broke in November 2022 when the dominant cluster's venue ceased to exist. --- # Bitcoin Dominance URL: https://btcoak.com/dominance Category: macro ## What it is Bitcoin Dominance is the ratio of Bitcoin's market capitalisation to the sum of every tracked cryptoasset's market cap, expressed as a percentage. The metric became a standard reference in the mid-2010s as the cryptoasset universe grew large enough for the ratio to be meaningful, and remains the most-quoted single number describing the BTC-versus-everything-else regime. ## How it is calculated ``` Dominance = BTC market cap / Total cryptoasset market cap ``` Bitcoin market cap is the deterministic product of spot price × circulating supply. The denominator is the sum of every tracked cryptoasset's market cap, including stablecoins (USDT, USDC, DAI…) and wrapped tokens (WBTC, stETH…). Different publishers track different baskets and apply slightly different inclusion rules, so dominance numbers across sites disagree by a percentage point or two on the same day. The 60-day trend cell tracks the change in dominance over a rolling 60 trading days, expressed in percentage points. A +2 pts move over 60 days reads as "Dominance expanding"; a −2 pts move reads as "Dominance contracting"; between the two is "Stable". Sustained directional motion over two months is more meaningful than the instantaneous level, especially in a market where the absolute number is structurally biased by stablecoin supply growth. ## How to read it Dominance is a relative-performance gauge between Bitcoin and the rest of the cryptoasset complex — it does not forecast Bitcoin's USD price. The two regimes the chart resolves cleanly are the high band (BTC-led) and the low band (Altcoin-led); everything between is transitional. | Reading | Regime | What it has historically meant | | -------- | --------------- | ------------------------------------------------------------------------------------------------ | | ≥ 70% | BTC-led | Capital concentrated in Bitcoin to the exclusion of the broader complex. | | 60 – 70% | BTC-leaning | BTC outperforming the rest meaningfully — late-bear bottoming or first leg of a new cycle. | | 50 – 60% | Mid-band | The transition zone. No strong directional read on the BTC-vs-alts axis. | | 40 – 50% | Altcoin-leaning | Rotation has carried meaningful share to the rest of the basket. | | < 40% | Altcoin-led | Bracketed only the 2017–2018 ICO mania (ATL 31.1% on 16 Jan 2018) and the 2021 alt-season floor. | ## The denominator changed character every cycle The dominance line moved from 95% to 38% over 2013–2018 partly because of Bitcoin, partly because of what entered the denominator. The basket gained new verticals every cycle: - **Pre-2017** (~80–95%): Bitcoin plus a few dozen forks and clones — Litecoin, Namecoin, Dogecoin, Peercoin, XRP. Ethereum's 2015 mainnet was the only structurally new entrant. - **2017–2018 ICO era** (85% → 31%): Ethereum's ERC-20 standard made it trivial to mint tokens; thousands of ICOs joined the basket. Dominance reached an [all-time low of 31.1% on 16 January 2018](https://www.coingecko.com/research/publications/bitcoin-dominance-history). Bitcoin made a new ATH on the same chart; the denominator was the story. - **2020 DeFi summer** (~60–65%): COMP, YFI and UNI launched mid-2020; total value locked in DeFi went from ~$700M to ~$15B in three months. The denominator gained a new vertical that did not exist in 2018. - **2021 L1 Cambrian** (69% → 38%): Solana, Avalanche, Polygon, Fantom, BNB chain, Terra/Luna gained material market cap. Year-end 2021 closed at ~38%. - **May 2022 Luna collapse** (step up): UST broke peg on 9 May 2022; LUNA fell from $119.51 to effectively zero within four days. Tens of billions of dollars of UST and LUNA market cap were erased from the denominator over the four-day collapse — the [Harvard Corporate Governance Forum](https://corpgov.law.harvard.edu/2023/05/22/anatomy-of-a-run-the-terra-luna-crash/) post-mortem covers the run mechanics. - **2023–2024 stablecoin growth** (structural drag): The aggregate stablecoin float rebuilt to ~$160B by August 2024 and ~$315B by April 2026 — a structural denominator addition not present in 2017. - **2024+ ETF era** (~52% → 60%+): U.S. spot Bitcoin ETFs launched 11 Jan 2024, opening a regulated demand channel that did not exist in 2021. The post-halving cycle has held above 50% throughout — the first cycle peak above that level since 2017. ## When it fails Dominance is not a directional Bitcoin signal. It is a relative-performance ratio. Reading high dominance as "bullish for Bitcoin price" or low dominance as "bearish" mistakes the ratio for the numerator. The 2018 cycle low printed dominance back above 50% with Bitcoin capitulating; the 2017 cycle top printed dominance below 40% with Bitcoin at a new ATH. Stablecoin issuance moves dominance mechanically. A $5B mint enters the denominator as new market cap; dominance compresses on a market-flat day where nobody traded Bitcoin. The [Altcoin Season Index](https://btcoak.com/altcoin-season) excludes stablecoins by construction; when dominance and the ASI disagree, the disagreement is the signal. The basket's character changes cycle to cycle. Pre-2017 dominance had a few dozen alts; 2026 dominance has a ~$300B stablecoin float, several thousand DeFi tokens, and dozens of L1 chains. Cross-cycle comparisons are robust to trend; not robust to what specifically the rest of the basket contains. The [Real Bitcoin Dominance Index](https://charts.bitbo.io/bitcoin-dominance/) strips stablecoins and non-PoW assets and prints in the 70–76% range. ## Frequently asked **What is Bitcoin dominance?** The ratio of Bitcoin's market capitalisation to the sum of every tracked cryptoasset's market cap. **What does it mean when Bitcoin dominance goes down?** Falling dominance means the rest of the cryptoasset complex is gaining market-cap share faster than Bitcoin. Mechanically it can also fall on stablecoin issuance. **Is high Bitcoin dominance bullish?** Bullish for Bitcoin relative to the rest of the cryptoasset complex; not directly bullish for Bitcoin's USD price. The 2018 bear-market low printed dominance back above 50% even as BTC was capitulating. **What is the lowest recorded Bitcoin dominance?** 31.1% on 16 January 2018, the peak of ICO mania. --- # Stablecoin Market Cap URL: https://btcoak.com/stablecoin-supply Category: macro ## What it is Stablecoin Supply is the aggregate USD market cap of every tracked dollar-pegged stablecoin (USDT, USDC, DAI, FDUSD, and the long tail). Each dollar of float is a dollar already inside the cryptoasset perimeter — converted from fiat, held on-chain or on a venue, ready to rotate into Bitcoin or alts without the bank-rail wait. ## How it is calculated The construction is a straight sum: ``` Aggregate float(t) = Σᵢ circulating_supplyᵢ(t) ``` across every tracked USD-pegged stablecoin _i_, with each issuer's circulating supply valued at the $1.00 peg. Wrapped representations of the same underlying token (bridged USDC on multiple chains, etc.) are deduplicated by issuer to avoid double-counting. The regime classifier is keyed off the trailing 30-day percent change. Aggregate float has grown by orders of magnitude since 2015 — sub-billion to hundreds of billions — so a fixed dollar threshold would be meaningless; the percent-change framing stays calibrated. Thresholds: > +5% reads as Expansion; < −2% reads as Contraction; between those is Stable. Tether CEO Paolo Ardoino has framed the company's on-chain mints as ["authorized but not issued… this amount will be used as inventory for next period issuance requests and chain swaps"](https://x.com/paoloardoino/status/1823331317624311866) (X.com, 13 August 2024). A $1B mint shows up as a clean step in the float regardless of whether the underlying counterparty is buying Bitcoin, paying for soybean shipments, or settling an inter-exchange redemption. ## How to read it | Reading | Regime | What it has historically meant | | ----------- | ----------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- | | > +5% / 30d | Expansion | Float growing faster than its long-run pace. Has clustered with bull-cycle phases — the 2020 DeFi-summer kickoff, the 2024 ETF-launch wave. | | −2% to +5% | Stable | The chart's default regime — structural drift, no fresh contrarian read. The float rarely flips between the two outer regimes without a discrete event. | | < −2% / 30d | Contraction | Net redemptions or rotation out. Has clustered with stress windows — the 2018 bust, the 2022 Luna/FTX cascades, the March 2023 USDC depeg. | ## The USDC SVB week — what aggregate supply means under stress The cleanest stress test of "aggregate float at a $1 peg" in the indicator's history is the second weekend of March 2023. Silicon Valley Bank closed on Friday 10 March 2023; later that night, Circle [disclosed](https://www.cnbc.com/2023/03/11/stablecoin-usdc-breaks-dollar-peg-after-firm-reveals-it-has-3point3-billion-in-svb-exposure.html) that $3.3 billion of USDC reserves — roughly 8% of total — were stuck at SVB. By 2am ET Saturday 11 March, USDC was changing hands on secondary venues at $0.87, a 13% break from the peg. The Treasury, Federal Reserve and FDIC issued a [joint statement](https://www.federalreserve.gov/newsevents/pressreleases/monetary20230312b.htm) on Sunday 12 March guaranteeing all SVB depositors. Circle restored redemptions on Monday 13 March and USDC re-pegged. The aggregate-float chart did not record the depeg — it values every coin at $1.00 by construction — but the underlying issuer was, for 36 hours, an open question of redeemability. Two takeaways. First, **the peg is an issuer-credit assertion, not a market clearing.** During a 36-hour window, the secondary-market clearing price said $0.87 and the aggregate-float chart said par. Aggregate supply expanding does not always mean dry-powder accumulating; aggregate supply flat does not always mean a stable peg. Second, **composition risk is hidden in the aggregate.** One issuer stumbling does not register on the float for as long as the peg holds; only when redemptions happen at the wrong end of the issuer's balance sheet does it. The [NYDFS BUSD wind-down](https://www.dfs.ny.gov/consumers/alerts/Paxos_and_Binance) of February 2023 took roughly $16B out of the float over several months — a regulatory event, not a market event — and the chart records it as a slow Contraction. ## When it fails Aggregate supply at $1.00 is not the same as redeemable-at-$1.00. The float values every issuer's circulating supply at the peg, regardless of whether the underlying redemption is being honoured. The USDC SVB week saw secondary markets clear at $0.87 while the aggregate chart kept printing par. Issuance-vs-redemption asymmetry breaks signal symmetry. A Tether mint can be inventory replenishment with no immediate market expression. A USDC redemption by a single institutional treasury for routine cash management can compress the aggregate without changing crypto demand. The indicator is descriptive on the order of months and noisy day-to-day. Composition is invisible in the aggregate. The NYDFS BUSD wind-down (13 February 2023) removed ~$16B of float over several months. The chart records that as a slow Contraction; per-issuer breakdown would have shown the structural shift more clearly. Not every dollar of float is waiting to buy crypto. A meaningful fraction of USDT circulates inside emerging-market trade finance and remittance rails — the "digital dollar" use case. There is no clean way to net that fraction out from the aggregate. ## Frequently asked **What does stablecoin supply tell me?** Each dollar of stablecoin supply is a dollar already inside the cryptoasset perimeter. Rising supply is interpreted as accumulating dry powder; contracting supply implies redemptions back to fiat or rotation into risk assets. **Which stablecoins are included?** Every tracked USD-pegged stablecoin — USDT, USDC, DAI, FDUSD, PYUSD, TUSD, and the long tail of smaller issuers. **How do I read the regime labels?** A trailing-30-day change above +5% reads as Expansion; below −2% as Contraction; between those, Stable. **What is the biggest stablecoin?** Tether (USDT), by a margin. As of early 2026, USDT is roughly 60% of the aggregate float; USDC is the second-largest at roughly 24%. **Is stablecoin supply a leading indicator for Bitcoin?** Mildly — descriptive, not predictive. Aggregate supply expansion has clustered with bull-cycle phases, but the lead/lag is unstable. --- # Altcoin Season Index URL: https://btcoak.com/altcoin-season Category: macro ## What it is The Altcoin Season Index is a 0–100 composite. It scores how much of the broader altcoin complex is outperforming Bitcoin over a trailing 90-day window. Stablecoins and asset-backed tokens (USDT, USDC, DAI, FDUSD, WBTC, stETH, cLINK…) are excluded so the score isolates direct rotation between Bitcoin and the rest of the cryptoasset complex. ## How it is calculated The methodology is short enough to quote intact. The [canonical statement](https://www.blockchaincenter.net/altcoin-season-index/) reads: > If 75% of the Top 50 coins performed better than Bitcoin over the last season (90 days) it is Altcoin Season. With a single exclusion clause: > Excluded from the Top 50 are Stablecoins (Tether, DAI…) and asset backed tokens (WBTC, stETH, cLINK,…). Cryptoasset markets clear continuously, so "90 days" is 90 calendar days, not 90 trading days. The Top 50 is rebalanced daily as market-cap ranks shift. The index value at any date _t_ is the share of the eligible sample whose trailing 90-day return has beaten Bitcoin's, multiplied by 100. ## How to read it The index is a directional rotation signal, not a price target. Sustained readings at or above 75 mark Altcoin Season — three of every four top-50 cryptoassets, ex-stables, ex-wrapped, have outperformed Bitcoin on a trailing 90-day return. Sustained readings at or below 25 mark Bitcoin Season — only one in four. Crossings of the 50 line mark the regime handover, but the meaningful regime shifts happen at the 25 and 75 thresholds. | Reading | Regime | What it has historically meant | | ------- | --------------- | ------------------------------------------------------------------------------------------------------------------------------------------------ | | ≥ 75 | Altcoin Season | Three of every four top-50 cryptoassets have beaten Bitcoin on a trailing 90-day return. Historically a late-cycle distribution regime. | | 50 – 74 | Altcoin-leaning | A majority of the sample is beating Bitcoin but the supermajority threshold has not cleared. | | 25 – 49 | BTC-leaning | A minority of the sample is beating Bitcoin. Most of the chart's history sits in this band — the "quiet" default regime. | | < 25 | BTC Season | One in four or fewer of the eligible sample has outperformed Bitcoin. Brackets late-bear capitulation windows and the opening leg of new cycles. | ## Time in each zone, by cycle A different cut of the same data: how many days each cycle has spent in each zone. The index series begins April 2017, so the 2012 and 2016 halving cycles predate coverage. Cycles are anchored on the 2020 and 2024 halving block-confirmation dates — block 630,000 on 2020-05-11 and block 840,000 on 2024-04-19. The pattern is asymmetric: the BTC-leaning band hosts more days than any other zone in every cycle; full Altcoin Season (the 75+ band) is rare in absolute terms. ## The index is a definitional artefact "The Altcoin Season Index" refers to a methodology, not a single canonical reading. The 75-of-50 construction was popularised by [blockchaincenter.net](https://www.blockchaincenter.net/altcoin-season-index/), and a top-100 variant runs on [CoinMarketCap's panel](https://coinmarketcap.com/charts/altcoin-season-index/). Same threshold, same window, different basket — and noticeably different readings during transitions. When the 51st-100th-cap cohort outperforms the top-50 (the usual mid-cap-front-running pattern early in a rotation), the top-100 index turns altcoin-season first. Readers cross-checking against the top-100 panel will see a different number with a parallel meaning. The Top-50 universe also drifts cycle-over-cycle. In 2021 the sample leaned heavily DeFi (UNI, AAVE, COMP) and meme-tokens (DOGE, SHIB); in 2024 the sample leaned L1-infrastructure and restaking (SOL, AVAX, ETH-restakers). Comparisons across cycles are sample-robust but not cohort-identical. ## When it fails The 90-day window lags meaningful regime shifts by 30–60 days. The index is a backward-looking rolling counter; rotations show up in smaller-cap market-cap aggregates and in the dominance line weeks before the trailing-90-day return calculation can register. Treat 75 crossings as confirmation, not signal. Top-50 versus top-100 basket sizing produces different readings during transitions. We ship the top-50 form because the methodology page's verbatim definition specifies it; the top-100 panel will print a different number on the same day with a parallel meaning. Read either; do not splice the two. Stablecoin issuance does not move this chart. Bitcoin Dominance reacts mechanically to stablecoin mints and burns — the stablecoin slug sits in dominance's denominator. The Altcoin Season Index excludes stablecoins entirely, so a $5B USDT issuance on a quiet Sunday rotates dominance without touching this index. When the two charts disagree, the disagreement is the signal. ## Frequently asked **What is the Altcoin Season Index?** A 0-100 composite scoring how much of the top-50 cryptoassets, ex-stables and ex-wrapped, are outperforming Bitcoin on a trailing 90-day window. ≥75 is Altcoin Season; ≤25 is Bitcoin Season. **What number is altcoin season?** A reading at or above 75 — three of every four top-50 cryptoassets, ex-stables and ex-wrapped, have outperformed Bitcoin on the last 90 calendar days. **How is altcoin season calculated?** Take the top 50 cryptoassets by market capitalisation, exclude stablecoins and asset-backed tokens, count how many have outperformed Bitcoin on the trailing 90-day return, divide by the eligible sample, multiply by 100. **Are we in altcoin season right now?** Read the live cell on /altcoin-season. An Altcoin Season call requires the index to clear 75 and stay there; the index lags the underlying rotation by 30-60 days because of the 90-day window. --- # BTC vs Global M2 URL: https://btcoak.com/btc-vs-m2 Category: macro ## What it is Two log-scaled series on one plot: Bitcoin's USD price on the left, the chosen M2 money-supply stock on the right. The thesis is simple — if Bitcoin is a monetary hedge, price should compound with the denominator. The chart lets you see whether it does. The basket toggle switches between the global aggregate (a USD-translated sum of major central-bank money supplies) and the US series (the Federal Reserve's H.6 release). ## How it is calculated The chart pairs three derived values: the M2 supply level on the right axis, the year-over-year growth rate of that supply, and the rolling Pearson correlation between Bitcoin price and M2 supply. **M2 YoY is locally derived.** Rather than trusting an upstream year-over-year field, we compute YoY from the supply level itself: for each observation, supply is divided by the closest available print to the same calendar day one year prior, minus one. That keeps both baskets on a single auditable definition and lines the US read up with the [Federal Reserve's H.6 series methodology](https://www.federalreserve.gov/releases/h6/about.htm) — verbatim: > M1 consists of the most liquid forms of money, namely currency, demand deposits, and other liquid deposits… The non-M1 components of M2 are small-denomination time deposits and retail money market funds. **Global M2 is a provider construction.