Fluuk · our probability, band by band

When we say 85%,
it wins about 85%.

Every match we price gets our chance of each player winning. Below, every completed match since our current model went live, grouped by how confident we were, set against how often that player actually won.

87.4% → 87.2%ATP 85–90% band · said → won · n 94

87.4% → 89.9%WTA 85–90% band · said → won · n 79

3,484completed matches since 16 Sep 2026

What we said (average in band)How often it wonn = matches

ATP · men

2,311 matches · said 66.0%, won 66.8% overall

90%+n 45

said 91.5%won 97.8%+6.3 pts±8.1 pts by chance

85–90%n 94

said 87.4%won 87.2%-0.1 pts±6.7 pts by chance

80–85%n 156

said 82.4%won 85.3%+2.9 pts±6.0 pts by chance

65–80%n 796

said 72.0%won 73.2%+1.3 pts±3.1 pts by chance

50–65%n 1,220

said 57.4%won 57.5%+0.1 pts±2.8 pts by chance

WTA · women

1,173 matches · said 67.6%, won 66.8% overall

90%+n 61

said 92.4%won 95.1%+2.7 pts±6.7 pts by chance

85–90%n 79

said 87.4%won 89.9%+2.5 pts±7.3 pts by chance

80–85%n 105

said 82.6%won 79.0%-3.5 pts±7.3 pts by chance

65–80%n 337

said 71.7%won 71.2%-0.5 pts±4.8 pts by chance

50–65%n 591

said 57.5%won 56.2%-1.3 pts±4.0 pts by chance

What this shows

Whether our numbers mean what they say. A band is honest when the dot sits on the ring: of the favourites we gave 85–90%, about that share won. We measure the probability we actually served at the time, on completed matches only, and count each match once.

What it does not show

It is not a betting return. An honest probability is the starting point; whether following it grows a kitty faster than the ~5% a bank pays depends on the prices taken and the staking, and we are measuring that separately before we say anything about it.

ATP · men

WTA · women

What this shows

What it does not show

Bet responsibly