About & Methodology
Methodology
If a number on this site looks wrong, this is where to check it: the formula behind it, where its data comes from and when it was last fetched.
Data sources and update times
Times are UTC.
Bitcoin price (ticker bar, every page)
- Source
- OKX public market data: BTC-USDT for US dollars, BTC-EUR for euros
- Updated
- Every minute; an open page refreshes it every minute
Swiss franc rate
- Source
- European Central Bank, daily EUR/CHF reference rate (OKX has no franc market)
- Updated
- Fetched every hour; the ECB publishes it once per working day
Bitcoin daily closes (Power Law, analysis, calculators)
- Source
- From 11 January 2018 OKX BTC-USDT daily candles (close of each UTC day); 18 July 2010 to 10 January 2018 Coin Metrics' community reference price
- Updated
- New closes every six hours; the model's formulas stay, our own regression is redone
Prices of shares and preferred shares, after hours, volume, exchange rates for them
- Source
- Yahoo Finance (public chart data)
- Updated
- Every 15 minutes
Daily closes and dividends since issuance; rates of STRC, STRF and SATA
- Source
- Yahoo Finance (public chart data)
- Updated
- Every six hours
Bitcoin holdings of companies, ETFs, exchanges and states
- Source
- The holders' own publications: annual reports, regulatory filings, press releases, ETF issuers' daily holdings
- Updated
- By hand whenever the source publishes something new, each entry with its source and reference date
Terms of the preferred shares (stated value, rate, ex-date, rank, debt of the issuer)
- Source
- The issuers' own filings and announcements
- Updated
- By hand whenever the source publishes something new, each entry with its source and reference date
Bitcoin bought and sold by country (globe on the Power Law page)
- Source
- Public trades on the exchanges' markets in a local currency (OKX, Binance and Binance.US candles, Coinbase, Upbit, bitFlyer, Bitso); the treasuries' holdings per country
- Updated
- Trades every minute, kept for 24 hours; holdings as in the Treasuries list
Bitcoin blocks (difficulty adjustment on the Power Law page)
- Source
- mempool.space public API: height, time and difficulty of the newest block and of the current period's first block
- Updated
- Every two minutes
Fear & Greed, market cap, dominance and next halving (home page)
- Source
- Alternative.me's Crypto Fear & Greed Index; Bitcoin's share of all crypto assets' market cap from CoinGecko; the newest block's height (mempool.space) for the bitcoin mined so far (with the halving schedule, times the OKX price) and the blocks to the next halving
- Updated
- Every 10 minutes; the Fear & Greed Index changes once a day
World map and globe
- Source
- Natural Earth (public domain)
- Updated
- Fixed
Mining pools, hashrate and block reward (home page, solo mining calculator)
- Source
- mempool.space (mining statistics)
- Updated
- Every 10 minutes
Liquidity heatmap: prices and open positions
- Source
- OKX: hourly candles and open positions (open interest) of the BTC-USDT perpetual, three weeks back; the liquidation prices estimated from them (5×, 10×, 25×, 50×, 100× leverage)
- Updated
- Every minute
Strategy's bitcoin bought with preferred shares (Digital Credit)
- Source
- Strategy's weekly Form 8-K filings at the SEC (EDGAR)
- Updated
- Checked every 6 hours; Strategy files every Monday
Bitcoin news (headlines)
- Source
- RSS feeds of BTC-ECHO and Blocktrainer (German), Bitcoin Magazine and Cointelegraph (English); only title, source and time
- Updated
- Every 30 minutes
- Bitcoin price last fetched
- 2026-11-13 23:01 UTC
- Latest daily close
- 2026-11-12
If a fetch fails, the site keeps the last good figures and tries again in the next round.
Power Law
The Power Law describes Bitcoin's price as a straight line on a log–log chart: price against the days since the genesis block on 3 January 2009. Our floor is the published formula that other Power Law pages use too, checked with our own regression. The top band is a curve through the cycle tops: it comes down toward the floor and never reaches it. The trend is the straight Power Law line until the middle between floor and top band lies below it, and that middle since Jan 2022. The top band always stays above the trend, and the trend above the floor.