** No central authority publishes "global M2." The aggregate is a USD-translated sum of major central-bank money supplies — the choices are which countries to include, how to translate non-USD M2 into USD, and how to handle definitional differences across jurisdictions (European M3 ≠ U.S. M2; Japanese M2 excludes postal-savings deposits in some periods). Different providers get different numbers from the same underlying source data. **The correlation is Pearson.** Standard Pearson on the daily-level (price, M2-supply) pairs, rolled forward to the latest observation across a window. The pipeline ships the 180-day rolling correlation; the page recomputes 90-day and 365-day correlations inline against the same series. ## How to read it The 180-day rolling correlation is the chart's primary regime classifier. | Reading | Regime | What it has historically meant | | ------------ | --------------- | ------------------------------------------------------------------------------------- | | ρ ≥ +0.8 | Strong positive | Bitcoin and M2 are compounding together. The canonical "liquidity beta" regime. | | +0.5 to +0.8 | Positive | Meaningful co-movement, with other factors perturbing the relationship. | | −0.5 to +0.5 | Decoupled | Bitcoin's own demand curve has taken over from monetary tailwind. | | ρ < −0.5 | Negative | Bitcoin moving opposite to M2. Rare. Usually a liquidation-cascade or risk-off shock. | ## The correlation decays across windows The chart's headline contribution: rolling correlations _at three windows simultaneously_. Most published Bitcoin-vs-M2 explainers show one window — usually 90 days. We compute all three nightly. The 90-day correlation swings between −0.9 and +0.95; the 365-day correlation smooths nearly all signal; the 180-day window is the conventional middle ground. ## The liquidity thesis is contested The Bitcoin-as-liquidity-beta argument has a primary author. Lyn Alden's September 2024 piece _Bitcoin: A Global Liquidity Barometer_, written by Sam Callahan, frames the case directly: > During this period [May 2013–July 2024], Bitcoin's price exhibited a correlation of 0.94 with global liquidity, reflecting a very strong positive correlation. > Bitcoin moved in the same direction as global liquidity in 83% of 12-month periods and 74% of 6-month periods. Read in full at [lynalden.com](https://www.lynalden.com/bitcoin-a-global-liquidity-barometer/). The counterpoint is worth carrying alongside, not buried. Joseph Wang — a former Federal Reserve open-market trader writing as [Fed Guy](https://fedguy.com/the-mechanics-of-quantitative-easing-and-m2/) — argues that M2 is the wrong measure of monetary conditions: > QE essentially converts Treasury securities into bank deposits, which is basically one form of money to another. > Money that was saved in Treasuries was not money that was going to be spent on goods and services. It seems more likely to be moved into other financial assets, like corporate debt or equities. By Wang's framing, the apparent Bitcoin-vs-M2 correlation in the post-2020 window is partly a relabelling of asset-rotation behaviour, not a deep monetary mechanism. Both views can be true at the same time — Bitcoin can correlate with M2 as a matter of fact while M2 misrepresents the underlying drivers of asset prices. ## When it fails "Global M2" is a provider construction. No central authority publishes a single canonical global aggregate; every series is a sum-of-jurisdictions with provider-specific country lists, FX-translation methodologies, and definitional reconciliation. Two reputable global M2 series can disagree by trillions of dollars on the same day. The thesis is contested. Wang's 2020 piece on QE and M2 argues that M2 doesn't measure liquidity in the way the Bitcoin-as-liquidity-beta thesis assumes — Treasuries are money for the institutional cohort that drives Bitcoin's marginal pricing, and a QE reshuffle from Treasuries to bank deposits expands M2 without changing "real" liquidity. Take the correlation reading as evidence; do not take it as a closed thesis. Correlation is not causation, and the 180-day window is short. Two assets can correlate strongly for a quarter purely because they both respond to a third factor. The 365-day correlation is the closer-to-structural read; the 90-day correlation is the closer-to-tactical read. Reporting cadence introduces a known lag. M2 supply data lags real-time by several days for the U.S. series and up to a week for the global aggregate. The rightmost edge of the M2 line is systematically a few days behind the rightmost edge of the spot price line. ## Frequently asked **Why plot Bitcoin against M2?** M2 is the broadest conventional measure of money supply. Bitcoin's narrative as a monetary hedge predicts price should move with the denominator. **What is the difference between Global and US M2?** Global M2 aggregates money supply across the major central banks, translated into USD. US M2 is the Federal Reserve's H.6 series — the narrower, more frequently updated, and more politically watched number. **Is Bitcoin correlated with M2?** Lyn Alden's September 2024 piece reports a whole-period correlation of 0.94 between Bitcoin and global liquidity from May 2013 to July 2024, with directional alignment at 83% over rolling 12-month windows. The instantaneous correlation decays sharply at shorter windows. **Does Bitcoin lead or lag M2?** Some analysts argue Bitcoin tracks global liquidity with a lag of roughly 70 to 100 days. The optimum lag is unstable across regimes — reported in the 56-to-107-day range across analyst writeups — and the relationship is descriptive rather than predictive. **How do I read the correlation regime?** The 180-day rolling Pearson correlation is the headline regime classifier. Above +0.8 is Strong positive; +0.5 to +0.8 is Positive; −0.5 to +0.5 is Decoupled; below −0.5 is Negative. --- # Drawdown from All-Time High URL: https://btcoak.com/drawdown Category: cycles ## What it is The Drawdown from All-Time High plots, for every Bitcoin daily close since 18 July 2010, the percent gap between that day's close and the highest close the network had ever recorded up to that point. The series is non-positive by construction. It sits at zero on every day a new all-time high prints, and it grows more negative the further price retreats from the running maximum. It is a backward-looking statistic with no view on where price goes next, only on where it has been relative to its own ceiling. ## How it is calculated The input is a single series: every Bitcoin daily close in USD. For every day _t_: ``` ath_t = max(price_0, price_1, …, price_t) dd_t = price_t / ath_t − 1 ``` Both quantities expand monotonically: `ath` can only go up, `dd` only ever crosses back to zero by setting a new `ath`. The series carries one extra column — `days_since_ath` — which counts calendar days since the running maximum was last set; that is the number that runs into the hundreds during deep bear phases and resets to zero on every fresh peak. The chart uses the merged daily-close history (Mt. Gox-era 2010–2013 from a one-time CSV, post-2013 from the live daily feed). There is no smoothing, no model fit, and no live API at compute time — drawdown is a pure pass over the daily-close file. ## How to read it Locate today's mark on the right edge and trace the basin to the left: how far below the top edge it sits is the current drawdown; how far back it stretches before crossing the zero line is the time since the last all-time high. The five regime bands carry the historical context for each depth band. Thresholds are conventional and community-set, not original to this site. | Reading | Regime | What it has meant | | ------------ | -------------------- | --------------------------------------------------------------------------------------- | | 0% to −5% | Near ATH | Within touching distance of the high. Routine inside trending markets; resets often. | | −5% to −15% | Shallow pullback | Mid-bull pause. Smaller than any cycle correction on record. | | −15% to −35% | Correction | Fires one to two times inside most bull markets. The 2021 mid-cycle crash printed here. | | −35% to −55% | Bear territory | The band that has preceded every major recovery on the record so far. | | −55% to −80% | Deep bear | The 2018 and 2022 troughs finished in this band, at −84% and −77% respectively. | | ≤ −80% | Historic bottom zone | Only the 2014 cycle reached this depth (−86% in early 2015). | ## Historical readings Walking the daily series and emitting one row for every multi-month underwater stretch that hit at least a 20% drawdown surfaces the cycle anatomy directly. Every cycle's deep drawdown is here, plus the 2021 mid-cycle correction (the one that briefly took spot from the April $64,863 high to roughly $30k that summer) and any post-2024 dips. The page emits each peak date, the trough date, the drawdown depth, and the days underwater — all computed from the daily-close file, never hard-coded. By cycle on close-of-day data: - 2013–2015 — peak ~$1,163, trough roughly −86% in early 2015 ($152 area) - 2017–2018 — peak $19,783, trough −84% on 15 Dec 2018 ($3,122) - 2021–2022 — peak $68,789, trough −77% on 21 Nov 2022 ($15,587) - 2024+ — current cycle still in progress, with new all-time highs resetting the running maximum ## Bear-market anatomy Each cycle trough has been shallower than the last. Time-underwater shows the same compression, slightly less cleanly: roughly 1,180 days from peak to fresh ATH after the November 2013 top, 1,060 days after the December 2017 top, and 850 days after the November 2021 top. Cycle drawdowns are still long — measured in years, not months — but each cycle's bear has been shorter and shallower than the last, which is what a maturing asset class is supposed to look like. | Cycle peak | Peak close | Cycle trough | Drawdown | Days underwater | | ---------- | ---------- | ------------ | -------- | --------------- | | Nov 2013 | ~$1,163 | early 2015 | ~−86% | ~1,180 | | Dec 2017 | $19,783 | Dec 2018 | −84% | ~1,060 | | Nov 2021 | $68,789 | Nov 2022 | −77% | ~850 | The honest qualifier: four full cycles is not a sample size to extrapolate from. Whether the next bear bottoms at −65% or −80% is the kind of question this chart cannot answer. ## When it fails Drawdown is descriptive, not predictive. It tells a reader where price sits relative to the highest close ever observed. It does not constrain the next print: a −55% drawdown can deepen to −80% before any recovery, and a −80% drawdown can sit at −80% for another six months before the bottom is in. History says where the historical floor has been. It does not say where this floor will be. The four-year cycle that produced these troughs may itself be ending. Matt Hougan argued in late 2025 that the cycle pattern under which prior drawdowns sat may itself be ending: "The bitcoin halving is by definition half as important as it was four years ago" (https://experts.bitwiseinvestments.com/cio-memos/the-four-year-cycle-is-dead-welcome-to-the-ten-year-grind), with ETF flows, regulation, and corporate adoption substituting for the older halving-driven dynamics. Lyn Alden has framed the same point as a liquidity-cycle reframe of the four-year mantra. If both are right, future drawdowns may not reach the −75-to-−85% range historical readings normalise around. Daily closes round off intraday lows. The 2010 and 2011 Mt. Gox-era flash crashes produced intraday drawdowns deeper than the daily-close numbers above. The 9 June 2011 incident is the canonical example. The page uses daily closes for the entire series for consistency with the rest of the site; intraday-true drawdown depth at the 2011 trough was meaningfully worse than the close-on-close figure. ## Frequently asked **How much is Bitcoin down from its all-time high?** The page computes spot's distance from the running all-time high inline and labels the regime — Near ATH, Correction, Bear territory, Deep bear, or Historic bottom zone. The spot price refreshes a few times a day in the browser; the historical series is recomputed nightly so the running maximum always reflects the latest daily close. **What is the biggest drawdown Bitcoin has ever had?** On daily-close data, Bitcoin's deepest drawdown was roughly −86% in early 2015, after the 2013 Mt. Gox-era $1,163 peak collapsed to about $152. The 2018 cycle came in at −84% (December 2018, $3,122 low against the 2017 $19,783 peak); the 2022 post-FTX trough at −77% (November 2022, $15,587 against the 2021 $68,789 peak). Each cycle's deepest mark has been a notch shallower than the last. **How long does a Bitcoin bear market last?** On the daily-close record, the 2013–2017 underwater run lasted about 1,180 days from peak to fresh ATH, the 2017–2020 run about 1,060 days, and the 2021–2024 run about 850 days. Time-underwater has compressed cycle by cycle alongside drawdown depth, mirroring the maturation pattern visible across most cycle metrics. **Why does drawdown reset when Bitcoin sets a new high?** Drawdown is a backward-looking statistic: it measures distance to the largest close ever observed. When today's close exceeds yesterday's running maximum, the maximum updates and the gap snaps to zero. The chart therefore alternates between long basins underwater and short flat segments at the top edge whenever a new high prints; that is the construction, not a glitch. **Is drawdown a good buy signal?** Drawdown describes the past, not the future — it has no notion of how much further price could fall before recovering. The deepest readings (−80% and below) have bracketed every prior cycle low, which is suggestive but not predictive: the sample size is four full bear markets. Read it alongside the 200-week MA, MVRV-Z, and Pi Cycle, not on its own. --- # MVRV Ratio URL: https://btcoak.com/mvrv Category: on-chain ## What it is MVRV — Market-Value-to-Realized-Value — divides Bitcoin's market cap by its realized cap. Market cap is today's price times circulating supply; realized cap values each coin at the price it last moved on-chain. The ratio is Bitcoin's analogue of a price-to-book ratio: when market value sits below cost basis (MVRV < 1) the network is at aggregate paper loss; when it sits well above cost basis, unrealised profit is building up. The framework is from Murad Mahmudov and David Puell's October 2018 essay _Bitcoin Market-Value-to-Realized-Value (MVRV) Ratio_, building on the realized-cap framework Antoine Le Calvez and Nic Carter introduced at the Baltic Honeybadger 2018 conference, itself crediting Pierre Rochard for the originator-of-ideas. ## How it is calculated The formula is short, but the inputs deserve attention: ``` MVRV = Market Cap / Realized Cap ``` Market cap is the deterministic product of spot price and circulating supply (the latter from the canonical halving schedule, including any coins long since lost). Realized cap values each unspent transaction output at the price on the day it last moved. btc oak does not have a per-UTXO last-spent-price feed. Instead we rebuild realized cap as a two-bucket weighted sum: ``` Realized Cap ≈ STH supply × STH realized price + LTH supply × LTH realized price ``` This approximation tracks the per-UTXO ground-truth realized cap to within a few percent across the full history; the gap is documented on the methodology page. A per-UTXO ground truth would shift today's MVRV print by roughly a percent or two — meaningful in regimes near the 1.0 or 3.7 thresholds, not at typical mid-cycle reads. ## How to read it MVRV is most informative at the extremes. Sustained sub-1 readings have only fired at cycle bottoms — the regime where the average coin on the network is underwater is statistically rare, compresses sellers, and historically marks the start of recovery. Sustained readings above 3.7 have only fired in 2013 and 2017 cycle peaks; the original framework anchored that ceiling on those two cycles, and it has not fired since. Between 1 and 3.7 covers roughly four-fifths of all trading days and carries weak signal on its own. | Reading | Regime | What it has meant | | ---------------- | ----------------- | ------------------------------------------------------------------------------------------------------- | | MVRV ≤ 0.8 | Deep capitulation | Network deeply at aggregate loss. 2015 trough at 0.54; 2018 trough at 0.69. Most stressed cycle floors. | | 0.8 < MVRV ≤ 1.0 | Below cost basis | Average coin underwater. The 2022 post-FTX trough at 0.80 — the shallowest cycle bottom on record. | | 1.0 < MVRV ≤ 2.4 | Fair-to-elevated | The mid-cycle range. Bitcoin spends more days here than in any other band by a wide margin. | | 2.4 < MVRV < 3.7 | Extended | Late-cycle expansion. Both 2021 peaks lived here; the 2024 pre-halving high topped at 2.76. | | MVRV ≥ 3.7 | Cycle-top zone | Mahmudov & Puell's original topping signal. Last fired Dec 2017 at 5.11×. | ## Historical readings Reading every canonical cycle anchor against the live series surfaces the regime-decay pattern with no further commentary needed. Cycle peaks at 5.63×, 5.39×, 5.11×, 3.94×, 2.93×, 2.76×. Cycle troughs at 0.54×, 0.69×, 0.80×. Both extremes shrinking each cycle, with remarkable regularity. Cycle anchors btc oak computes against: - 2013-04-10 — 2013 Apr peak - 2013-12-04 — 2013 Nov peak - 2015-01-14 — 2015 cycle low - 2017-12-17 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2021-04-14 — 2021 Apr peak - 2021-11-10 — 2021 Nov peak - 2022-11-21 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## The 3.7 ceiling decay The cleanest way to see the regime shift is the per-cycle MVRV peak in order: | Cycle | Peak | Hit 3.7? | | ------------- | ----- | -------- | | Apr 2013 peak | 5.63× | Yes | | Nov 2013 peak | 5.39× | Yes | | Dec 2017 peak | 5.11× | Yes | | Apr 2021 peak | 3.94× | Yes | | Nov 2021 peak | 2.93× | No | | Mar 2024 peak | 2.76× | No | Three of six cycle peaks fired the 3.7 ceiling — all in the pre-2018 era. The April 2021 peak at 3.94 was the only post-2017 incursion, and even that print sat closer to the 2.4 lower-bound than the 5+ readings of 2013 and 2017. November 2021 fell short by almost a full unit. March 2024 fell short by 0.94. The trend is monotonic. Two structural causes drive the decay. First, lost coins have grown as a share of issued supply: every Patoshi-era coin that has not moved since 2010 sits in realized cap with a near-zero per-coin valuation, biasing the denominator down. Chainalysis research has estimated 2.78 to 3.79 million BTC permanently lost; Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have never moved. Second, the market has matured: institutional inflows damp peak-cycle euphoria, and the same blow-off shape that produced 5×+ MVRV reads in 2013 and 2017 simply does not characterise the 2021 or 2024 cycles. A practical corollary: a "wait for MVRV ≥ 3.7" rule would have missed the entirety of the 2021 cycle and is currently on track to miss this one. The thresholds Mahmudov and Puell anchored on 2010s data describe the 2010s. For the 2020s, the Extended band (2.4–3.7) is the new de-facto cycle-top range. ## When it fails The 3.7 topping ceiling has stopped firing. Three of six cycle peaks cleared it (2013-Apr at 5.63, 2013-Nov at 5.39, 2017-Dec at 5.11). Three did not (2021-Apr at 3.94, 2021-Nov at 2.93, 2024-Mar at 2.76). The threshold was anchored on the 2013 and 2017 cycles and is now outside the modern operating range. A "wait for 3.7" rule would have missed two cycles consecutively. Realized cap inherits lost-coin distortion. The denominator includes every coin priced at the day it last moved — including Patoshi-era coins priced at $0 and a long tail of coins last moved years ago at near-zero prices. Both biases are one-directional: realized cap reads lower than "active float realized cap" would, and MVRV reads higher. The effect is structural, not stochastic. Our realized-cap reconstruction is not per-UTXO. btc oak builds the denominator from STH/LTH cohort series rather than from per-UTXO last-spent prices. The approximation tracks ground truth within a few percent across the historical record; readings near the 1.0 or 3.7 thresholds are the regimes most sensitive to the gap. Treat MVRV near a regime boundary as "within 0.05 of either side" rather than as a precise cross. Cohort migration noise — coins ageing past 155 days into the LTH bucket — adds a small persistent bias in strong bull regimes. ## Frequently asked **What is the Bitcoin MVRV ratio?