Live prices are currently unavailable — the exchange feed could not be reached and no recent cached reading is held.

Research hub

Data sources and methodology

Every figure published on this site comes from somewhere, and this page says where. It sets out the two kinds of data the site keeps, the sources behind each curated table, the vintage of every series, and the places where the compiled record is genuinely uncertain. If you intend to rely on a number here, read this page first.

Two kinds of data, kept apart on purpose

The site publishes two categories of figure, and conflating them is the most common way to misread what is here. The first is the live series. The BTC/USD spot price and the daily candle series are fetched from the Coinbase Exchange public API by the site's canister, cached for a few minutes, and re-fetched on demand. They are current readings rather than settled records, they change between visits, and each one carries the timestamp of the moment it was fetched. When the upstream feed cannot be reached, the canister serves the last cached value and labels it as cached rather than pretending it is fresh.

The second category is the curated historical record: yearly returns, the halving history, the drawdown and recovery table, the milestone timeline and the glossary. These were compiled once from published sources, checked against one another where two sources overlap, and then fixed. They change only when the dataset is deliberately republished, which is why each carries a vintage range rather than a fetch time. The curated tables are the stable spine of the site and the only material offered for download.

A practical consequence follows from that split. A live figure quoted in an article will be wrong by the time you read it, and the site treats it as a reading rather than a fact. A curated figure will still be right in a year, provided the source it came from was right in the first place — which is the question the rest of this page addresses.

The source register

Every dataset the site publishes, with the source it was compiled from and the period it covers. The register is served from the same metadata the tables themselves carry, so it cannot drift out of step with the data.

The source register is unavailable

The canister did not return the source metadata. The datasets themselves remain available from the research hub.

Open the research hub

How the curated tables were prepared

The curated tables were assembled from published exchange and block-chain records, then cross-checked wherever two independent sources cover the same period. The method was deliberately conservative: where sources disagreed, the site took the series that could be re-fetched and re-verified, and recorded the disagreement rather than averaging it away. No figure on this site has been interpolated, smoothed or modelled. If a price is not attested by a source, it does not appear.

The early years required the most judgement. Before 2015 there was no single dominant exchange, and the quoted price of bitcoin differed materially from venue to venue — sometimes by more than twenty per cent on the same day. For 2010 to 2014 the site uses Bitstamp and CoinDesk historical series, which are the most complete surviving records for that period, and every table that reaches back that far says so in its provenance line. From 2015 onward the figures come from Coinbase Exchange daily candles, which are published in a consistent format and can be re-fetched at any time.

The halving records are the most secure of the curated tables, because the underlying facts are on the block chain itself. The dates, block heights and subsidy amounts are protocol facts that anyone can verify against a node. Only the price around each halving is a market figure, and it is labelled as approximate for that reason. The milestone timeline is the least quantitative table on the site: it is a dated narrative drawn from contemporaneous public reporting, and its value lies in the chronology rather than in any number it contains.

What the record does not tell you

A reference site is only as honest as its account of its own limits. These are the ones that matter most.

  • These are compiled figures, not audited ones

    The curated tables were assembled from public sources by hand and checked against one another. They have not been audited by a third party, and they carry no warranty. A reader who needs a number to a particular standard should go back to the primary source listed in the register above.

  • Early prices are approximate by nature

    Before 2015 the market was thin, fragmented and unregulated. A single daily close for 2011 or 2012 is a reasonable summary of a day's trading, not a precise record of it. Treat the pre-2015 figures as indicative and the post-2015 figures as reliable.

  • Nothing is adjusted for inflation, fees or taxes

    Every figure is nominal US dollars. A return calculated from these tables is a gross price return: it ignores trading fees, custody costs, spreads and the tax a real holder would owe. Real returns over a fifteen-year period are materially lower than the nominal ones.

  • Lost coins are not accounted for

    The supply figures count coins that have been mined, not coins that can still move. Estimates of permanently lost bitcoin run into the millions, and the site does not attempt to adjust for them. Circulating supply here means mined supply.

  • The live series depends on one venue

    The spot price and daily candles come from Coinbase Exchange alone. That is a deep and well-run market, but it is one venue, and its price is not necessarily the price available to a buyer elsewhere. The site does not compute a cross-exchange index.

  • Survivorship shapes the history

    The milestone timeline records events that turned out to matter. Exchanges that failed, proposals that were abandoned and forks that faded are represented only where they changed the course of the asset. The record is a history of what happened, not a complete inventory of everything that was attempted.

Revisions and corrections

Curated data does get revised. When a source publishes a correction, or when a re-fetch against the Coinbase API returns a candle that differs from the one previously recorded, the site updates the affected table and moves its vintage label forward. It does not silently overwrite the old figure and pretend the change never happened: a revision is a fact about the record, and the vintage range on each table is there so a reader can tell which version they are looking at.

The live series are not revised, because they are not stored as history. Each fetch replaces the cached value, and the previous reading is gone. If you need a permanent record of a particular day's price, take it from the curated daily candles rather than from the spot panel, and note the date you captured it.

The editorial standards page sets out how corrections are handled in the writing itself, and what the site does when it gets something wrong. The short version is that it says so, on the page, rather than quietly editing the claim away.