Rotator snapshots its top bullish, rotate-out and underperforming signals every day. After 30 days, we compare the call against the median coin over the same 30 days — beating the market is the whole test, and half of all coins beat it by definition. The result is published here: the calls that beat it, the ones that moved our way but trailed it, and the ones that went the wrong way outright. The counter below restarted on with scoring engine ; the retired v2 engine's final record is in part 4. Every model change is logged here.
The best 20, the worst 20 and every pick are just below ↓
Our old engine (April to September 2026) was confirmed on 76.2% of 863 graded calls. That grading was generous: it counted the best price reached inside a 7 to 14 day window, where calling every coin would have scored about 68%. Read it as an upper bound, roughly 8 points better than chance.
The current engine is held to a stricter test: a call counts only if the coin beats the median coin over the next 30 days, where chance is 50%. It started on 11 September, so its first results arrive from 11 October. Until then there is nothing to claim yet.