** MVRV divides Bitcoin's market cap by its realized cap. Market cap is today's price times circulating supply; realized cap values each coin at the price it last moved on-chain. The framework is from Murad Mahmudov and David Puell's October 2018 essay, building on the realized-cap framework Antoine Le Calvez and Nic Carter introduced at the Baltic Honeybadger 2018 conference, with originator-of-ideas credit to Pierre Rochard. **What does an MVRV above 3.7 mean?** In Mahmudov & Puell's original framing, MVRV ≥ 3.7 marks "overvaluation" — historically a cycle-top zone. The 3.7 figure was anchored on the 2013 and 2017 cycle peaks. On btc oak's daily-close series the 3.7 threshold has not fired since December 2017: April 2021 topped at 3.94 (the only post-2017 incursion), November 2021 at 2.93, and March 2024 at 2.76. Treat the 3.7 ceiling as a 2010s-era convention rather than a modern signal. **How is MVRV calculated?** MVRV = Market Cap / Realized Cap. Market cap is spot price × circulating supply. Realized cap values each unspent transaction output at the price it last moved on-chain. btc oak reconstructs realized cap as a two-bucket weighted sum — short-term-holder supply × STH realized price plus long-term-holder supply × LTH realized price — because a per-UTXO last-spent-price feed is not available to us. Empirically the reconstruction tracks a per-UTXO ground truth within a few percent. **What does MVRV below 1 mean?** MVRV below 1 means Bitcoin's market cap is smaller than its realized cap — the average coin is held at a paper loss. The regime has bracketed every cycle bottom on btc oak's record: 0.54 in January 2015, 0.69 in December 2018, and 0.80 in November 2022 (post-FTX). Capitulation has shrunk cycle by cycle: the 2015 bottom dipped well below 0.6, while the 2022 trough barely crossed 1.0. A "wait for sub-1" rule still works as a generational-bottom signal but takes shallower readings each cycle. **Who created the MVRV ratio?** Murad Mahmudov and David Puell published _Bitcoin Market-Value-to-Realized-Value (MVRV) Ratio_ on Medium on 1 October 2018. The piece explicitly credits Pierre Rochard as the originator-of-ideas for the realized-cap framework and Antoine Le Calvez and Nic Carter as the pair who developed and debuted the metric publicly at the Baltic Honeybadger 2018 conference on 23 September 2018. MVRV was the first metric built on top of realized cap; the rest of the realized-cap family followed. --- # MVRV Z-Score URL: https://btcoak.com/mvrv-z Category: on-chain ## What it is The MVRV Z-Score restates the MVRV ratio in standard-deviation units. Where MVRV simply divides market cap by realized cap, the Z-score subtracts realized cap from market cap and then divides by the expanding-window standard deviation of market cap. The result is a scale-adjusted overvaluation gauge introduced by the pseudonymous analyst Awe & Wonder in 2018, building on the MVRV ratio Murad Mahmudov and David Puell had published earlier that year — itself built on the realized-cap framework Antoine Le Calvez and Nic Carter debuted at the Baltic Honeybadger 2018 conference. ## How it is calculated The formula is the Awe & Wonder original: ``` Z = (Market Cap − Realized Cap) / Stdev(Market Cap) ``` The numerator is just the unrealised gap: how far above (or below) cost basis the network sits in dollar terms. The denominator is what gives the Z-score its cross-cycle bite — the expanding-window standard deviation of market cap, computed over every observation in the series up to and including today. The original Awe & Wonder series, and the popular dashboard reconstruction that propagated it, both use the expanding window rather than a rolling one. btc oak follows that convention so the +7 / +0.1 reference lines remain comparable across cycles. Why standardise? Raw MVRV is a unitless ratio, but its empirical thresholds drift. As Bitcoin's realized cap matures, the same level of unrealised profit per realised dollar becomes harder to reach. The Z-score sidesteps that drift by translating the gap into the network's own historical volatility units. A +7 print in 2017 (when stdev was ~$30B) and a +7 print in 2025 (when stdev is several hundred billion) both mean the same thing in standard-deviation terms. btc oak does not have a per-UTXO last-spent price feed. We rebuild realized cap as a two-bucket weighted sum — STH supply × STH realized price plus LTH supply × LTH realized price. This approximation tracks per-UTXO ground-truth realized cap to within a few percent, which propagates to a sub-decimal-point shift in Z at typical reads. ## How to read it The Z-score is most informative at extremes. The +7 band has only ever fired at blow-off tops (2011, 2013, 2017); the +0.1 band has bracketed every cycle bottom on the record. The middle zone — between roughly +0.1 and +5 — covers most days and carries weak signal on its own. | Reading | Regime | What it has meant | | ------------- | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------- | | Z ≤ +0.1 | Historic bottom | Market cap within 0.1 stdev of realized cap (often below). Bracketed 2015 (−0.61), 2018 (−0.46), 2020 Covid (−0.16), 2022 FTX (−0.35). | | +0.1 < Z ≤ +2 | Undervalued | Below long-run mean but above deep-bottom. The transition zone after every cycle bottom. | | +2 < Z ≤ +5 | Neutral / mid-cycle | The bulk of trading days — useful as a low-conviction baseline. | | +5 < Z < +7 | Overextended | Late-cycle expansion. April 2021 peak topped at +6.90 here; 2017 cycle passed through on the way to +10.40. | | Z ≥ +7 | Historic top | Awe & Wonder's extreme-overvaluation zone. Fired only in 2011, 2013 (twice), and 2017. Last cleared 19 December 2017. | ## Historical readings Reading every canonical cycle anchor against the live series surfaces both halves of the regime-decay pattern at once. Cycle peaks at +8.68, +7.98, +10.40, +6.90, +3.54, +2.97 — declining each cycle since 2017. Cycle troughs at −0.61, −0.46, −0.35 — also compressing each cycle, though more gently. Cycle anchors btc oak computes against: - 2013-04-09 — 2013 Apr peak - 2013-11-29 — 2013 Nov peak - 2015-01-14 — 2015 cycle low - 2017-12-07 — 2017 cycle top - 2018-12-15 — 2018 cycle low - 2020-03-17 — 2020 Covid low - 2021-02-22 — 2021 Apr peak - 2021-10-21 — 2021 Nov peak - 2022-11-10 — 2022 cycle low (post-FTX) - 2024-03-14 — 2024 pre-halving high ## The +7 ceiling decay The cleanest way to see the regime shift on this chart is the per-cycle Z-score peak, in order: | Cycle | Peak Z | Hit +7? | | ------------- | ------ | ------- | | Apr 2013 peak | +8.68 | Yes | | Nov 2013 peak | +7.98 | Yes | | Dec 2017 peak | +10.40 | Yes | | Apr 2021 peak | +6.90 | No | | Nov 2021 peak | +3.54 | No | | Mar 2024 peak | +2.97 | No | Three of six cycle peaks on record cleared +7 — all of them in the pre-2018 era. Three did not. April 2021 at +6.90 came within breathing distance of the ceiling but missed. November 2021 topped at +3.54, less than half the threshold. March 2024's pre-halving high reached +2.97 — well into the mid-cycle band, despite printing a fresh all-time price high at $73,000. The trend is monotonic since 2017. Two structural causes drive the decay. First, the denominator grows. The expanding stdev of market cap is computed over every observation since August 2011 — and as market cap balloons, so does its variance, faster than the numerator (the dollar gap) can keep pace. A market-cap-realized-cap gap of $1 trillion in 2025 maps to a smaller Z than a $200B gap mapped to in 2017, because the historical stdev anchor under each is wildly different. Second, lost coins distort realized cap. Every Patoshi-era coin that has not moved since 2010 sits in the denominator at a near-zero per-coin valuation, biasing realized cap down. Chainalysis research has estimated 2.78 to 3.79 million BTC permanently lost; Sergio Demian Lerner's Patoshi research identifies roughly 1.1 million Satoshi-era coins that have not moved since 2010. The bias is one-directional and accumulates slowly with circulating supply. ## When it fails The honest answer: the indicator is not broken — but the threshold has drifted, and treating +7 as a current signal is a category error. The Z-score is doing exactly what it's designed to do, which is express the gap between market and realized cap in standard-deviation units. What's changed is the denominator's scale, not the indicator's integrity. The denominator drift is structural. Awe & Wonder calibrated +7 against Bitcoin's 2011–2018 history. In that era, market-cap variance was dominated by 2013 and 2017 blow-off tops, so the stdev anchor was relatively small. Each additional bull cycle adds another extreme observation to the expanding window, enlarging stdev and lowering the Z-equivalent of any future gap. The new working threshold is roughly +5. Five of the six modern peaks are at or above +3, and the April 2021 peak cleared +5 (topping at +6.90). The November 2021 and March 2024 peaks fell short — but those cycles also topped raw MVRV well below the historical 3.7 ceiling, so the shortfall is consistent across both metrics. Treating Z > +5 as "Overextended" rather than waiting for > +7 captures the modern shape of the cycle. Realized cap inherits lost-coin distortion. Both biases are one-directional and accumulate slowly. The bottom-side Z threshold (+0.1 / negative prints) has held up better than the top-side; the deep-bottom band has fired every cycle, including the gentle 2022 trough. And our realized-cap reconstruction is not per-UTXO — readings near the +0.1 threshold are the regimes most sensitive to the few-percent reconstruction gap. ## Frequently asked **What is the Bitcoin MVRV Z-Score?** The MVRV Z-Score restates the MVRV ratio in standard-deviation units. Where MVRV divides market cap by realized cap, the Z-score subtracts realized cap from market cap and then divides by the expanding-window standard deviation of market cap. The result is a scale-adjusted overvaluation gauge introduced by the pseudonymous analyst Awe & Wonder in 2018. **What does an MVRV Z-Score above 7 mean?** In Awe & Wonder's original calibration, Z ≥ +7 marked the historical extreme-overvaluation zone. The threshold cleanly identified the 2011, 2013, and 2017 blow-off tops on the daily-close record. It has not fired since December 2017: April 2021 topped at +6.90, November 2021 at +3.54, and March 2024 at +2.97. Treat the +7 ceiling as a 2010s-era reference rather than a modern signal. **How is the MVRV Z-Score calculated?** Z = (Market Cap − Realized Cap) / Stdev(Market Cap), where the stdev is computed on every historical market-cap observation up to and including today (the expanding window). btc oak follows the original Awe & Wonder definition rather than a rolling 4-year window, so historical reads remain comparable across cycles. The denominator scales with Bitcoin's own volatility regime. **What does an MVRV Z-Score near zero mean?** Z near zero means market cap and realized cap are within a small standard deviation of each other — the network is roughly at fair value relative to its own cost basis. Sustained readings at or below +0.1 have bracketed every cycle bottom on the record: the deepest negative prints fired in January 2015 (−0.61) and December 2018 (−0.46); November 2022 post-FTX touched −0.35; the 2020 Covid flush dropped to −0.16. Roughly 17% of all trading days on btc oak sit at or below the +0.1 deep-bottom line. **Who created the MVRV Z-Score?** The pseudonymous analyst Awe & Wonder introduced the metric in 2018, building on the MVRV ratio Murad Mahmudov and David Puell had published earlier that year — itself built on the realized-cap framework Antoine Le Calvez and Nic Carter debuted at the Baltic Honeybadger 2018 conference. The original Medium post was later removed; an open mirror of the methodology is preserved on the Cryptowords archive at https://cryptowords.github.io/bitcoin-mvrv-z-metric. --- # HODL Waves URL: https://btcoak.com/hodl-waves Category: on-chain ## What it is An age-stratified view of Bitcoin's circulating supply. Coins are bucketed by how long they have sat on chain since their last spend; the chart collapses the full age distribution into two bands at the 155-day cohort boundary — long-term holders (≥155 days unmoved) and short-term holders (<155 days unmoved). Together they sum to 100% of circulating supply at every daily close. ## How it is calculated ``` LTH(t) = supply held in UTXOs last spent ≥ 155 days ago / circulating(t) STH(t) = 1 − LTH(t) change90d(t) = LTH(t) − LTH(t−90d) ``` The 155-day boundary is the empirical threshold Rafael Schultze-Kraft and Kilian Heeg established in their 2020 cohort-analysis paper as the day count at which an unspent transaction output's conditional spend probability flattens to a near-constant baseline. Below 155 days a coin is statistically much more likely to be moved next week than next year; above it, conditional spend probability barely changes with further age. The same threshold drives SOPR, cohort realized price, MVRV, and RHODL. The framework itself is from Dhruv Bansal's April 2018 essay _Bitcoin Data Science (Pt. 1): HODL Waves_. Bansal's original chart slices supply into a dozen age bands; the two-bucket form on this site collapses those bands at 155 days and is sufficient for cycle-level conclusions. ## Regimes | Reading | Regime | What it has meant | | --------------- | ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | | LTH ≥ 78% | Deep accumulation | Long-term holders control four-fifths of circulating supply. Has fired around the late stages of cycle bottoms — late 2015, mid 2019, mid 2023. | | 70% ≤ LTH < 78% | Accumulating | Default range across most of the post-2017 history. | | 60% ≤ LTH < 70% | Mid-cycle | Cohort balance is unsettled. Common in the 12 months following a cycle low and again in the months leading into a cycle top. | | LTH < 60% | Distribution | Long-term holders are spending. Has bracketed the 2013 and 2017 topping windows; the 2021 and 2024 tops did not fall this far. | ## The per-cycle peak-to-trough swing Every modern cycle has redistributed less than the one before it. The cycle-swing table walks each cycle's accumulation peak (LTH at its cycle maximum, typically 6–12 months before the price top) and distribution trough (LTH at its cycle minimum, typically within a quarter of the price top). The difference is the magnitude of cohort rotation in that cycle. The 2013 distribution drained over 25 percentage points of dominance from accumulation peak to topping trough; the 2017 cycle was similar; the 2021 cycle drained roughly 15–18 points; the 2024 cycle to date has drained the smallest band of any cycle on record. The pattern fits the broader maturation story — MVRV, Puell, NUPL, and Reserve Risk all show the same shrinking-amplitude trajectory at the network scale. The institutional-custody era complicates the read. A spot-ETF custody wallet that holds coins for years on behalf of underlying shareholders ages those coins through the 155-day boundary identically to a self-custody cold-storage wallet. The on-chain record shows accumulation; the underlying ownership may have churned through hundreds of distinct shareholders without a single chain transaction. Cohort dominance is a supply-age statement, not an ownership statement, and the gap between the two has widened materially since January 2024. ## When it fails **Cohort dominance is not ownership.** A spot-ETF custodian holding coins on behalf of underlying shareholders shows on-chain as a single LTH wallet; the underlying ownership may rotate through every redemption cycle without any chain transaction. The 2024-onward reading is structurally less informative about distinct-holder behaviour than the pre-2024 reading was. **The 155-day boundary is empirical, not load-bearing.** A coin crossing the cohort line on day 156 reads as long-term; the same coin on day 154 reads as short-term. The boundary is a convenient cut, not a phase transition. **Lost coins inflate LTH dominance.** Every Patoshi-era coin that has not moved since 2010 is permanently in the long-term bucket. The persistent ~1.1 million Satoshi-era coins, plus the broader 2.78–3.79 million BTC that Chainalysis estimates as lost, all sit in the LTH band by construction. The structural LTH baseline is a few percentage points above what an active-supply-only series would read. **Custodial reshuffles fire false signals.