Every calendar month of bitcoin since the tape opens, one bar per year: pick a month and see all of its years at once, with the median of them as a dashed line. The running month is hollow and counts for nothing.
+12.8 % October · median of 16 years · 11 of 16 up · best +211.0 % (2010) · worst -36.8 % (2011) · last closed month Sep 2026 · October 2026 is running, hollow, in no count, Sep-30 2026, measured and published by askthetape.com.
«December is bullish». «September is the weak one». The claim is everywhere and the instrument behind it usually is not, so here it is on our own tape: pick a month above and every one of its years is a bar, the plain close-to-close move from the last close of the previous month to the last close of that one. The dashed line is the MEDIAN of those years, not the average — one 1987-style year can carry an average on its own, and on this tape it does: November averages +34.3 % and its median is +8.5 %, because of 2013 alone. Under the bars the plate says how many of those years finished up, and which year was the best and the worst. Read the count before you repeat a season: 16 or 17 observations per month is a handful, not a law, and nothing here is a forecast. The month we are in right now is drawn hollow and is in no median, no count and no record until it closes.
| leg | Bitstamp daily closes (data/px_daily_100.csv), the same leg as «How stretched the price is» and «Days people watch» · month close = the last daily close of that calendar month, UTC · move = that close over the previous month's close − 1, one bar per year, close to close · nothing is interpolated and no second source is used |
|---|---|
| cadence | one bar per calendar month; a month's bar appears the day the month closes. A single month's view therefore gains one bar A YEAR by construction — that is the instrument, not a stale feed, and no «source is quiet» plate is drawn on this page for that reason. The running month is drawn hollow from the last daily close on Free; on Tape it is measured to the hourly Bitstamp pull, and on Desk to the live price — the same bar, a fresher price. |
| since | August 2010 — the first CLOSED month of the leg, measured on this build (the tape itself opens mid-July 2010, so that first partial month has no previous close to measure against and is not drawn) |
| 🩸 how few observations this is | 16 or 17 observations per month. Thirteen Decembers is not a law and this page will not pretend otherwise: significance — none claimed, no test is run, no p-value is printed, and the median is shown instead of the average precisely because a single year can own an average (November: mean +34.3 % against a median of +8.5 %, driven by 2013 alone). Two months carry 17 years rather than 16 simply because the tape opens in August 2010. |
| 🩸 the running month is hollow | October 2026 is not judged: it is drawn as a hollow dashed bar at -2.69 % so far (22 of 31 days), and it is excluded from every median, every «N of M up» count and every best/worst on this page. A month joins the numbers the day it closes. |
| 🩸 the first view follows the calendar | the view that opens the page and the card is the month we are in — in September it is September — then the other eleven in calendar order, and «every closed month» last. That rule was set on 2026-09-19 and is machine-checked, so the page never opens on a month nobody is asking about. |
| the whole ruler in one line | Jan +5.2%, 9 of 16 up · Feb +11.6%, 11 of 16 up · Mar -0.6%, 8 of 16 up · Apr +10.0%, 11 of 16 up · May +7.1%, 9 of 16 up · Jun +2.3%, 9 of 16 up · Jul +8.2%, 11 of 16 up · Aug -8.5%, 6 of 17 up · Sep -1.4%, 8 of 17 up · Oct +12.8%, 11 of 16 up · Nov +8.5%, 10 of 16 up · Dec +2.2%, 8 of 16 up |
| next door | the last view here — every closed month in a row — is the SAME row that already sits on «Days people watch» (cal_btc_month): one series, two pages, nothing duplicated and nothing recomputed. |
| 🩸 the four quarters | Q1 to Q4 are the same instrument one step up: the close-to-close move from the last close of the previous quarter to the last close of that one, one bar per year, 64 closed quarters since December 2010. A quarter carries 16 observations here against 16 or 17 for a month — the same handful, four buckets instead of twelve, and thirteen of anything is not a law: read the count before you repeat a season. The quarter we are in right now is hollow and is in no median, no count and no record until it closes. Ether has the same four views on «Which months were kind to ether». |
| the quarter ruler in one line | Q1 +1.2%, 8 of 16 up · Q2 +18.7%, 10 of 16 up · Q3 +2.2%, 9 of 16 up · Q4 +28.2%, 10 of 16 up |
| policy | no backfill of this page's own cut · read it here · export: whole history |
| this build | We show October 2025 as the latest complete October and September 2026 as the latest complete month. |
| policy | no backfill · read it here · export: whole history |
| month | median | finished up | best year | worst year |
|---|---|---|---|---|
| January | +5.2 % | 9 of 16 | +83.4 % (2011) | -32.4 % (2015) |
| February | +11.6 % | 11 of 16 | +63.5 % (2013) | -31.3 % (2014) |
| March | -0.6 % | 8 of 16 | +178.7 % (2013) | -32.8 % (2018) |
| April | +10.0 % | 11 of 16 | +346.1 % (2011) | -17.3 % (2022) |
| May | +7.1 % | 9 of 16 | +149.7 % (2011) | -35.4 % (2021) |
| June | +2.3 % | 9 of 16 | +84.2 % (2011) | -37.3 % (2022) |
| July | +8.2 % | 11 of 16 | +39.8 % (2012) | -17.1 % (2011) |
| August | -8.5 % | 6 of 17 | +65.8 % (2017) | -38.6 % (2011) |
| September | -1.4 % | 8 of 17 | +22.0 % (2012) | -37.3 % (2011) |
| October | +12.8 % | 11 of 16 | +211.0 % (2010) | -36.8 % (2011) |
| November | +8.5 % | 10 of 16 | +450.0 % (2013) | -37.0 % (2018) |
| December | +2.2 % | 8 of 16 | +59.0 % (2011) | -34.6 % (2013) |
The chart itself is drawn in the browser: turn JavaScript on to hover a day, pick a window and draw your line. The numbers above come from the last build, Oct-23 2026 21:24Z.
What the tape says today · what's new · all 111 metrics