The research behind this model, with every test, table and reference: our paper →
days = days since the genesis block (3 January 2009, 00:00 UTC)
Floor (USD) = 0.42 × 10^−16.493 × days^5.688
Top band (USD) = Floor (USD) × e^(3.56 ÷ (1 + (days ÷ 4692)³))
Trend (USD) = min(Floor (USD) ÷ 0.42, √(Floor (USD) × Top band (USD)))
Residual = log10(daily close ÷ Trend (USD))- Intercept a (log10)
- −16.493
- Exponent b
- 5.688
- R²
- 0.961
- Daily closes
- 5’962
- From – to
- 2010-07-18 – 2026-11-12
- Our own regression on these closes
- 10^−16.328 × days^5.634
Method: a and b are the published Power Law fit that other Power Law pages calculate with too. They are fixed, so our floor agrees with those pages and a day's floor never changes afterwards – in the ticker bar too. We check them with our own regression: least squares of log10(closing price) on log10(days since genesis), every daily close weighted equally, redone whenever a new daily close arrives. The regression is the check and not the model, because it follows the market: after months of low prices it moves the whole line, and with it every earlier day's floor. The trend has no number of its own: it is the lower of the straight Power Law line and the middle between the floor and the top band on the log chart, √(floor × top band). Until Jan 2022 that was the Power Law line, at 2.38 × the floor. Since then it is the middle of the band: today the trend is 1.62 × the floor, which is 0.68 × the straight Power Law line. Of the trends we compared it fits the daily closes best, over the whole history and since 2022. The top band's two numbers change only when a new cycle top joins, a year after it was reached; the trend since 2022 then moves with them.
Check: our own regression
The same regression on the daily closes up to the end of each year and on every close so far, with the value its line gives for today beside the formula's (the floor is 0.42 × that). While the price stays low for months, the regression drifts; the formula does not.
| Daily closes up to | Intercept a (log10) | Exponent b | Its line today | Against the formula |
|---|---|---|---|---|
| End of 2023 | −16.588 | 5.714 | €146’869 | +1.2 % |
| End of 2024 | −16.540 | 5.699 | €143’927 | −0.8 % |
| End of 2025 | −16.540 | 5.700 | €143’979 | −0.8 % |
| 12 Nov 2026 (every close) | −16.328 | 5.634 | €132’460 | −8.7 % |
Floor and top band: measured against the straight Power Law line, the bear markets' lowest closes were at ×0.43 (2011), ×0.53 (2015), ×0.56 (2018), ×0.41 (2022), and 3.7 % of all days closed below the floor at 0.42 × that line. The lows have kept their distance to it; the cycle tops – the all-time highs that stood for at least a year – have not: measured against the floor they were at ×36.9 (2011), ×26.0 (2013), ×14.9 (2017), ×6.55 (2021), ×2.67 (2025). The top band follows that: floor × e^(3.56 ÷ (1 + (days ÷ 4692)³)), the curve of this form with the least squares of the tops' distance to the floor on the log chart, through the cycle tops only. Half of that distance was gone by November 2021; what is left shrinks ever more slowly and never reaches zero. The band comes closer to the floor every year and stays above the trend: today it is 2.63 × the floor, which is 1.62 × the trend.
Is the growth slowing down?