** A single exchange cold-wallet rotation can move tens of thousands of BTC across the cohort boundary in a day. The 90-day change smooths most of that noise; spot reads around large custody movements should be treated as transitional rather than substantive. ## Frequently asked LTH dominance has rebuilt to multi-year highs through every post-2014 bear market — 2015, 2018-19, 2022-23 — and contracted into every cycle top. James Check has documented that long-term-holder supply remained anchored in the high 70s through a drawdown of roughly 21% from the 2024 all-time high — behaviour that prior cycles did not exhibit at comparable price drawdowns. Whether this reflects structurally stronger holder conviction, the new mechanics of ETF custody, or simply the not-yet-completed distribution phase of an in-progress cycle is the question this chart will answer over the next several quarters. --- # Daily Issuance URL: https://btcoak.com/issuance Category: stats ## What it is Bitcoin Daily Issuance plots the protocol-determined supply emission per day on a logarithmic axis, with the USD value of that flow on a second panel and the annualised inflation rate alongside. The BTC count per day is a step function set entirely by the halving schedule — there is no policy committee, no discretionary stimulus, no emergency override. The chart shows the entire monetary policy of Bitcoin in one frame: a step function whose remaining slope is fully knowable two centuries out. ## How it is calculated ``` btc(t) = reward(t) × 144 blocks usd(t) = btc(t) × price(t) inflation(t) = (btc(t) × 365) / supply(t) ``` The reward at any date is fixed by the halving schedule encoded in `GetBlockSubsidy()` in Bitcoin Core: 50 BTC per block from genesis (3 January 2009), halving to 25 at block 210,000 on 28 November 2012, to 12.5 at block 420,000 on 9 July 2016, to 6.25 at block 630,000 on 11 May 2020, and to **3.125** at block 840,000 in **April 2024**. The 144-blocks-per-day target is the protocol's 10-minute mean; realised cadence drifts a few percent each month around that mean. ## How to read it The inflation strip is the cleanest read of the three series. It steps down at every halving (the new lower issuance overnight cuts the numerator in half), then bleeds slowly through the following era as the denominator continues to grow. Today's reading sits well below gold's historical mining rate of ~1.5%, and well below every major fiat currency. | Reading | Regime | What it has meant | | ---------- | ----------------- | --------------------------------------------------------------------------------- | | > 4% | Pre-2016 era | Annualised supply growth in line with broad fiat aggregates. | | 1.5 – 4% | Above gold parity | The 2016–2020 era. At or above the gold mining rate. | | 0.5 – 1.5% | Below gold parity | The post-2020 era. Below the leading non-fiat reference asset. | | < 0.5% | Asymptote regime | Post-2032. The BTC count is, for portfolio-modelling purposes, effectively fixed. | ## The post-halving inflation-decay The distinctive number this chart produces is not the headline inflation rate on any given day but the **rate at which inflation decays in the year following a halving**. Each halving cuts the numerator in half overnight; the denominator keeps growing through the new era at the new lower rate, producing a slow further bleed. The size of that further bleed has shrunk dramatically with each halving. The 2012 halving took inflation from 12.51% to 11.12% over the following year (a 1.4-pp slide). The 2016 halving took it from 4.17% to 4.00% (a 0.17-pp slide). The 2020 halving took it from 1.79% to 1.76% (a 0.03-pp slide). The 2024 halving took it from 0.83% to 0.83% (a 0.01-pp slide, two orders of magnitude smaller than 2012's). The annual decay is now vanishing into the noise floor. ## When it fails The USD panel is **not** the policy. It is price times quantity, dominated entirely by spot. A 30% drawdown halves the dollar revenue without any change to the schedule. The deflationary story this chart tells lives in the BTC series and the inflation strip; the dollar overlay is for miner economics, not for monetary modelling. The "days until next halving" countdown is a calendar projection — four years from the most recent halving. The actual on-chain schedule is block-based (every 210,000 blocks at the 10-minute target), and the realised halving date typically drifts a few weeks from the calendar estimate. The 2020 halving landed within 24 hours of the calendar projection; the 2024 halving landed about a month earlier than a naive +4y projection would have suggested. Subsidy is only one half of miner revenue. Transaction fees are the second leg, currently a single-digit percentage in steady state and far higher around blockspace-demand spikes (Ordinals/Inscriptions in 2023, the Runes launch in April 2024). The post-2140 economy lives entirely on the fee column. ## Frequently asked **How much new Bitcoin is issued per day?** Currently 450 BTC — 3.125 BTC per block × 144 blocks. The figure halves at every protocol halving. **What is Bitcoin's annualised inflation rate?** Below 1% as of 2026 — well below gold's historical mining rate of ~1.5%. **When is the next Bitcoin halving?** Projected for April 2028, four years after the April 2024 halving. The schedule is block-based on-chain, so the actual date drifts a few weeks from the calendar estimate. **How is Bitcoin's supply schedule decided?** It is encoded in the protocol. Changing it would require a hard-fork agreed by every node operator — the equivalent of the entire user base voting to debase. **When will all Bitcoin be mined?** The 21-million ceiling is approached asymptotically; the last fractional satoshi is mined around 2140. Roughly 99% of all Bitcoin will be mined before 2036. --- # Rolling CAGR URL: https://btcoak.com/rolling-cagr Category: stats ## What it is Rolling CAGR plots Bitcoin's compound annual growth rate over four lookback windows — 1, 3, 5, and 10 years — recomputed on every daily close. The chart answers the realised-return question at four horizons simultaneously: what would a buyer have earned, annualised, holding from any past date through to today. ## How it is calculated ``` CAGR_n(t) = (price(t) / price(t-n))^(365/n) - 1 ``` The four lookback windows in days are 365, 1,095, 1,825, and 3,650 — the calendar-day equivalents of 1, 3, 5, and 10 years. Each line plots the compound annualised return over that window ending on day _t_. The headline reading at the top of the page recomputes each window against live spot, so the four CAGRs tick with price through the day. The cycle-anchor table holds each anchor's reading constant for cross-cycle comparability. ## How to read it The four windows answer different questions. The 1-year CAGR is essentially a phase indicator — strongly positive in late-cycle expansion, deeply negative in mid-bear, near zero in the transition phases. The 3- and 5-year windows damp the cycle into a slower-moving signal. The 10-year window is the one most worth lingering on: it has converged into a tight cross-cycle band even when the shorter windows span an order of magnitude. | Window | Cycle-top print | Cycle-bottom print | Range | | ------ | --------------- | ------------------ | ------------------------------------ | | 1y | +400% to +800% | −60% to −80% | Volatile, full cycle range | | 3y | +120% to +180% | −10% to +20% | Cycle phase indicator | | 5y | +90% to +120% | +20% to +50% | Smooth, late-cycle weighted | | 10y | +40% to +55% | +25% to +35% | Convergent — the long-horizon answer | ## The 10-year convergence The chart's quiet contribution is the 10-year band. Across every cycle on record, the 10-year CAGR has stayed inside roughly the 25–55% range — even when the 1-year CAGR span through bear-and-bull cycles is six times wider. The compression is the empirical case for long-horizon return modelling: holding periods of a decade or more produce a tight, well-bounded annualised return distribution where shorter periods do not. The mechanism is straightforward. A 10-year window crossing two halvings averages into the long-run network-growth path — the 1- and 3-year windows are dominated by the cycle they sit in. As Bitcoin's cycles mature and the long-run trend dominates more of the variance, the 10-year band has narrowed slightly cycle-over-cycle, which is itself a finding about market maturation. ## When it fails CAGR is a realised-return metric, not a forecast. The three-year reading at any cycle top is by construction a record of how well a buyer three years ago has done — it says nothing about what the next three years deliver. Treat the chart as a description of past success, not a projection. Compounding mechanics make CAGRs fragile near zero. A 1-year window starting after a 70% drawdown shows a nominally large positive CAGR even when the absolute price is far below where it started; comparing the 1-year reading across cycle phases without that context is misleading. The 10-year window's convergence is descriptive, not law. Past long-horizon outcomes have lived in this range; the next decade has no obligation to. Treat the band as a structural prior, not a guarantee. ## Frequently asked **What is Bitcoin's annual return?** Depends on the window. The 1-year CAGR ranges from −80% to +800% across the cycle; the 10-year CAGR has stayed in the 25–55% band on every cross-cycle reading on record. **What is the best long-term return on Bitcoin?** The highest 10-year CAGR on the daily-close history sits in the mid-50% range; the lowest, around 25%. Both numbers anchor on date pairs spanning two halving cycles. **Is Bitcoin a good 10-year hold?** Historically every 10-year holding period has delivered a positive annualised return. The chart is the descriptive record; future windows have no obligation to repeat the pattern. **How does CAGR differ from total return?** CAGR is the geometric annualisation of total return: a 10-year holding period that produced a 10× total return prints a CAGR of ~26%. CAGR is the comparable metric across windows of different length. --- # Rolling Volatility URL: https://btcoak.com/volatility Category: stats ## What it is Rolling Volatility plots Bitcoin's annualised realised standard deviation of daily log returns over three lookback windows — 30, 90, and 365 days. The chart records how much price has moved getting wherever it sits, on the same scale that traditional finance uses for equity-index and commodity volatility. Bitcoin's vol has compressed monotonically across halving epochs, and the chart is the primary visual evidence of that maturation. ## How it is calculated ``` r_i = ln(price_i / price_{i-1}) vol_n(t) = stdev(r over the trailing n days) × sqrt(365) ``` Log returns rather than arithmetic returns to keep the additive scaling correct under annualisation. The annualisation factor is √365 (continuous-compounding convention; cryptoasset markets clear seven days a week, so the calendar-day and trading-day counts collapse). Three rolling windows produce three lines that converge during quiet regimes and diverge sharply through events that move recent days but not the trailing year. ## How to read it The 30-day window captures regime changes within a few weeks; the 365-day window is the slow-moving baseline. The four-band classifier on the 30-day series: | Reading | Regime | What it has meant | | --------- | ---------- | --------------------------------------------------------------------------------------- | | ≤ 30% | Compressed | Late-2026 regime. Sustained calm; below the historical equity-index event-window range. | | 30 – 60% | Normal | Steady-state Bitcoin. The bulk of post-2020 days live here. | | 60 – 100% | Elevated | Active drawdown or expansion. Fits 2018, 2022, parts of 2024. | | > 100% | Extreme | The pre-2016 norm; rare in the post-2020 era. | ## The per-halving-epoch compression The chart's distinctive cut is the per-epoch table — mean and standard deviation of each volatility window across five eras keyed on the halving schedule. Realised vol has compressed monotonically by epoch: - **Pre-2013 era** (genesis → 2012-11-28): mean 30-day vol ~110% - **2013–2016 era** (after halving #1): mean ~70% - **2016–2020 era** (after halving #2): mean ~65% - **2020–2024 era** (after halving #3): mean ~55% - **2024+ era** (after halving #4): mean ~45% The compression is structural, not a single-cycle anomaly. It tracks roughly with deepening market liquidity, broader institutional participation, and the 2024 spot-ETF demand channel. Today's 30-day reading sits in the same band as historical S&P 500 vol around major event windows (Q1 2020, Q4 2008) — Bitcoin used to live two to four times above that level. ## When it fails Realised vol is backward-looking by construction. A reading of 35% today says price was calm over the last 30 days; it says nothing about whether the next 30 days will be. Treat the chart as a descriptive record, not a forecast. Cross-cycle comparison is fragile when the underlying market structure is changing. Comparing 2014's 30-day vol to 2026's is comparing two materially different markets — 2014 had thin order books, retail-only participation, and minimal hedging supply; 2026 has CME futures, US spot ETFs, and active options markets. The compression is real, but its decomposition (market depth vs hedging supply vs participant mix) is not what this chart tells you. The √365 annualisation is convention; alternative formulations using √252 (US-equity trading-day count) produce readings about 20% lower. Cross-asset comparisons need to align on the same annualisation choice. Implied vol (Deribit DVOL, CBOE BITX) has converged with realised vol in the late-2020s; the gap was much wider pre-2018. The two now move together within a few volatility points outside event windows. ## Frequently asked **How volatile is Bitcoin?** As of 2026, the 30-day annualised realised vol is in the 30–50% range — comparable to historical equity-index event-window vol. In 2014 the same metric was 100–120%. **Is Bitcoin's volatility decreasing over time?** Yes. The per-halving-epoch table shows monotonic compression: ~110% pre-2013 down to ~45% in 2024+. The compression tracks deepening market liquidity and broader participation. **How is Bitcoin volatility calculated?** Standard deviation of daily log returns over a rolling window, annualised by √365. Three windows: 30, 90, 365 days. **What is the difference between realised and implied volatility?** Realised vol is computed from past price moves (this chart). Implied vol is the market's forward-looking estimate, computed from options prices. The two have converged in late-2020s readings. --- # Profitable Days URL: https://btcoak.com/profitable-days Category: stats ## What it is Bitcoin Profitable Days plots every daily close since 18 July 2010 on a logarithmic price axis, coloured by a single test: would today's spot price show a paper profit if you had bought on that day? Sage segments are wins; rust segments are coins still underwater at the current price. The reference price is literally today, so the chart updates as spot moves — a path-dependent picture of who has and hasn't made money so far. ## How it is calculated ``` profitable(d) = price(today) > price(d) lifetime_ratio = #{profitable(d)} / total_days ``` The test is intentionally trivial. For every historical close *d*, the day is sage if today's spot is higher and rust if it isn't. The lifetime ratio is the share of days where the test resolves green; the 30-day rolling ratio is the same share over a trailing 30-day window. There is no smoothing, no return-window anchor, no cohort framing — just spot today against close on day *d*. ## How to read it Profitable Days is most informative at the extremes. A lifetime ratio above 97% has only fired during the most expansive cycle phases. A ratio below 80% has only fired during deep bears. Between 85% and 95% covers the bulk of trading days and carries weak signal on its own. | Reading | Regime | What it has meant | | --- | --- | --- | | ≥ 97% | Late expansion | Most of the historical record sits below current spot. Has bracketed every cycle top since 2017. | | 92 – 97% | Mid expansion | A working ratio with a thin rust band. The largest share of post-2015 history. | | 85 – 92% | Drawdown | The rust band is widening as recent highs sit above current spot. Fits 2018, 2022, 2025–2026. | | < 85% | Deep bear | Sustained readings below 85% have only fired in the 2014–2015 and pre-2013 bears. | ## The all-time-low ratio The cleanest single number this chart produces is the all-time-low buyer ratio — the date in Bitcoin's history where the smallest fraction of preceding closes would have been profitable buys. After a 365-day warm-up to give the early history room to develop, the minimum is **58.04%** on **20 November 2011**, with spot at $2.20. That print sits in the depths of the post-Mt Gox 2011 bear — the slide from $32 in June 2011 to $2 in November 2011, after the Mt Gox 19 June 2011 hot-wallet incident and the cascade of secondary failures that followed. With sixteen months of trading on record, an anchor at the bottom left more than four-tenths of the preceding closes above current spot. The 2013 bull run pushed Bitcoin from $2 to over $1,100 in roughly two years, mechanically converting the 2010 and 2011 rust segments to sage as spot cleared every prior daily close. The lifetime ratio has not revisited the sub-60% regime since. Subsequent cycle bottoms (2015, 2018, 2022) shaved the lifetime ratio by 10–15 percentage points but stayed above 80%. The all-time low belongs to the era when there was simply not enough history to absorb a single-cycle drawdown. ## When it fails Path-dependent — every new ATH wipes prior rust. Today's lifetime ratio is a function of today's spot, not a fixed historical truth. Any cross-cycle comparison of the lifetime ratio has to anchor on the same spot, otherwise it is comparing two different questions. The ratio is anchored on today, not on horizon. Coins bought near a cycle peak read as "unprofitable" here even if a buyer has held them for ten years and harvested several intermediate highs. The chart does not know about realised returns. Daily closes only. An intra-day wick that filled by close does not register; an intra-day flash above today's spot does not flip the day to profitable. Bitcoin's intra-day range is materially wider than its close-to-close range. ## Frequently asked **What does "profitable day" mean for Bitcoin?** A day is profitable if today's spot price is higher than the closing price on that historical day. The reference is literally today, so the chart updates as spot moves. **How often has buying Bitcoin been profitable?** Across the full daily-close history, around 95% of days have been profitable at most reasonable spot prices over the last decade. The lifetime ratio rarely drops below 90% outside cycle bottoms. **What is the longest Bitcoin unprofitable stretch?** Depends on today's spot — the metric is path-dependent. The current stretch is the longest since the 2018 bear; every new all-time-high will clip its leading edge. **Is Bitcoin profitable to hold long term?** Historically yes: of every Bitcoin daily close since July 2010, more than nine in ten sit below today's price. The chart says nothing about the future. **How does this compare to drawdown?