The average return and win rate for each calendar month, computed from every completed occurrence on record.
For every calendar month that has fully completed in the price history, this page takes the first available daily close and the last available daily close of that month and computes the percentage change. Those changes are then averaged across every historical occurrence of that same calendar month — every January averaged with every other January, and so on — to produce the number shown for each of the twelve rows.
A month only counts once it has near-full coverage (at least 20 days of span and 4 data points), and the month currently in progress is left out of its own average entirely, since a partial month isn't a finished one. It still appears in the table, marked as the current month, so you can see where "now" sits relative to the historical pattern.
Alongside each month's average return, the table shows its win rate — the share of historical occurrences of that month that closed higher than they opened — and n, the number of times that month has actually completed since price history began. Bitcoin has traded for a bit over a decade, so even the longest-running month has well under twenty samples. A win rate built on a handful of years is a real number, but it is not the same kind of evidence as a win rate built on decades of data.
Seasonality here is descriptive, not predictive: it reports what has happened, not what must happen next. A month with a strong historical average can still close red, and a month with a weak one can still close green — the average is a summary of a small, specific set of past years, not a rule the market is obligated to follow. The Yearly Returns page answers a related but different question — how each full calendar year performed, rather than how a given month tends to perform across years.
Related: Yearly Returns — the same underlying price history broken down by calendar year instead of by month. Read more: the four-year cycle and every bear market measured.