At the tops clearly, in the middle of the prices since the last cycle, at the lows not so far. Every cycle top was a smaller multiple of the floor than the one before, while the lows stayed where they were: the lowest close since the top of 2025, in June 2026, was at 0.41 × the straight Power Law line like the low of 2022. So the floor keeps the published formula, and the range of prices narrows from above only. The top band measures that: a curve through the five cycle tops, fitted by least squares to their distance from the floor on the log chart. Half of that distance was gone by November 2021; the rest shrinks ever more slowly, in the long run like 1 ÷ days³, and never reaches zero. The band passes within 13 % of each top. Fitted on the first four tops it was 22 % above the fifth, and leaving out any one top moves it by at most 22 % at that top. The trend follows from the two lines and has no fitted number of its own: it is the straight Power Law line as long as the booms rose far above it, and since Jan 2022 the middle between floor and top band on the log chart. In the last cycle the closes lay evenly between the floor and the top, without the short spikes of earlier booms, so that middle is where the prices are: in the four years to November 2026 the price closed above the straight line on only 22 % of days and averaged 28 % below it, but above the trend on 46 % of days and 8 % below it on average. Over the whole history the trend fits the closes better than the straight line (a deviation of 0.295 instead of 0.302 in log10), since 2022 clearly better (0.156 instead of 0.195), and better than a curve of its own fitted to the closes. Taken from the cycle tops known at dates from 2014 to 2024 and compared with the four years that followed each, it was closer to the later prices than the straight line and than a trend fitted to the closes. Certain it is not. The evidence that the middle of the prices has left the straight line comes from one cycle, and the lower half of the closes has not come down against the straight line over the years; what came down is the upper half. Five tops are few, and several curves pass through them about as well. What decided between them is how far the price moves within a year: the curves closest to the five tops end on the floor within a few years, nearer to it than the price moves in one calm year. The band in use is 2.63 × the floor today, 2.0 × in 2029, 1.5 × in 2033 and 1.2 × in 2045. That assumes booms keep shrinking about as fast as the price's swings have over the last ten years; if they shrink more slowly, tops will be above the band, and a new cycle top moves it and the trend with it. In dollars every line keeps getting steeper, the top band too: each is a multiple of the floor, and the floor grows by 36 % over the next twelve months. From top to top the price rose 39-fold, 17-fold, 3.4-fold and 1.85-fold; that cannot go on slowing, because from 2025 to 2029 the floor alone rises 3.4-fold. The clearest test is the floor's: on 2 Dec 2028 it reaches the highest close so far, $124’673 on 6 Oct 2025. If Bitcoin has not passed that price by then, it is below the floor, and the Power Law no longer describes it.
Zones
The zones lie between the model's own lines. The floor zone is below the floor. Accumulation reaches from the floor halfway to the trend on the log chart, to √(floor × trend), today 0.79 × trend. A year after a day in accumulation the close was higher on 100 % of days. Fair lies around the trend and ends halfway between the trend and the top band on the log chart, at √(trend × top band), today 1.27 × trend; hot reaches from there to the top band. That is where the record turns: a year after a day in fair the close was higher on 78 % of days, after a day in hot on 6 %. Euphoria is above the top band; after a day there the close a year later was higher on 0 % of days. These shares count every daily close since July 2010 that has a close a year later, sorted with today's lines. The days of one cycle resemble each other, so they describe five cycles, not thousands of separate cases. The lines come closer together every year, and the zones with them: they have no fixed multiple of the trend. Since Jan 2022 the trend is the middle between floor and top band, so on the log chart accumulation is the lowest quarter of the way from the floor to the top band, fair the middle half and hot the top quarter.
- Floor zoneTodayHeavy buy
The price is lower than at almost any time in Bitcoin's history, measured against its trend. The model sees it as extremely undervalued. So far, the price was higher a year later on 100 % of the days in this zone.
below the floortoday below €60’778 - AccumulationTodayHeavy buy
The price is well below its long-term trend, closer to the floor than to the trend. Historically, this has been a good range for buying and holding for years. So far, the price was higher a year later on 100 % of the days in this zone.
from the floor to halfway to fair value on the log charttoday €60’778 to €77’378 - FairTodayBuy
The price is close to its long-term trend: the model's neutral estimate of what Bitcoin is worth today. So far, the price was higher a year later on 78 % of the days in this zone.
around fair value: from halfway down to the floor to halfway up to the top band on the log charttoday €77’378 to €125’415 - HotTodayHold
The price is near the top band, the level the cycle tops have reached. In the past, this has usually come before larger corrections. So far, the price was higher a year later on only 6 % of the days in this zone.