** Profitable Days reads against today's spot; drawdown reads against the running maximum at every historical date. Different questions. Profitable Days asks "what fraction of past days would a buyer have made money on, at today's price?"; drawdown asks "how far is BTC below its peak right now?" --- # Halving-Cycle Overlay URL: https://btcoak.com/halving-cycles Category: stats ## What it is The halving-cycle overlay plots each of Bitcoin's four post-halving return curves on a shared "days since halving" x-axis, normalised to the day-0 price of each cycle. A point at (365, 3×) means that on day 365 of that cycle, price stood at three times its halving-day close. With four cycles overlaid — 28 Nov 2012, 9 Jul 2016, 11 May 2020, and 19 Apr 2024 — a reader can compare the active cycle against its three predecessors at the same days-elapsed and read the diminishing-returns pattern across cycles directly. ## How it is calculated Inputs are the merged Bitcoin daily-close series and the four halving dates, hard-coded from consensus block heights: ``` 2012-11-28 block 210,000 2016-07-09 block 420,000 2020-05-11 block 630,000 2024-04-19 block 840,000 ``` For each cycle _i_ the close on the halving date is `p₀ᵢ`, and for every day _d_ from 0 to min(1,460, today − halvingᵢ), `multiple(d) = price(halvingᵢ + d) / p₀ᵢ`. Where the requested day falls between two daily closes (or after the last close), the most recent prior close is used; that nearest-prior policy keeps each line continuous on the chart even when the underlying series has gaps. The day-0 anchor is the daily close on the halving date — not the halving-week median, the 30-day mean, or the day-0 open. Each of those alternatives shifts every line by a few percent; the shape of the comparison barely moves. The close is the most objective single number. ## How to read it The chart answers one question: at the same days-elapsed, how does the active cycle compare to the previous three? A line above 1.0× means price is above its halving-day close; below 1.0× means it has fallen back. Read each cycle's shape end-to-end — the common pattern is a slow drift higher for the first six to twelve months, an acceleration into the cycle peak in the back half of year one or early year two, then a year or more underwater before the next halving begins the cycle again. | Reading | Regime | What it has meant | | ---------------- | -------------------- | ---------------------------------------------------------------------------------------------------- | | 0 — 90 days | Post-halving drift | Range-bound or shallow uptrend. The 2024 cycle finished day 90 at ~1.5×; 2020 at ~1.0×. | | 90 — 365 days | Cycle expansion | The fastest-multiplying phase historically. Each prior cycle has broken 2× by day 365. | | 365 — 730 days | Peak window | Where every prior cycle has topped: 2012 near day 365, 2016 near day 525, 2020 near day 545. | | 730 — 1,100 days | Cycle drawdown | Bear-market basin, often with a 70% retrace from the cycle peak. The 2018 and 2022 troughs sit here. | | > 1,100 days | Pre-halving recovery | Recovery to halving-day price and beyond. Each cycle has set a new ATH inside this window. | ## Historical readings Snapshotting each cycle at fixed days-elapsed checkpoints — +90, +180, +365, +500, +730 — and recording the eventual peak multiple makes the diminishing-returns pattern unambiguous. Multiples are computed from each cycle's day-0 close using the daily-close history, with no smoothing. Halving anchors: - 2012-11-28 — block 210,000, day-0 close ~$12.35 - 2016-07-09 — block 420,000, day-0 close ~$650 - 2020-05-11 — block 630,000, day-0 close ~$8,600 - 2024-04-19 — block 840,000, day-0 close ~$63,800 ## Peak-multiple decay By peak multiple, cycle by cycle: roughly 94× from the 2012 day-0 close (peak in November 2013 near $1,163); roughly 30× from the 2016 day-0 close (peak in December 2017 near $19,783); roughly 8× from the 2020 day-0 close (peak in November 2021 near $68,789); the 2024 cycle is still in progress with a peak multiple well below the prior progression. Each cycle has produced a smaller percentage move from day 0 than the cycle before it. The classic explanation: it is far easier to multiply a $12 asset by 90× than a $63,800 asset by even 10×; the marginal capital required to move the market grows super-linearly with the market cap. That mechanical argument is one half of the story. The other half is that some of the older cycles' returns came from raw adoption — a population pricing Bitcoin from the first time. That is a one-time effect. The peak timing is the chart's other regularity. Each cycle has topped roughly twelve to eighteen months after its halving: late 2013 for the 2012 cycle, December 2017 for the 2016 cycle, November 2021 for the 2020 cycle. The 2024 cycle's price action through 2025 sits inside the historical peak-window range — but the ATH set in October 2025 was the most modest of any post-halving high relative to the day-0 anchor. ## When it fails The halving as a structural driver is contested. Matt Hougan published the most direct critique to date in his December 2025 memo _The Four-Year Cycle Is Dead. Welcome to the Ten-Year Grind_ (https://experts.bitwiseinvestments.com/cio-memos/the-four-year-cycle-is-dead-welcome-to-the-ten-year-grind): "I think blindly assuming the four-year cycle will repeat is foolish… The bitcoin halving is by definition half as important as it was four years ago." Hougan substitutes ETF flows, regulatory clarity, and corporate-treasury adoption as the new structural drivers. Lyn Alden made the parallel point: the four-year framing "doesn't really have a reason to exist anymore other than some degree of self-prophecy" (https://cryptobriefing.com/lyn-alden-the-four-year-bitcoin-cycle-is-losing-relevance-institutional-access-is-reshaping-market-dynamics-and-the-current-bear-market-may-be-shorter-than-previous-ones-the-wolf-of-all-streets/). She frames the dominant driver as the global liquidity cycle rather than Bitcoin's issuance schedule. Cycle anchoring is one assumption among several. Anchoring at the halving date is not the only honest choice — researchers have anchored cycles at successive ATHs, at successive 200-week MA crossings, or at calendar quarters. Each anchoring tells a slightly different story. Nico Cordeiro put the structural problem cleanly in a 2020 critique of cycle-extrapolation work: "Bitcoin grew by X in the past, it will grow by X in the future. One should remember that past results are not representative of future returns" (https://strixleviathan.com/a-chameleon-model-why-bitcoins-stock-to-flow-model-is-fatally-flawed/). Sample size. Three completed cycles is not a base from which to fit anything. The current cycle's multiple is one observation; it could be early in a classic late-cycle run-up, or it could be the first full halving since the issuance-driver lost its leading role. The chart tells you what has happened. It does not tell you what will. ## Frequently asked **What does the halving-cycle overlay show?** Each of Bitcoin's four halving cycles — 28 Nov 2012, 9 Jul 2016, 11 May 2020, 19 Apr 2024 — is plotted as a multiple of its halving-day price on a shared "days since halving" x-axis. A point at (365, 3×) means that on day 365 of that cycle, price was three times the halving-day close. With four cycles overlaid you can read where the current run sits against history at the same days-elapsed. **Which Bitcoin cycle has performed best?** By peak multiple from halving day, the 2012 cycle ran the furthest — roughly 94× before topping near $1,163 in November 2013. The 2016 cycle peaked near 30× in December 2017; the 2020 cycle near 8× in November 2021; the 2024 cycle has so far reached a peak multiple well below the 2020 progression. Every cycle since the first has printed a lower peak multiple than the one before. **Does the four-year Bitcoin cycle still work?** Increasingly contested. Matt Hougan argued in late 2025 that "the bitcoin halving is by definition half as important as it was four years ago" and that ETF flows, regulation, and corporate adoption are now the structural drivers of price — not the issuance halving. Lyn Alden has framed the same point as a liquidity-cycle reframe of the four-year mantra. The data on this chart is consistent with both takes: each post-halving multiple has been a fraction of the last. **How is "days since halving" computed?** Calendar days from the halving block's confirmation date forward, capped at 1,460 (≈ four years) per cycle. Each cycle starts at day 0 with multiple 1.0× and runs forward as the cycle ages; cycles that haven't completed yet just end early. Halving block confirmations: block 210,000 on 28 Nov 2012, 420,000 on 9 Jul 2016, 630,000 on 11 May 2020, and 840,000 on 19 Apr 2024. **Why is the 2012 cycle line so noisy?** The daily-close history starts 18 Jul 2010 — Bitcoin's first week of liquid trading. The 2012 halving lands ≈ 850 days into that history, which means day-0 is well-covered, but the surrounding Mt. Gox-era prints had thin order books and wide spreads, so the percentage moves on individual days were dramatic. That volatility is part of the historical record and stays unsmoothed. --- # Time in Band URL: https://btcoak.com/time-in-band Category: stats ## What it is Time in Band is a histogram of how many days Bitcoin has spent inside each of the nine Rainbow valuation bands — both across the full daily-close history and across each halving cycle separately. The chart's headline finding is that the distribution is dramatically asymmetric: years in the lower bands, weeks in the Euphoric. The asymmetry has deepened cycle by cycle. ## How it is calculated ``` band(d) = which Rainbow band the close on day d sits in days_in_band[k] = count of d where band(d) == k share[k] = days_in_band[k] / total_days ``` The Rainbow regression is refit nightly via OLS on `log10(price) = a × ln(days_since_genesis) + b`; the nine bands sit at parallel `0.1`-log offsets above and below the centre line (`BAND_OFFSETS` in `$lib/theme`). Each daily close is classified into its current band and tallied. The full-history histogram covers genesis through today; the per-cycle histograms slice the same data into five windows defined by the halving schedule. The cycle windows are: pre-2013 (genesis → 2012-11-28), 2013–2016 (after halving #1), 2016–2020 (after halving #2), 2020–2024 (after halving #3), and 2024+ (the current cycle). ## How to read it The all-time histogram shows roughly three-quarters of all Bitcoin trading days clustering in the middle five bands (3 through 7). The lower bands (1 and 2) are oversampled by 2010–2011 history when the regression's centre line had not yet stabilised. The upper bands (8 and 9 — Hot Stuff and Euphoria) are the tails of the distribution; combined they account for under 5% of all days. The cycle-level distribution is where the analytical contribution lives. Each cycle's band 9 frequency is the page's distinctive: how often did Bitcoin trade in the topmost band, normalised to that cycle's length? ## The Euphoria-band frequency drop | Cycle | Window | Band 9 days | Share | | ------------------ | ----------- | ---------------- | ------- | | Cycle 1 (pre-2013) | ~1,600 days | hundreds of days | ~10–15% | | Cycle 2 (2013–16) | ~1,300 days | dozens | ~3–5% | | Cycle 3 (2016–20) | ~1,400 days | a handful | ~0.5% | | Cycle 4 (2020–24) | ~1,500 days | zero | 0% | | Cycle 5 (2024+) | running | zero | 0% | The frequency of Euphoria-band readings has fallen monotonically across cycles. Cycle 1 hosted Euphoria-band days in the double-digit percent range; cycle 4 did not produce a single Euphoria-band day across roughly 1,500 trading sessions. This is the visual case for the broader cross-cycle observation that each cycle is smaller in ratio space than the last — the same finding that NUPL, MVRV, and several other indicators record in their own units. The mechanism is structural. The Rainbow centre line drifts upward over time on the OLS refit, so each cycle's prices have a higher bar to clear before band 9 is reached. Combined with the cycle-decay observed in absolute peak multiples (94× in 2013, ~30× in 2017, ~8× in 2021, lower still in 2024), the Euphoria band has become statistically out of reach in the post-2020 era. ## When it fails Bands move when the regression refits. A cell's per-cycle count can shift by a few days between consecutive nightly builds; cross-cycle comparisons rest on the bands at refit-time, not at the time the cycle was unfolding. The histogram is path-dependent on the OLS fit. As more data accumulates, the fit's centre line drifts; today's band 5 is not the same dollar range as 2015's band 5. Treat the chart as a relative-band classifier, not an absolute price reference. The pre-2013 cycle is genuinely different in character — thin order books, exchange instability, single-event price moves. Its band-9 hosting rate is not directly comparable to later cycles' for that reason; it is part of the historical record but the more interesting question is the cycle-2-onwards monotonic compression. Layer-2 migration drains base-layer activity that would otherwise contribute to the cycle's running average. Whether that biases the residency distribution is an open methodological question; the chart records the residency, not the underlying activity. ## Frequently asked **What are the Rainbow bands?** Nine valuation bands fit by OLS regression on Bitcoin's full price history in log space, refit nightly. The bands sit at parallel offsets above and below the regression centre line. **How much time has Bitcoin spent in the Euphoria band?** Across the full history, roughly 1–3% of all trading days. The fraction has dropped to zero in the most recent cycle. **Why does Bitcoin spend so little time at the top?** Late-cycle blow-off phases are short by construction — the price has to outrun its own log-regression centre line by a meaningful multiple, which has only happened during the 2013, 2017, and (briefly) 2021 cycle tops. **Is Bitcoin underpriced when it's in the lower bands?** The chart records residency, not directional signal. Lower-band readings have historically clustered around cycle bottoms; they have also persisted for years in mid-cycle phases. --- # Monthly Returns Heatmap URL: https://btcoak.com/monthly-returns Category: stats ## What it is Bitcoin Monthly Returns is a year-by-month grid of Bitcoin's monthly returns since 2010, with annual totals down the right edge and per-month averages across the bottom. Cells are coloured by sign and magnitude. The chart is the evergreen seasonal-performance reference for Bitcoin — and the page's distinctive contribution is the per-month statistical table that tests whether any of the visible patterns survive a basic dispersion check. ## How it is calculated ``` monthly_return(y, m) = close(last day of month m, year y) / close(last day of prior month) − 1 annual_total(y) = product over m of (1 + monthly_return(y, m)) − 1 ``` Month-boundary daily closes only; partial-month cells (the current trailing month) are flagged in the chart's tooltip and excluded from the per-month statistics. The grid spans every full month from August 2010 through to the most recent completed month. ## How to read it The qualitative reading is what most monthly-returns charts surface: October has historically been positive on average, December has had several large ups and large downs, summer months are mixed. The chart's distinctive contribution is the per-month statistical table that tests whether any of these visible patterns survive a basic dispersion check. For each calendar month, the table reports sample size _n_, mean return, standard deviation, median, t-statistic against zero, and positive count. The t-critical at _n_ = 14 (df = 13) for a two-tailed test at _p_ < 0.05 is 2.16; for _n_ = 15 (df = 14) it is 2.14. | Month | Mean | Stdev | t-stat | Significant at p<0.05? | | ----- | ----- | ----- | ------ | ---------------------- | | Jan | ~+5% | ~22% | < 1 | No | | Feb | ~+8% | ~18% | ~1.6 | No | | Mar | ~+6% | ~16% | ~1.4 | No | | Apr | ~+15% | ~24% | ~2.4 | Yes (marginal) | | May | ~+5% | ~20% | < 1 | No | | Jun | ~−2% | ~14% | < 1 | No | | Jul | ~+8% | ~18% | ~1.7 | No | | Aug | ~−2% | ~16% | < 1 | No | | Sep | ~+1% | ~16% | < 1 | No | | Oct | ~+18% | ~30% | ~2.3 | Yes (marginal) | | Nov | ~+12% | ~28% | ~1.6 | No | | Dec | ~+5% | ~22% | < 1 | No | (Numbers are illustrative bands; the live page recomputes on the actual data.) ## The seasonality skeptic's read The headline finding is that **almost every month's t-statistic fails the conventional two-tailed 95% significance threshold**. October's reputation as "Uptober" rests on a small-sample mean that does not robustly survive the dispersion test. April reads similarly. Both months produce positive averages, but the standard deviation across years is wide enough that the average is not statistically distinguishable from zero at any _n_ < 30. The same problem applies to most equity-market seasonality claims. Sullivan, Timmermann, and White (2001) — _Dangers of data mining: the case of calendar effects in stock returns_ — found that the apparent equity calendar effects in their study did not survive proper data-snooping correction at much larger sample sizes than Bitcoin's. Bouman and Jacobsen (2002) — _The Halloween indicator, "Sell in May and go away"_ — survives some such tests in equity data; the equivalent test on Bitcoin's much smaller sample produces a marginal _t_-stat that is not robust to sample-period choice. The honest reading: Bitcoin's monthly grid records what has happened, with substantial dispersion year to year. Seasonality narratives ("Uptober", "summer doldrums", "Santa rally") are stories told over a small sample with high variance; they are not statistically robust patterns. ## When it fails Small sample size is the structural problem. Each calendar month has at most 16 observations (2010 through 2025); after dropping outlier years that include known structural events (Mt Gox collapse 2014, COVID flush March 2020), the effective sample shrinks further. No conventional statistical test can produce strong claims at that _n_. Survivorship and data-period bias matter. The 2017–2021 stretch was the most internet-visible portion of Bitcoin's history; "Uptober" gained currency during a window when October was indeed mostly positive. Including the post-2021 stretch dilutes that pattern; including pre-2014 data (which has its own structural issues) dilutes it further. Cross-asset comparison. Equity-market seasonality has been studied for decades at much larger sample sizes; even there, after data-snooping correction, most named effects do not survive. Bitcoin's _n_ is small enough that the priors should be even more sceptical. ## Frequently asked **What is the best month for Bitcoin?** October and April have the largest mean returns historically, but neither survives a conventional statistical test of dispersion against zero. **Is "Uptober" a real pattern?** Statistically, no — the average is positive over the available history, but the standard deviation across years is too wide for the mean to be distinguished from zero at conventional significance levels. **How many months are in the data?** Roughly 180 monthly observations across 15 years. Per-month sample size ranges from 14 to 16. **Should I trade Bitcoin based on the monthly returns heatmap?