Open aggregates across all finalized signals. The data updates automatically as new results come in.
Share of hits among all finalized signals in the selected period.
Last 12 weeks. Each line shows the Win Rate of an algorithm by week for weeks that had signals.
All aggregates on the page are calculated using a simple and verifiable formula — without smoothing, Bayesian corrections, or other statistical adjustments.
Share of hits out of total completed signals. Formula: WR = W ÷ (W + L), where W is hits and L is misses.
Only finalized signals — those with a definitive outcome — count toward the numerator and denominator. Active signals and pushes (refunds) are not included in the Win Rate calculation.
The return on a hypothetical flat stake of one unit per signal. For a hit, the return equals the odds taken; for a miss, it is zero. Formula: ROI = (sum of returns − N) ÷ N, where N is the total number of signals.
Example: 100 signals, of which 60 hit at average odds of 1.65, ROI = (60 × 1.65 + 40 × 0 − 100) ÷ 100 = −1%. A positive Win Rate does not guarantee a positive ROI — the difference in average odds can outweigh accuracy.
A refund applies only in the Tau algorithm and only on whole match totals (1.0 and 2.0), when the actual total exactly matches the bookmaker's line. Refunded signals are excluded from both the numerator and the denominator of both metrics: they affect neither Win Rate nor ROI. In the archive such signals are marked with the letter P.
The Results page updates in real time as signals are finalized. Each signal receives its status (W, L, or P) when the match ends or its time window closes, and its result is immediately reflected in the aggregates on this page.
The "Week" period includes signals from the last seven calendar days. "Month" — the last thirty. "All time" — the entire period of public statistics, since the start of 2024. Active and pending signals are excluded from the aggregates so the numbers do not depend on matches that have not yet been played.
Before making decisions based on these statistics, keep in mind a few fundamental limitations.