the upper half of the way from fair value to the top bandtoday €125’415 to €159’668 - EuphoriaTodaySell a small portion only
The price is above the top band, the model's line through the earlier cycle tops. By the model's measure it is extremely high. So far, the price was higher a year later on 0 % of the days in this zone.
above the top bandtoday above €159’668
“The floor reaches today's price in about X months”
day = (price ÷ (0.42 × 10^−16.493))^(1 ÷ 5.688)
X months = (day − days today) ÷ 30.442-year MA band
2Y MA = average of the last 730 daily closes
x = 730 ÷ days
level = 6.688 × x ÷ (1 − (1 − x)^6.688)
Upper line = 2Y MA × (level + max(0, 5.0401 − 0.0007178 × days)²)The 2-year MA band compares the price with its own average over the last two years. The lower line has held up: the price closed below its 2-year average on 28 % of days, and a year later it was higher on 97 % of them. The upper line is usually drawn at a fixed five times the average. We checked that against every daily close. The price last closed that high in Jan 2018, and what each cycle's boom reached has shrunk by about half every time: ×18.0 (2013), ×9.96 (2017), ×4.88 (2021), ×2.48 (2024) the average. Three halvings in a row are a trend, not an anomaly of the last cycle, and the first came before the indicator was widely followed, so front running alone does not explain it. It is what the Power Law says: the cycle tops come down toward the floor, and the floor grows more slowly every year, so the price runs less far ahead of its own past. The upper line follows the peaks: least squares on √(peak ÷ 2Y MA − level) through each cycle's highest multiple, a straight line that reaches zero in Mar 2028. The level is how far a price that rises exactly with the Power Law is above its own 2-year average, ×1.37 today; the line does not go below it, or plain Power Law growth would count as a top. Today the upper line is at ×1.49 the average. Four peaks are few: the first three put the line 9 % above the fourth. And the multiple peaks early in a boom, in 2024 a year and a half before the top, so the line marks a market that has run hot rather than the top itself.
Analysis views
Bear markets and recovery
A fall of at least 70 % from an all-time high whose low comes at least 90 days after the high. Recovery is the first close above the old high.
Ladder
Today's price against every daily close: the share of days that closed higher; the rungs are the deciles of all closes.
Cycle tracker
Price ÷ price on the halving day (28 Nov 2012, 9 Jul 2016, 11 May 2020, 20 Apr 2024), and the distance to each cycle's all-time high, by days since.
CAGR
The average yearly return over 4, 5 and 10 years and since the first close: (price ÷ close at the start)^(1 ÷ years) − 1, for every day and today at the live price.
Yearly returns
Each calendar year: last close of the year ÷ last close of the year before − 1; the current year at the live price.
Difficulty adjustment
Average block time = (time of the newest block − time of the period's first block) ÷ blocks since. Expected change = 2’016 × 10 minutes ÷ (2’015 × average) − 1, at most ×4 and ÷4, Bitcoin's own rule. Previous change = difficulty of the period's first block ÷ the block before − 1. Expected date = newest block's time + blocks to go × average.
Peak multipliers
Each cycle top's close ÷ the trend on that day.
Drawdown
Each day's close ÷ the highest close up to that day − 1. The largest is the deepest of these, the average their mean over every day.
Projection
Floor and trend at the start of each quarter for ten years, calculated in US dollars; euros and francs at today's rates.
To the Power Law page →Download every daily close with trend, floor and residual (CSV)
Ticker bar and currencies
The bar above every page and every amount in EUR, CHF or USD.
Distance to the floor = price ÷ Floor (USD) − 1
Sats per unit of currency = 100’000’000 ÷ Bitcoin price in that currency
Bitcoin in CHF = Bitcoin in EUR × ECB EUR/CHF rate
Dollars per euro = Bitcoin in USD ÷ Bitcoin in EURThe model calculates in US dollars. Its figures (fair value, floor, top band) are converted into euros and francs at the ratio of today's Bitcoin prices in those currencies.
Bitcoin Companies
The holdings come from the holders' own publications: annual reports, regulatory filings, press releases and the holdings ETF issuers publish daily. Every entry names its source and the date it is valid for. Share prices and exchange rates come from Yahoo Finance, the Bitcoin price from OKX.