** No — the dispersion is too wide and the sample too small to support reliable seasonal trades. Treat the grid as descriptive history, not as forward signal. --- # Hash Rate URL: https://btcoak.com/hash-rate Category: mining ## What it is Bitcoin hash rate is the aggregate computational throughput of the network's mining hardware, measured in hashes per second. The current scale is exahashes per second — one EH/s equals 10¹⁸ hashes per second, or one quintillion. Hash rate is not measured directly; it is estimated from the protocol's difficulty target and the realised block-interval distribution, then smoothed over a daily or seven-day window. Today's network hashes roughly 10⁵× faster than it did in 2011. ## How it is calculated Two facts about hash rate need stating up front. First, hash rate is not measured. There is no oracle that polls every miner and sums their throughput. Second, every published hash-rate number is therefore a _model_, fitted to two observable quantities: the protocol's current difficulty target and the time interval between blocks the network actually produces. Ben Celermajer captured the principle as "in a distributed process like mining, it is near impossible to obtain reliable hash rate figures from the universe of miners." The standard derivation, documented on the Bitcoin wiki, is: `hashrate ≈ D · 2³² / 600` where _D_ is the network's current difficulty target and 600 is the protocol's ten-minute target block interval in seconds. Daily hash rate is backed out from those two quantities, then smoothed into a continuous EH/s series — one gigahash per second is 10⁻⁹ EH/s. Days where converted hash rate sits below one terahash per second are dropped to keep the log axis legible. Short-window readings carry meaningful variance from the Poisson statistics of block discovery alone. Most reference dashboards smooth seven days for that reason; this page surfaces the daily series and the 30-day rate of change. ## How to read it Two readings carry the chart. The first is the absolute level — today's EH/s versus the all-time high — for context on where the network sits in its long-run growth curve. The second is the 30-day rate of change, which surfaces the local regime: rebuilding hard, growing steadily, or capitulating. | Reading | Regime | What it has meant | | ----------------- | --------------- | ------------------------------------------------------------------------------------------------------------------------------- | | ≤ −10% / 30d | Capitulation | Inefficient hardware retiring at scale. The June 2021 China shutdown and the August 2024 post-halving slow-bleed printed here. | | −10% to −2% / 30d | Decline | Margin compression bleeds older rigs offline. Most multi-week post-halving and post-difficulty-spike windows live in this band. | | −2% to +5% / 30d | Steady | Marginal capacity trading places. The dominant regime — Bitcoin spends most of its days in this band. | | +5% to +15% / 30d | Acceleration | New-generation ASIC rollouts and post-capitulation rebuilds. Sustained acceleration here typically precedes new ATHs. | | > +15% / 30d | Rapid expansion | Unusual; only the immediate post-China-ban rebuild has printed here. | ## Historical readings Six order-of-magnitude milestones bracket the network's history, plus the single-day all-time high. Crossings are computed inline from the daily-close hash-rate series: - 1 TH/s — CPU/GPU-era mining - 1 PH/s — first-gen ASIC rollout - 1 EH/s — industrial mining at scale - 100 EH/s — pre-pandemic baseline - 500 EH/s — post-2022-bear rebuild - 1 ZH/s — first single-day print The pre-2020 cadence stitches in coarser data, so the early crossing dates resolve to the day data is available rather than the day each threshold was first breached intraday. ## The ASIC efficiency curve The headline story of the post-2016 hash-rate climb is not energy. It is efficiency. Each generation of flagship Bitcoin ASIC has hashed more per joule than the last, and the cumulative improvement from the Antminer S9 in 2016 to the S21 Pro in 2024 is roughly six- to seven-fold. | Model | Released | Hash rate | Efficiency | | ---------------- | -------- | --------- | ---------- | | Antminer S9 | Jun 2016 | 13.5 TH/s | ~98 J/TH | | Antminer S17 Pro | Apr 2019 | 53 TH/s | 39.5 J/TH | | Antminer S19 Pro | May 2020 | 110 TH/s | 29.5 J/TH | | Antminer S19 XP | Jul 2022 | 140 TH/s | 21.5 J/TH | | Antminer S21 | Oct 2023 | 200 TH/s | 17.5 J/TH | | Antminer S21 Pro | Jul 2024 | 234 TH/s | 15 J/TH | Hash rate is a power × efficiency product. If energy spend stayed flat from 2016 to 2024, the same farms running S9s would have produced one-sixth the EH/s the same farms running S21 Pros produce today. The efficiency gain is therefore an upper bound on the share of today's ATH that is "free" hash — bought by replacing rigs rather than adding wattage. A 10× rise in hash rate over a half-cycle is not a 10× rise in real-resource commitment. Hash rate is a _biased_ security-budget proxy, and the bias has gotten stronger every cycle. ## The 2021 China-ban shock Bitcoin's network hash rate fell roughly half over two months as Chinese provinces shut down domestic mining. The pre-shock peak was ~180.7 EH/s on 14 May 2021; the trough was ~86 EH/s in early July 2021 — a ~52% drop in roughly seven weeks. Recovery to the pre-ban level reached late November to early December 2021; a fresh ATH printed shortly after. The redistribution moved Chinese ASICs to North America, Kazakhstan, and Russia. Read the 2021 collapse as an inventory-relocation event, not an economic capitulation. ## When it fails The data is modelled, not measured. The same daily series can swing several percent on lucky or unlucky block-interval runs without any change in real-world hashpower. The chart on this page presents both the daily series and the smoothed 30-day rate of change so the noise is visible alongside the signal. Hash rate lags price, not the other way around. Fantazzini and Kolodin's 2020 paper _Does the Hashrate Affect the Bitcoin Price?_ finds unidirectional Granger-causality from price to hashrate in their second sub-sample (December 2017 to February 2020); their first sub-sample (August 2016 to December 2017) shows neither direction as significant. The popular hash-rate-is-bullish narrative, which treats hash-rate ATHs as a leading indicator for price, has weak empirical footing. The headline number sits inside two big mining pools. As of late 2025, Foundry USA and AntPool together account for roughly half of the network's hash rate — Foundry runs in the 30% range, AntPool in the high-teens, with combined share fluctuating around 50–60%. The security-budget reading of an all-time-high hash rate ought to be discounted by that intermediary concentration. ## Frequently asked **What is Bitcoin hash rate?** The aggregate computational throughput of the network's mining hardware, measured in hashes per second. The current scale is exahashes per second — one EH/s equals 10¹⁸ hashes per second. **Is Bitcoin hash rate at an all-time high?** The most recent ATH lives at the top of the daily series. Hash rate refreshes nightly; the spot-price comparison in the reading row above auto-refreshes a few times a day in the browser. **How is Bitcoin hash rate calculated?** Hash rate is back-calculated from on-chain data using `hashrate ≈ difficulty × 2³² / 600`, where 600 seconds is the protocol's target block interval. In a distributed process like mining, reliable hash-rate figures cannot be obtained directly from the universe of miners. **Does hash rate predict Bitcoin price?** Academic work points the other way. Fantazzini and Kolodin's 2020 paper finds unidirectional Granger-causality from price to hashrate, not the reverse, in the relevant sub-sample. Hash rate lags price. **What happened to Bitcoin hash rate during the 2021 China ban?** It fell roughly half over two months. Pre-shock peak ~180.7 EH/s in mid-May 2021; trough ~86 EH/s in early July 2021. Recovery to the pre-ban level reached late November to early December 2021, mostly through Chinese ASICs being containerised and shipped to North America, Kazakhstan, and Russia. --- # Hash Ribbon URL: https://btcoak.com/hash-ribbon Category: mining ## What it is The Hash Ribbon plots two simple moving averages of Bitcoin's network hash rate — a 30-day line and a 60-day line — on a logarithmic right-hand axis in exahashes per second. Pale-grey shading covers every day the 30-day sits below the 60-day; in Charles Edwards's terms, that shading is the "miner capitulation" window. Accent dots mark days where the 30-day crosses back above the 60-day — the published recovery upcross. Edwards introduced the indicator in his October 2019 Capriole essay _Hash Ribbons & Bitcoin Bottoms_. ## How it is calculated Two SMAs run over the daily hash-rate series: - `ma30(t) = mean(hashrate[t−29..t])` - `ma60(t) = mean(hashrate[t−59..t])` A recovery upcross prints on day _t_ when `ma30(t−1) ≤ ma60(t−1)` and `ma30(t) > ma60(t)`. Capitulation begins when the inequality flips the other way. The piece almost every reproduction omits is Edwards's _second_ step. Edwards observed that buying on the bare hash-rate cross produced one bad print — the January 2015 false positive — and that the majority of such drawdowns "can be eliminated by simply adding a price action indicator. Such an indicator could include the famous Bitcoin 10- and 20-day SMA cross over." His full buy signal, in his own words, is then defined as "Purchasing during miner capitulation, as the Hash Rates start to 'recover' and only once price momentum has gone positive (using the 10–20 SMA cross) yields the results below." In other words: the 30/60 hash-rate cross marks the end of capitulation; the 10/20 price-SMA cross is what turns it into a buy. btc oak plots every 30/60 hash-rate upcross, which matches the dated buy-signal list circulating on Bitcoin Magazine and on Capriole's own TradingView script. The full two-step buy signal is the more disciplined read; the bare cross is what gets published. ## How to read it Three regimes resolve from the chart. The grey-shaded capitulation window is patient territory — miners who can't cover power costs are switching off, the back of the queue is being cleared, and the indicator's job is to wait for the cross. The accent dots mark that cross. The unshaded stretch above the ribbon is the "normal" regime: the indicator carries no fresh information until the next capitulation prints. | State | Reading | What it has historically meant | | ------------ | ------------------------- | -------------------------------------------------------------------------------------------------------------------- | | Capitulation | 30d < 60d | Miners under stress. Patience phase; every cycle low since 2015 has been preceded by one. | | Recovery | Fresh 30d→60d cross | Edwards's recovery upcross. Roughly two-thirds of the historical fourteen prints have closed positive at six months. | | Normal | 30d ≥ 60d, no fresh cross | Above the ribbon. The chart carries no fresh signal. | ## Historical readings Eight cycle-anchor windows surface the indicator's character. Each row pins a published Capriole-list upcross date; the price at signal and the 180-day forward return are computed inline from the daily-close series. Cycle anchors: - 2019-05-23 — 2018–19 bear-end recovery - 2020-06-23 — Post-COVID recovery - 2020-12-02 — 2020 cycle ramp - 2021-08-07 — Post-China-ban rebuild (fired despite the geographic-relocation cause) - 2022-07-04 — 2022 LUNA-crash recovery - 2023-01-14 — 2023 banking-crisis lead-in - 2024-06-17 — 2024 post-halving capitulation - 2025-12-01 — 2025 late-cycle leg ## The 2021 China-ban edge case The cleanest example of the ribbon firing for a non-canonical reason printed in mid-2021. Bitcoin's network hash rate fell roughly 50% in two months from the May 2021 peak as Chinese provinces — chiefly Inner Mongolia, then Sichuan — pushed local mining operations to shut down. The 30-day average collapsed under the 60-day, capitulation shading filled the chart, and a recovery upcross duly printed on 2021-08-07. Bitcoin then ran to roughly $69k over the following three months. The signal worked. The cause did not match the indicator's pedagogical setup. The Chinese hash drop was not unprofitable miners switching off — it was profitable miners boxing up rigs and shipping them to Texas, Kazakhstan, and Russia. The canonical mining-map record shows China's network share falling from roughly 46% in April 2021 to effectively zero by July, then climbing back over 20% by year-end as underground operations restarted. The ribbon measured a hash-rate dip; the dip measured ASIC logistics, not miner economics. The honest framing is that the Hash Ribbon detects something correlated with cycle lows — reduced realized network throughput, broadly — without requiring the cause to be miner capitulation in the strict economic sense. That is a feature for a cycle-bottom hunter and a bug for anyone reading the indicator as a pure miner-stress proxy. ## When it fails The bare cross has explicit false positives. Edwards flagged a January 2015 buy in his own article as the indicator's worst historical print — the cross fired before the cycle low and was followed by a substantial drawdown. A December 2019 cross "produced the lowest returns recorded" for the indicator before COVID compressed any recovery. An August 2022 cross fired weeks before the FTX collapse rather than after the cycle low. Edwards's full two-step would have excluded the January 2015 print and softened the others; the bare cross plotted on this chart catches them. The cause can be non-canonical. The May–July 2021 hash-rate collapse was a geographic relocation event, not miner economics. The recovery upcross still fired and the trade still worked, but the reason had nothing to do with marginal miners going broke. Treat the indicator as a cycle-low coincidence detector, not a strict miner-capitulation reading. The data is modelled, not measured. Network hash rate is a back-out from realised difficulty and observed block intervals. A run of lucky or unlucky blocks can swing daily readings several percent without any change in real-world hashpower. The ribbon's 30-day and 60-day averages absorb most of the noise but nothing eliminates it. Critics also argue the indicator's relevance softens as institutional spot demand swamps miner-driven sell-pressure as a price input. ## Frequently asked **What is the Hash Ribbon?** Two moving averages of Bitcoin's hash rate — 30-day and 60-day — that flag when miners are under stress (30d below 60d) and when the worst of that stress is over (30d crossing back above 60d). Charles Edwards introduced the indicator in his 2019 Capriole essay. **What is the buy signal for the Hash Ribbon?** Edwards's published spec is two-step: the 30/60 hash-rate cross _plus_ a 10-day-over-20-day price-SMA confirmation. The unfiltered hash-rate cross alone is what most live charts plot, because it matches the Capriole-published dated buy-signal list. The full Edwards filter would have excluded the January 2015 false positive. **Who created the Hash Ribbon?** Charles Edwards of Capriole Investments. The canonical write-up — _Hash Ribbons & Bitcoin Bottoms_ — was published in October 2019 with the full 30/60 plus 10/20 specification. **Has the Hash Ribbon ever flashed a false signal?** Yes. January 2015 (Edwards's own example), December 2019, and August 2022 were all false positives. The reported 64% hit rate over fourteen historical signals is honest framing — the indicator works on average, not every time. --- # Active Addresses URL: https://btcoak.com/active-addresses Category: on-chain ## What it is The daily count of unique Bitcoin addresses that appeared as either input or output of a confirmed transaction. Each distinct address counts once per UTC day, even if it transacts multiple times. Address-set, not transaction-count — one address transacting twenty times still counts once. The 30-day simple moving average is the trend signal. ## How it is calculated ``` AA(t) = | { addr : addr appears as input or output of a confirmed tx on day t } | sma30(t) = mean(AA(t-29) ... AA(t)) yoy(t) = sma30(t) / sma30(t-365) - 1 ``` Coinbase issuance rows with no sender address are excluded. Addresses are the network's transaction-graph nodes, not user accounts — the divergence between the two has grown structurally since exchange custody became dominant in 2018-19 and again since spot-ETF custody became material in 2024. ## Regimes | Reading | Regime | What it has meant | | ---------------- | ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | YoY > +30% | Expansion | The 30-day SMA is materially above its level a year ago. Has bracketed the 2017 run-up, the 2020-Q4 to 2021-Q1 surge, and the 2024 post-halving early window. | | −10% to +30% YoY | Stable | The largest share of post-2017 history sits in this band. Says little on its own. | | YoY < −10% | Contracting | The 30-day SMA is materially below its level a year ago. Fits the 2018 and 2022 deep-bear stretches; can also fire in the post-Runes 2024+ window when inscription activity drops out of the trailing year. | ## The post-2023 decoupling Through 2022 the chart read as a clean adoption proxy — address counts rose with cycles, fell with bears. Post-2023 it stopped doing that cleanly. The cycle-peak table walks the maximum 30-day SMA in the window around each cycle top and computes the implied USD-per-active-address ratio. The 2017 cycle peak SMA, the 2021 dual peaks, and the 2024 pre-halving window all printed in a tight band of roughly 900k to 1.2M addresses per day — despite price tripling between cycles. USD-per-active-address has risen by roughly an order of magnitude across cycles. The 2017 reading printed in the high teens; the 2021 readings in the mid-to-high fifties; the 2024 reading in the high sixties. Each cycle, the same unit of base-layer address activity has been assigned threefold more market value. The post-2023 window introduced a different distortion. Inscriptions — beginning with Casey Rodarmor's January 2023 Ordinals genesis — and Runes, the token protocol that launched at the April 2024 halving in block 840,000, both consume base-layer block weight without growing the unique-address footprint. Many inscription transactions reuse a small handful of inscriber and marketplace addresses for thousands of mints. Address counts compress while transaction counts and fees rise; the active-address chart diverges from the transaction count and fee-revenue charts in ways no prior cycle exhibited. ## When it fails **Address ≠ user.** Custodial exchange hot wallets route transactions for thousands of customers through a small handful of addresses (under-counts retail). A privacy-conscious user can rotate through a fresh address per receive (over-counts that one user). Both biases are material, and the mix shifts cycle to cycle. **Inscription protocols reuse addresses.** Ordinals (early 2023 onward) and Runes (April 2024) consume base-layer block weight at high transaction counts but with significant address reuse — a single marketplace address mints, transfers, and settles thousands of inscriptions. The active-address signal compresses while transaction count and fee revenue rise. **Layer-2 activity is invisible.** Lightning Network routing, sidechain (Liquid, Rootstock) transfers, and federated-bridge activity do not touch the base layer and do not appear in the count. Lightning capacity in the tens of thousands of BTC routed annually is real economic flow that the active-address chart cannot see by design. **The pre-2014 history is sparse.** Address-set telemetry on the first three years of the chain reflects much smaller absolute participant counts than the post-2014 era; comparisons across the boundary should weigh structural growth, not just headline-number ratios. ## Frequently asked The 2017 and 2021 cycle tops both printed sustained 30-day averages above one million. Cycle troughs have consistently dragged the average below 700k, with the 2018 and 2022 lows both bottoming near 500k. The post-Runes window (April-October 2024) compressed materially — daily counts fell well below the multi-year baseline as inscription activity absorbed block weight without growing the address set. The cleanest cross-comparison is against new addresses (first-touch only), transaction count (the throughput cap is set by block weight), and daily fees (the dollar gauge of blockspace pricing). Read together, the four resolve into the components the active-address number used to track on its own. --- # New Addresses URL: https://btcoak.com/new-addresses Category: on-chain ## What it is The daily count of Bitcoin addresses that appear on the ledger for the first time. First-touch only — an address that has ever previously appeared in any confirmed transaction is excluded, even if it transacts again. The signal measures the rate at which the address set itself is growing, which is a cleaner on-boarding proxy than total active-address count. ## How it is calculated ``` NA(t) = | { addr : addr appears on chain for the first time on day t } | sma30(t) = mean(NA(t-29) ... NA(t)) sma90(t) = mean(NA(t-89) ... NA(t)) yoy(t) = sma90(t) / sma90(t-365) - 1 ``` The first daily print of an address counts toward NA on that day; every subsequent transaction contributes to active addresses but not to NA. The 90-day window is the primary trend smoother because the raw daily series has substantial weekly periodicity from exchange-side wallet rotations. ## Regimes | Reading | Regime | What it has meant | | ---------------------- | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | YoY > +30% | Accelerating | On-boarding pace materially faster than a year ago. Has bracketed run-ups into the 2017 cycle top, the 2021 spring rally, and the 2024 post-halving early window. | | −15% to +30% YoY | Normal | The largest share of post-2018 history sits here. | | YoY < −15% w/ price up | Adoption divergence | On-boarding is well below a year ago even as price rises. A late-cycle marker — capital rotating inside the existing address set rather than coming from new participants. | ## The cycle-low lead The 30-day SMA of new addresses has bottomed before each Bitcoin cycle price low on the daily-close record. The cycle-lead table walks each major price low (2015-01-14, 2018-12-15, 2020-03-12 Covid flush, 2022-11-21 post-FTX), finds the local minimum of the 30-day SMA in the surrounding window, and reports the lead in days. The lead has averaged two to four months across the four cycles in the table, with no cycle where the SMA bottomed _after_ the price. The signal is not actionable on a daily horizon — the SMA bottom date itself is only knowable retrospectively, after enough subsequent prints have confirmed the trough — but it is a reproducible regime confirmation that the on-boarding floor has been set. The magnitudes are diverging from prior cycles. The 2015 SMA bottom printed in the low six figures; the 2018 bottom in the mid six figures; the 2022 bottom near 200k. Each cycle's on-boarding floor has been higher in absolute terms but lower as a fraction of the prior cycle's peak. The addressable-bound side of this argument is the same one driving the active-address decoupling — block weight is fixed, transaction shape is roughly fixed, address counts saturate around what each block can contain. ## When it fails **A new address is not always a new user.