Patterns · what usually happens after each kind of event · 621 past events behind 223 patterns
Event patterns and their odds
What the last events like this actually did: every pattern below groups past events by kind, venue, market-cap size, category, unlock size and Bitcoin's trend, and is kept once at least 8 past cases have a measured next-day move. Prediction from past patterns, not financial advice.
Patterns
223
with odds from past cases
Past events
621
measured, each counted in every pattern it fits
Checked on recent cases
150
patterns with 20+ cases
Flipped
48
newer cases disagree with older ones
Coming up
66
scheduled events that match a pattern
223 patterns have odds from past cases as of Oct 4, 2026.
A pattern is kept once at least 8 past cases of one kind of event have a measured next-day move.
150 of them have enough cases (20 or more) to be checked on their newest ones, and 48 of those flipped: the newer cases lean the other way.
The detailsevery pattern with its odds and all of its past cases
Every pattern223 patterns
Every pattern, grouped by event kind
as of · odds counted from past cases
Each row is one pattern: how many past cases it holds, the share that were higher the next day with its 90% range, the 30-day picture and whether it held up on recent cases. Ordered by measured cases — switch to share higher after 30 days.
Source: counted over every measured event on record · fetched
How the odds are counted
How the odds are counted
as of · methodology
Every figure counts past cases, with a 90% range; it is not financial advice, and past patterns can stop working.
A pattern is a set of conditions on an event and its token at the time: kind, venue, market-cap size at the event date (micro under $50M, small to $500M, mid to $5B, large above), category group, unlock size and Bitcoin's trend (close above or below its own 50-day average). Patterns go from broad to specific, and every combination with at least 8 measured next-day moves is kept.
Every share comes with a 90% range (a Wilson interval), so 60% on 10 cases prints a wide band and the same 60% on 60 cases a narrow one. Patterns with 20 or more measured cases are re-counted on their older 60% of cases versus the newest 40%; a pattern whose majority side flips is flagged as flipped everywhere it is shown.
A scheduled event is matched to the most specific pattern whose every condition it meets (ties go to the larger count). Full write-up on the methodology page.