Share of the 21 million
BTC holdings ÷ 21’000’000.
Value
BTC holdings × today's Bitcoin price.
Change over 7, 30 and 90 days
Holdings today minus the last recorded holdings 7, 30 or 90 days ago; without an entry that old it shows a dash.
Market cap
Share price × shares outstanding as of the company's latest report.
mNAV
Market cap ÷ (BTC holdings × today's Bitcoin price). Above 1 the market pays more than the Bitcoin are worth, below 1 less.
BTC per share and BTC yield
BTC per share = BTC holdings ÷ shares outstanding. The BTC yield is its growth since the last entry of the previous year.
US spot ETF flows
Per ETF the change in holdings from its previous recorded day, added up over all US spot ETFs; the value at that day's Bitcoin close.
Digital Credit
Every instrument fact is checked against the source named before it is published (mostly the issuer's own announcements and filings). An instrument without a source and date appears only as “being verified”, without figures. Prices, volume, history and dividends paid: Yahoo Finance.
Effective yield
What buying at today's price pays per year: dividend rate × stated value ÷ price. Below the stated value it is higher than the rate, above it lower.
BTC coverage
How many times the issuer's bitcoin covers the instrument: BTC holdings × BTC price ÷ (debt + the stated value outstanding of this and every preferred ranking before or alike).
BTC price at 1×
The BTC price at which coverage is exactly 1×: (debt + preferreds ranking before or alike) ÷ BTC holdings. Below it the bitcoin alone no longer covers the claims.
mNAV
Market value of the common shares ÷ value of the bitcoin held. Above 1× the market pays more than the bitcoin is worth.
Current rate (STRC, STRF, SATA)
Dividends of the last 12 months ÷ today's price. My Goal and the digital-credit calculator use it when you enter no rate yourself.
Prices, stated values and dividends are in the instrument's own currency; without the issuer's debt stated, no coverage is shown.
Calculators and My Goal
Calculated in your browser from the model, today's prices and rates; nothing you enter is sent to us.
- Spread σ of ln(close ÷ trend)
- 0.68
- Persistence after 30 days
- 0.90
Price paths
Fair value (trend), floor or top band of the model on each future day, or a price you enter.
Simulations (Monte Carlo, lump sum vs. DCA)
Monthly price paths that stray from the trend like Bitcoin since 2010, with the spread and persistence above, measured on all daily closes. They start from a fixed seed, so the same inputs always give the same result.
Currencies
Amounts are converted between EUR, CHF and USD at today's rate, from the Bitcoin prices of the ticker bar.
My Goal
Proceeds = BTC sold × target price. Cash flow per year = proceeds × the digital-credit rate. The date a target price becomes fair value solves trend(day) = target price. Every Power Law figure there is a model estimate.
Forecast: what if …
Price = price without the scenario × (bitcoin on exchanges today ÷ left on exchanges)^k. What buyers buy leaves the exchanges evenly over the chosen years. States buy as much per inhabitant as the BITCOIN Act's 1 million BTC for the USA; pension funds 1 % of about 58 trillion dollars.
Data we maintain by hand
Holdings and the terms of preferred shares are entered by our team from the source named, with the date they are valid for. A preferred share without source and date is listed as “being verified”, without figures. Every change to a holder's BTC is kept in its history, from which the changes, the BTC yield and the ETF flows are calculated.
We check new entries on our test site first; bitcoin-powerlaw.com shows them once they are published.
If a figure looks wrong, the source named beside it is the reference.
Principles
- The Power Law's floor uses the published formula that other Power Law pages calculate with too, so our floor agrees with theirs. The trend is fitted to every daily close since July 2010 and comes down toward the floor, as the booms have become smaller; the parameters and our own regression as a check are on the Power Law page.
- Every calculator prints its assumptions under the result, and the methodology page gives the formula behind every figure, so you can redo the arithmetic yourself.
- A model describes the past; it does not promise the future. Every Power Law figure is an estimate.