** HD-wallet software rotates a fresh receive address per incoming transaction by default; an exchange mints a fresh deposit address per customer; a custodian rotates addresses for operational hygiene. Each pattern inflates the count without adding a participant. The inflation is structural and persistent. **Inscription protocols spike the count.** Ordinals inscription activity (early 2023 onward) and the Runes launch (April 2024) generate single-use receive addresses for each mint. The 90-day SMA absorbs most of the spike but not all of it; the signal during inscription-active windows over-states first-time human on-boarding by a meaningful margin. **Layer-2 first-touch is invisible.** Lightning Network channel opens are base-layer transactions and do print as new addresses; subsequent Lightning routing inside a channel does not. Sidechain (Liquid, Rootstock) activity does not appear here at all. As Layer-2 share grows, base-layer first-touch under-represents true network on-boarding by a widening gap. **The pre-2014 history is sparse.** Daily first-touch counts in the 2010–2013 era reflect a much smaller absolute base; cross-era comparisons should weight the structural growth-rate side, not just the headline-number ratio. The cycle-lead table starts in 2014 for that reason. ## Frequently asked Active addresses counts every address that transacted on the day; new addresses counts only the first-touch appearance. The two diverge most cleanly during inscription waves (active rises while new flattens, because marketplace addresses re-use heavily) and during deep bears (both fall, but new tends to bottom slightly earlier). Inscription protocols generate single-use receive addresses for many mints. Casey Rodarmor's Ordinals genesis in January 2023 produced the first wave; the Runes launch at the April 2024 halving block 840,000 produced the largest single-day spike on record. Those spikes are real protocol activity, but they over-count first-time-user on-boarding by an order of magnitude during the spike itself. --- # Bitcoin Correlations URL: https://btcoak.com/correlations Category: on-chain ## What it is Bitcoin's rolling 90-day Pearson correlation with five reference markets — gold (GLD), the Nasdaq-100 (QQQ), the S&P 500 (SPY), small-caps (IWM), and 20+ Year Treasury bonds (TLT). Five lines, five macro stories, one chart. The signal answers the digital-gold-versus-tech-beta question in dated form, one regime at a time, and surfaces every regime crossover on the record. ## How it is calculated ``` ρ_90d(BTC, X)_t = Cov(r_BTC, r_X) / (σ_BTC · σ_X) ``` where `r` denotes daily log returns (`ln(close / close_-1)`) and the covariance and standard deviations are computed over the trailing 90 aligned trading days. Equity returns are New-York-close-to-New-York-close; Bitcoin's 24/7 returns are folded onto the equity calendar so the Friday-to-Monday bar carries the weekend price action. The 90-day window balances responsiveness against noise — shorter windows whip on every macro surprise, longer windows lag too much to read regime turns. The five reference markets are exchange-traded funds rather than the underlying indices: GLD (SPDR Gold Shares), IWM (iShares Russell 2000), QQQ (Invesco QQQ Trust), SPY (SPDR S&P 500), and TLT (iShares 20+ Year Treasury Bond). ETFs are used so the dividend-and-distribution-adjusted total-return matches what an investable benchmark would actually deliver. ## Regimes | Reading | Regime | What it has meant | | ------------------- | --------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Top \|ρ\| < 0.30 | Idiosyncratic | No reference asset dominates. Bitcoin is responding to its own drivers — halving cycle, on-chain regime, narrative shifts. | | Top is GLD | Gold-linked | The digital-gold thesis is live in the data. Rare regime. Has fired around the 2023 US regional-banking stress, brief windows of dollar weakness, and the months around the spot-ETF launch in early 2024. | | Top is SPY/QQQ/IWM | Risk-on linked | Equity-coupled. Bitcoin is trading on growth-equity beta — discount rates, earnings revisions, risk appetite. The default regime through most of 2022 and large stretches of 2024–25. | | Top is TLT | Risk-off linked | Long-duration coupling. The framing where Bitcoin trades as a duration trade rather than a hedge. | | Top correlate flips | Mixed | No single asset dominates persistently. | ## The crossover history The single observation most competitor pages omit: Bitcoin's gold correlation and Nasdaq correlation cross repeatedly. Each crossover is a regime flip from "digital gold" to "tech beta" or back. The crossover table walks the full daily-close history and emits a row for every regime that lasted at least thirty days. Across the post-2010 record, Bitcoin has spent roughly a fifth to a third of its trading days in a gold-led regime (BTC-GLD > BTC-QQQ) and the balance in a Nasdaq-led regime. Nasdaq-led runs dominate the time-in-regime budget — especially since mid-2020. The Covid-era liquidity flush, the 2022 Fed hiking cycle, and the 2024–25 AI-led rally all produced multi-month windows where the QQQ coefficient sat above +0.5 and the GLD coefficient was a sideshow. Lyn Alden's 2024 essay on Bitcoin as a global liquidity barometer argues the deeper driver is global money-supply growth: tech equities and Bitcoin both depend on cheap discount rates, and the equity-correlation is downstream of that shared sensitivity rather than the cause. The framing flips faster than narrative allows. Arthur Hayes frames Bitcoin as a global-fiat-liquidity barometer; the canonical institutional _Bitcoin First_ research note positions it as an uncorrelated portfolio diversifier; Peter Schiff has argued Bitcoin is "negatively correlated" with gold in his debates with Saifedean Ammous. Each framing is a window in this data, not a constant. An institutional research note on the factors driving gold and Bitcoin documents the long-run baseline plainly: across the post-2015 record, the 90-day BTC-GLD correlation has averaged about +0.10, with a peak of +0.57 and a trough of −0.37. The digital-gold thesis is a regime that fires, not a structural state. ## When it fails **Pearson is symmetric and direction-blind.** A high BTC–QQQ reading tells you the two co-moved; whether tech dragged Bitcoin or Bitcoin dragged tech is a question this coefficient cannot answer. For causality, the appropriate tool is a Granger-causality test or a vector autoregression, not rolling Pearson. **Rolling-window correlations capture distribution overlap, not tail coupling.** The 2020 Covid liquidity flush is the canonical case: Bitcoin and equities both fell sharply in March 2020 in a two-week window, but the 90-day coefficient barely moved because the bulk of the trailing window was pre-shock. **The reference universe is selective.** The five ETFs cover gold, US equities (broad, tech, small-cap), and long bonds. They do not cover the dollar index, oil, EM equities, the yen carry, or credit spreads — each of which has a real claim on Bitcoin's correlation surface in specific regimes. **Calendar-alignment loses weekend information.** Bitcoin trades 24/7; equity ETFs do not. Folding the 24/7 returns onto Monday-to-Friday closes parks weekend price action inside the Friday-to-Monday bar. Major weekend moves hit the chart as a single elevated Monday return rather than three separate prints. ## Frequently asked Both — sometimes. Bitcoin's 90-day correlation with the Nasdaq-100 and with gold can be high simultaneously, especially during macro liquidity-driven windows. Bitcoin spent a meaningful share of its daily-close history in a gold-led regime, and the balance in a Nasdaq-led regime. The framing is regime-dependent, not structural; position-sizing on a permanent +0.5 Nasdaq correlation, or on a permanent zero, will both be wrong inside two years. Larry Fink described Bitcoin in October 2024 as "an alternative to other commodities like gold," in the same year that the largest spot Bitcoin ETF launched. The data places that statement inside a real but bounded regime — not a permanent reclassification. --- # Daily Transaction Count URL: https://btcoak.com/transactions Category: stats ## What it is Daily Transaction Count plots the number of confirmed on-chain Bitcoin transactions per day, with a 30-day moving average overlaid for trend. The chart's value has shifted twice over its history — first by Lightning Network and exchange batching siphoning payment volume off the base layer, then by inscription protocols (Ordinals, BRC-20, Runes) re-inflating the count from a different mechanism. Today it is best read as a blockspace-consumption gauge, not an adoption proxy. ## How it is calculated ``` count(d) = number of confirmed transactions in blocks mined that day sma30(d) = mean(count(d-29) … count(d)) yoy(d) = sma30(d) / sma30(d-365) - 1 ``` The headline regime bucket comes from `yoy` on the 30-day SMA: **Expansion** above +30%, **Contracting** below −10%, **Stable** in between. The 30-day window damps the day-to-day noise from inscription waves; the YoY frame puts each reading in cycle context. ## How to read it Three useful lenses, in increasing order of analytic value. **Level reading** — today's 30-day average against the 250–500k payment-era band. **YoY reading** — the headline regime bucket, useful as a momentum filter. **Cycle-aligned reading** — comparing each cycle anchor's count against price at that anchor, where the decoupling story shows up unambiguously. | Reading | Regime | What it has meant | | ---------------- | ----------- | ---------------------------------------------------------------------- | | > +30% YoY | Expansion | Bracketed every cycle run-up since 2017 and the 2023 inscription wave. | | −10% to +30% YoY | Stable | The largest share of post-2017 history; says little on its own. | | < −10% YoY | Contracting | Fits the 2018 and 2022 deep-bear stretches. | ## The count-vs-price decoupling The single most important fact this chart records is that base-layer transaction count and price have **decoupled**. The 2013 and 2017 cycle tops printed record transaction counts alongside record prices — the count was a reasonable adoption proxy. The 2021 top did not. The 2024–26 cycle has not. | Anchor | Price | Count | Note | | ----------------------- | -------- | -------- | --------------------------------- | | 2017-12-17 (top) | ~$19,140 | ~392,000 | Pre-Lightning, pre-batching | | 2021-11-10 (top) | ~$69,000 | ~301,000 | Post-Lightning, post-batching | | 2024-04-23 (Runes peak) | ~$66,600 | ~927,000 | Inscription-driven, off the curve | Two structural shifts explain the 2017→2021 decline despite price growth. First, the **Lightning Network** mainnet launched in March 2018 and routed an unknown but non-trivial share of payment volume off-chain. Second, exchange-internal transfers and large-custody **batching** consolidated what used to be many user-visible movements into a single coinbase-to-coinbase or hot-wallet-to-cold-wallet transaction. Then 2023 and 2024 reversed direction for a different reason. **Ordinals** activity in early 2023, the **BRC-20** token wave starting in March 2023, and the **Runes** launch at the April 2024 halving each pushed the count above its prior payment-only ceiling. But the additional transactions are protocol metadata and small token mints, not user wallet payments. The chart no longer reads as a single coherent "adoption" gauge — it reads as a composite of payments + inscriptions + batching effects. ## When it fails **Lightning and L2 are invisible.** Lightning routing, sidechain (Liquid, Rootstock) activity, and federated-bridge transfers do not hit the base layer. Public estimates put Lightning capacity in the tens of thousands of BTC routed per year; none of that is reflected here. **Inscription protocols inflate the count.** Ordinals, BRC-20, and Runes each generate large numbers of small base-layer transactions during activity bursts. Real consensus-layer events, but not the wallet-payment activity the count historically tracked. **Batching consolidates many actions into one row.** Exchange sweeps, large custody re-balances, and CoinJoin coordination can collapse what a user-side observer would call "dozens of transfers" into a single transaction with many inputs and outputs. **The early-history resolution is sparser.** Pre-2020 the source data resolution drops to a four-day cadence; the chart preserves whatever the source provides, so SMA windows are effectively wider during that era. ## Frequently asked **How many Bitcoin transactions are processed per day?** As of 2026, the network confirms in the 400,000–700,000 range depending on whether inscription activity is firing. The base-layer ceiling for ordinary transactions sits near 500,000 per day. **Why does Bitcoin's transaction count saturate?** Block weight is capped at four million weight units per block, equivalent to roughly 1.5–3.5 MB of transaction data. With 144 blocks per day, that bounds throughput before the mempool starts to back up. **Why are 2021 transactions lower than 2017?** Lightning Network mainnet launched in March 2018; exchange-internal batching consolidated user activity into single transactions. The count is no longer a clean adoption proxy. **Are Lightning Network transactions counted here?** No. Base-layer only. --- # Daily Fees URL: https://btcoak.com/fees Category: stats ## What it is Daily Fees plots the network's aggregate fee revenue in USD per day on a logarithmic axis spanning $100 to $100M. The 7-day moving average is the trend line; the fee share of miner revenue alongside is the long-run security-budget gauge. Bitcoin's post-2140 economy lives entirely on this leg; the chart is the live indicator for whether fee revenue can carry the security budget once the subsidy reaches zero. ## How it is calculated ``` fees(d) = sum of (inputs - outputs) across all confirmed transactions in d subsidy_usd(d) = reward(d) × 144 × price(d) share(d) = fees(d) / (fees(d) + subsidy_usd(d)) sma7(d) = mean(fees(d-6) … fees(d)) ``` The fee surplus is the protocol-level definition: every transaction's inputs minus outputs, summed over the block, paid to the miner alongside the block subsidy. The share denominator uses the protocol's halving schedule for the subsidy in BTC and the day's closing price for the USD valuation. The page-level reading row recomputes the share against live spot so the headline ticks with the BTC price through the day. ## How to read it Fee data is too volatile to read off the latest day. A useful regime read needs both a level (the 7-day SMA) and a context (the share of miner revenue). Four named regimes cover the post-2017 history. | Reading | Regime | What it has meant | | --------------- | --------- | --------------------------------------------------------------------------------------------------------------------------------- | | < $1M & < 5% | Quiescent | Mempool cleared most blocks. Fits 2018–19 and post-2024-event lulls. | | $1–5M & 5–15% | Active | Default steady-state in the late-2020s. Some blocks fill; competition is real but not severe. | | $5–25M & 15–35% | Congested | Sustained mempool backlog. Fits late-2017, 2021 spring, 2023 Ordinals waves. | | > $25M or > 35% | Eruption | Fees within an order of magnitude of subsidy. Fits December 2017, the 2023 Ordinals second wave, and the April 2024 Runes launch. | ## When fees beat the subsidy The single rarest event in this chart is a day on which fees exceeded the block subsidy — share above 50%. It has happened only a handful of times on record, all clustered around the April 2024 halving. The single highest fee-share day is **20 April 2024**, the day block 840,000 was mined and the Runes protocol launched, when fees came in at roughly **$81 million** against a subsidy of roughly $29M — a fee share of **73.8%**. For one day the post-2140 economy was a working preview. The pattern matters because Bitcoin's long-run security model rests on that share rising structurally, not just spiking on event days. A subsidy halving every four years means the absolute USD figure of the subsidy compresses unless price is rising; if price stagnates, fees are the only remaining miner-revenue lever. The April 2024 cluster is the first time the chart has stress-tested that thesis at non-trivial scale. ## When it fails **USD framing mixes two volatilities.** Daily fees in dollars reflect both sats-per-vByte fee pressure and BTC price. A flat fee market with a rising price prints a rising USD-fee line even if no more transactions are being sent. The fee share of revenue partially neutralises this. **Single-day spikes are misleading.** A coordinated wallet movement, an inscription mint, or a low-fee mining-pool block can move the single-day total by a factor of three. Use the 7-day SMA for regime calls. **Fee revenue is one half of miner economics.** The other half is the block subsidy, which steps down at every halving and rounds to zero around 2140. The fee share rises mechanically as the denominator falls; reading the share trend without that context will overstate "structural" fee growth. **Pre-2020 resolution is sparser.** Source data drops to a four-day cadence pre-2020; SMA windows during that era are effectively wider than their names. ## Frequently asked **How much do Bitcoin transaction fees cost per day?** As of 2026, the network pays in the $100k–$5M range in aggregate on calm days, with single-day bursts up to $30M+ during inscription waves. The all-time-high single-day fee total is $81M on 20 April 2024. **What share of miner revenue comes from fees?** As of 2026, the share is in the 1–5% range in steady state. The historical range across regimes spans from < 1% in calm windows to 73.8% on the Runes launch day. **Why are Bitcoin fees so volatile?** Block weight is fixed; demand is not. When demand exceeds capacity, fees bid up sharply rather than linearly because users compete in a sealed-bid auction for inclusion. **When have Bitcoin fees exceeded the block subsidy?** On a handful of days, all clustered around the April 2024 halving. The single highest fee-share day is 20 April 2024 at 73.8%. **How are Bitcoin fees calculated?** Each transaction's inputs minus outputs is the fee, paid to the miner who confirms the transaction. The chart sums across every confirmed transaction in the day's 144 blocks. --- # NVT Ratio & Signal URL: https://btcoak.com/nvt Category: stats ## What it is NVT — Network Value to Transactions — is Bitcoin's analogue of a price-to-earnings ratio: market capitalisation divided by daily on-chain transaction volume in USD. High NVT means the market pays a lot for each dollar of utility moving over the network in a day; low NVT means the opposite. The chart plots two lines: Willy Woo's raw NVT (same-day denominator) and Dmitry Kalichkin's smoothed NVT Signal (90-day SMA denominator). ## How it is calculated ``` market_cap(t) = price(t) × circulating_supply(t) NVT(t) = market_cap(t) / tx_volume_usd(t) NVT_Signal(t) = market_cap(t) / SMA90(tx_volume_usd)(t) ``` The market-cap numerator uses the protocol's halving schedule for circulating supply — the same series that drives the daily issuance chart. The denominator is the network's estimated on-chain transaction volume in USD. Willy Woo introduced raw NVT in early 2017. Dmitry Kalichkin published _Rethinking NVT Ratio: introducing the NVT Signal_ on **4 February 2018**, substituting a 90-day moving average of on-chain volume for the same-day denominator. The smoothing serves two purposes: it suppresses single-day volume bursts that would otherwise drag the ratio down, and it gives the ratio a memory — the denominator reflects the last quarter of network activity, not just today's blocks. ## How to read it | Reading | Regime | What it has meant | | --------- | ----------- | ------------------------------------------------------------------------------------- | | < 100 | Undervalued | Bottom quintile. Has bracketed deep cycle lows in 2014–15 and 2018–19, parts of 2022. | | 100 – 220 | Normal | Middle 60%. The bulk of trading days live here; says little on its own. | | > 220 | Overvalued | Top quintile. Has bracketed every cycle top — 2013, 2017, 2021. | NVT Signal is a regime gauge, not a turning-point indicator. The 90-day SMA introduces a quarter-cycle of lag at regime turns; signal at the actual cycle low typically arrives two to three months after price has already turned. Use it for position sizing and late-cycle euphoria sanity-checks, not for entry timing. ## 100/220 versus Kalichkin's 45/150 The most important thing to understand about this chart is that the threshold numbers in our regime bands — 100 for Undervalued, 220 for Overvalued — do not match Kalichkin's canonical 45/150 from the original paper. The reason is the denominator, not the framework. Kalichkin computed NVT Signal against an _adjusted_ on-chain volume: his denominator nets out change outputs, common exchange self-transfers, and a heuristic correction for low-information mixer activity. The adjusted volume runs roughly 50–60% of the raw on-chain figure. Our denominator is the raw network estimated on-chain volume — well-defined, reproducible, and comparable across the entire history at daily resolution, but roughly **two times higher** than Kalichkin's adjusted volume by construction. A denominator that is twice as large produces an NVT Signal that is roughly half as large. Applied directly, Kalichkin's 45/150 thresholds on a 2×-bigger denominator would mark almost every post-2017 reading as Overvalued and lose the regime distinction entirely. We recalibrated to **100** and **220** to preserve Kalichkin's **20/60/20 percentile split** — bottom quintile, middle 60%, top quintile — on our series. The labels retain the analytical force Kalichkin assigned them; only the absolute numbers differ. If you compare our reading to one on a different framework and the numbers differ by a factor of two, the framework is the explanation, not a data error. ## When it fails **The denominator drifts with Layer-2 migration.** NVT counts base-layer on-chain volume only. As Lightning, Liquid, Rootstock, federated bridges, and large-custody batching absorb more payment activity, the denominator shrinks for reasons that have nothing to do with utility decline. A structurally rising NVT Signal in calm regimes is consistent with both "market overvalued" and "denominator under-counted" — the chart does not distinguish them. **Inscription waves inflate the denominator.** Ordinals and Runes activity packs many small data-carrying transactions into blocks. NVT Signal during the 2023–24 inscription periods reads lower than payment-only utility would suggest. **Threshold calibration is dataset-specific.** The 100/220 bands are tuned to our raw on-chain volume series. They do not transfer directly to adjusted-volume frameworks; do not compare absolute NVT Signal levels across providers without translating thresholds. **Cross-cycle decay is a real finding.** The 2022 cycle low printed an NVT Signal in the Normal band, breaking a pattern from the 2014–15 and 2018–19 lows. Cycles are getting smaller in ratio space across multiple metrics, not just NVT. ## Frequently asked **What is Bitcoin NVT?** Market cap divided by daily on-chain transaction volume in USD. Bitcoin's analogue of a P/E ratio. Introduced by Willy Woo in early 2017. **What is NVT Signal and how is it different?** NVT Signal substitutes a 90-day moving average of transaction volume for the same-day denominator. Less reactive to single-day volume bursts; reads regime shifts more cleanly. Introduced by Dmitry Kalichkin on 4 February 2018. **What does NVT Signal say about Bitcoin's price?** Above 220 has bracketed every cycle top. Below 100 has marked the deepest accumulation windows, though not every final low (the 2022 post-FTX bottom held in the Normal band). **Why are the thresholds 100 and 220, not Kalichkin's 45 and 150?** Our raw on-chain volume runs ~2× Kalichkin's adjusted volume. We recalibrated to preserve his 20/60/20 percentile split on our series. **Is NVT a good buy signal?** A poor entry trigger and a useful frame. Good for position sizing; weak as a market-timer. Pair with MVRV and realized price for cross-checks. --- # Bitcoin Total Supply URL: https://btcoak.com/total-supply Category: reference ## What it is Bitcoin's supply schedule is enforced by a single consensus rule: every block pays a per-block subsidy that halves every 210,000 blocks (≈ four years). The schedule runs for 33 non-zero halving epochs and then terminates with a per-block reward that rounds to zero satoshis. The asymptote is `Σ_{n=0..32} (210,000 × floor(50e8 / 2ⁿ)) sats = 2,099,999,997,690,000 sats = 20,999,999.9769 BTC` — the cap is hard, the round-number "21 million" is the rounding. ## How the math works The consensus rule is `nSubsidy = 50 BTC; nSubsidy >>= floor(height / 210000); return nSubsidy`. Every 210,000 blocks the right-shift effectively halves the reward; after 33 such halvings the per-block reward in satoshis rounds to zero and the schedule terminates. BIP-0042 (Pieter Wuille, 2014) formalised this and added a guard against over-shifting the reward past the 64-halving boundary. The cap was already enforced in code; the BIP made it spec. The four-year cadence is shorthand for ten-minute blocks × 210,000. The protocol re-targets difficulty every 2,016 blocks to hold the average interval near ten minutes, but the realised average has been closer to nine and a half. That undershoot is why every halving so far has arrived ahead of its four-year-from-prior date. ## Live circulating supply Around 19.85 million BTC are in circulation as of 2026-04-27 — about 94.5% of the protocol's asymptote of 20,999,999.9769 BTC. Roughly 1.15 million BTC remain to be issued, with the schedule tapering to a final non-zero subsidy of one satoshi per block in the late 2030s. ## Lost coins Issued isn't the same as accessible. Independent estimates of permanently-lost coins range from 2.78 to 3.79 million BTC — methodology varies and the upper end ages quickly. The Patoshi-era cluster work (Sergio Demian Lerner, 2013) attributes ~1.1 million BTC to early mining wallets that have never moved; dormancy heuristics on later coins push the estimate higher. Long-dormant isn't the same as lost: cold-storage coins from 2013–2016 vintages have been moved in 2024–2026 windows, and any one estimate ages quickly. ## Circulating vs total vs max These three terms mean different things across publishers, and Bitcoin's consensus rule has no burn mechanism, so: - **Circulating** is everything ever mined. - **Total** equals circulating in the BIP-0042 model — there's no burn-and-decrease pathway. - **Max** is the asymptote: 20,999,999.9769 BTC (or 21,000,000 BTC if you round). ## Common questions - Is the cap really 21 million? It converges to 20,999,999.9769 BTC, not exactly 21 million. The shortfall is the cumulative effect of integer-truncating right shifts inside the consensus rule. - When will all bitcoins be mined? The 33rd halving rounds the per-block reward to zero. Daily issuance reaches zero a few halvings before calendar 2140; the asymptote is approached, not crossed. - How many bitcoins are lost forever? Independent estimates range from 2.78–3.79 million BTC. This is a range, not a point figure. - What happens after all bitcoins are mined? Miners are paid only by transaction fees. The fee market is already real (see [/fees](/fees) for the daily series); whether fees alone are sufficient to secure the network at zero subsidy is the open empirical question. --- # Bitcoin Halving Dates URL: https://btcoak.com/halving-dates Category: reference ## What it is Bitcoin's halving is a protocol-level event in which the per-block subsidy paid to miners is cut in half. It happens every 210,000 blocks — roughly four years — and the schedule is enforced in the consensus rule that pays each block. The schedule runs for 33 non-zero halving epochs and converges on 20,999,999.9769 BTC. ## How the schedule works The consensus rule is `nSubsidy = 50 BTC; nSubsidy >>= floor(height / 210000); return nSubsidy`. Every 210,000 blocks the right-shift effectively halves the reward; after 33 halvings the per-block reward in satoshis rounds to zero and the schedule terminates. BIP-0042 (Pieter Wuille, 2014) formalised this and added a guard against over-shifting the reward past the 64-halving boundary. The four-year cadence is shorthand for ten-minute blocks × 210,000. The protocol re-targets difficulty every 2,016 blocks to hold the average near ten minutes, but the realised average has been closer to nine and a half. Every halving so far has arrived ahead of its four-year-from-prior date. ## Every halving on record | H# | Date | Block | Reward (before → after) | Price on day | | --- | ---------- | ------- | ----------------------- | ------------ | | H1 | 2012-11-28 | 210,000 | 50 → 25 BTC | $12.35 | | H2 | 2016-07-09 | 420,000 | 25 → 12.5 BTC | $657.61 | | H3 | 2020-05-11 | 630,000 | 12.5 → 6.25 BTC | $8,601.80 | | H4 | 2024-04-19 | 840,000 | 6.25 → 3.125 BTC | $63,512.75 | Projected: | H# | Date (projected) | Block | Reward (after) | | --- | ---------------- | --------- | -------------- | | H5 | ≈ 2028-04 | 1,050,000 | 1.5625 BTC | | H6 | ≈ 2032-04 | 1,260,000 | 0.78125 BTC | | H7 | ≈ 2036-04 | 1,470,000 | 0.390625 BTC | | H8 | ≈ 2040-04 | 1,680,000 | 0.1953125 BTC | Projection method: running average block interval since genesis, applied forward from the most recent halving block. Two- to four-week earlier than four-years-from-prior because realised intervals run below the ten-minute target. ## Forward-12-month return at each halving The first three halvings were each followed within 12–18 months by a new all-time high. The H4 cycle landed a softer forward return; several analysts now argue institutional flow has overtaken halving mechanics as the dominant price driver. The historical rhythm is documented; the forward extrapolation is the open question. H1 (2012): forward-12m ≈ +8,749% (massive — first cycle, thin starting price). H2 (2016): forward-12m ≈ +285%. H3 (2020): forward-12m ≈ +559%. H4 (2024): forward-12m positive but a fraction of the prior cycles' magnitudes. ## When the framing fails Matt Hougan argued in late 2025 that "the bitcoin halving is by definition half as important as it was four years ago" because the issuance reduction relative to circulating supply shrinks each cycle. Lyn Alden has framed the same point as a liquidity-cycle reframe of the four-year mantra: post-2020 cycles are increasingly synchronised with global liquidity rather than the halving alone. ## Common questions - When was the most recent Bitcoin halving? 2024-04-19 at block 840,000; the per-block subsidy dropped to 3.125 BTC. - When is the next Bitcoin halving? Around April 2028. The exact date is block-anchored, not date-anchored — see [/halving-countdown](/halving-countdown) for the live block-driven projection. - Why does the projection date differ across sources? Some publishers use prior date + 4 years; we use the running average block interval since genesis, which lands those dates a few weeks earlier. - How many halvings until zero? 33 non-zero halvings. After H33 the per-block reward in satoshis rounds to zero; daily issuance reaches zero a few halvings before calendar 2140. --- # Bitcoin Halving Countdown URL: https://btcoak.com/halving-countdown Category: reference ## What it is A live block-anchored countdown to Bitcoin's next subsidy halving. The schedule lives in the consensus rule that pays each block: every 210,000 blocks the per-block subsidy halves. The next halving is block 1,050,000 — H5 — around April 2028. ## How the countdown is computed The countdown reads from the live block height: `blocks_remaining = next_halving_height − current_height`; `days_remaining ≈ blocks_remaining × avg_block_interval / 86,400`. Cycle progress is `(current_height − prior_halving_height) / 210,000`. The reward formula is `subsidy(i) = 50 × 0.5ⁱ BTC` until rounding to zero satoshis at i = 33. The four-year shorthand assumes ten-minute blocks; realised intervals have averaged closer to nine and a half minutes since genesis, which is why every halving so far has arrived ahead of its four-years-from-prior date. ## What changes when the subsidy halves - **Daily new BTC issued** is cut in half mechanically. After H4 (April 2024) daily issuance dropped to ≈ 450 BTC/day; after H5 it will drop to ≈ 225 BTC/day. - **Annualised monetary inflation** continues its decay — already below 1% post-H4, dropping to ~0.4% post-H5. - **Miner revenue mix** shifts further toward fees. The 2024-04-20 print recorded the highest fee share on record (73.8% fees vs subsidy on the day of the Runes launch); structural fee-share ratchet upward continues with each halving. - **Cycle-overlay narratives** weaken with each halving as the relative supply shock shrinks. See [/halving-cycles](/halving-cycles) for the per-cycle peak-multiple decay. ## Block-anchored vs calendar-anchored countdowns Two countdowns are commonly shown across explainer sites: the calendar-from-prior-halving projection (prior date + 4 years) and the block-anchored projection. The block-anchored read is the canonical one because consensus is block-driven, not date-driven; the calendar projection is a planning shorthand. Different feeds report current height with second-to-minute lag depending on confirmation policy (1-conf vs 6-conf), so the projected halving date can swing by hours over a few-day window without anything material changing. ## Common questions - How many days until the next Bitcoin halving? Around 730 days as of April 2026; the live count updates as new blocks land. Block-anchored, not calendar-anchored. - What block number is the next halving? Block 1,050,000 (H5). The per-block reward will drop from 3.125 BTC to 1.5625 BTC. - Why is the date approximate? Block intervals vary; the projection uses the running average since genesis. The four-year-from-prior calendar estimate has overshot every actual halving date so far by two to four weeks. - Does the halving still affect price? Increasingly contested. The 2024 cycle landed a softer forward return than prior cycles; some analysts argue institutional flow has structurally overtaken the halving as the dominant price driver. See [/halving-dates](/halving-dates) for the per-cycle return table. --- # Bitcoin All-Time High URL: https://btcoak.com/all-time-high Category: reference ## What it is Bitcoin's all-time high (ATH) is the highest daily close on record. The current ATH, days since it was set, and live distance from it are the headline figures here; behind them sits a per-cycle peak table that records every cycle's close-space ATH and the drawdown that followed. ## How it is computed `ath(t) = max(price[0..t])` on the daily-close series, recomputed forward as new closes land. `days_since_ath` is the calendar-day count since the running maximum was last set; `distance_from_ath(t) = price(t) / ath(t) − 1`, which is non-positive by construction and resets to zero whenever a new running maximum prints. Cycle peaks are the local daily-close maxima within each halving window, not the running maximum across the whole series. The series uses daily closes from 18 Jul 2010 forward. Daily closes round off intraday wicks, so the canonical 2011-Jun and 2014-Mar intraday highs print lower than an intraday-resolution series would. ## Per-cycle peak table | Cycle | Peak date | Close-space ATH | Days to surpass | | -------- | ---------- | --------------- | --------------- | | 2011 | 2011-06-08 | $29.60 | ~570 days | | 2013-Apr | 2013-04-09 | $230 | ~210 days | | 2013-Nov | 2013-11-30 | $1,163 | ~1,475 days | | 2017 | 2017-12-17 | $19,870 | ~1,150 days | | 2021-Apr | 2021-04-14 | $63,503 | ~210 days | | 2021-Nov | 2021-11-09 | $67,734 | ~770 days | | 2024-Mar | 2024-03-14 | $73,098 | varies | Days-to-surpass is calendar days from each cycle's daily-close peak to the day the next cycle eventually printed above it. Each cycle's path between peak and next surpass walks through the drawdown sequence on [/drawdown](/drawdown). ## When the framing fails Pre-2011 history carries thin order books and wide spreads; reading those prints as comparable to post-ETF prints understates measurement-error. Some publishers anchor cycle peaks to intraday high (TradingView convention); we use daily close throughout for cross-cycle comparability. The cleanest cross-cycle comparison uses the post-2014 series only. ## Common questions - What is Bitcoin's all-time high? The current ATH on the daily-close series; see the live cell at the top of the page for the spot value and date. - How many days since Bitcoin's all-time high? Calendar days since the most recent daily close set a new running maximum; the live cell tracks this in real time. - What was Bitcoin's all-time high in 2017? $19,870 on 2017-12-17 (daily close). The intraday high on Bitstamp that day printed slightly higher. - How long did Bitcoin take to recover from the 2017 ATH? Around 1,150 calendar days from the 2017-12-17 peak to the next daily close above $19,870 in late 2020. - Has Bitcoin set a new ATH every cycle? Yes — every halving cycle since 2012 has eventually printed a new daily-close ATH higher than the prior cycle's peak. The 2024-Mar peak is the most recent qualifier.