Why doubles accuracy is higher: Doubles partnerships have more combined data points, reducing volatility. Singles matches are inherently less predictable — one off day swings everything.

Reading this chart: Each row is an SR™ difference bucket. The green bar shows actual win rate for the higher-rated player, and the orange line shows what our model predicted. When they line up, our model is well-calibrated.
Brier Score: 0.0931 — On a scale where 0 is perfect and 0.25 is a coin flip, our probability estimates are highly accurate.

Our prediction engine uses a proprietary hybrid system with proprietary K-factors and scaling parameters, with rating deviation tracking for uncertainty. Exact scaling factors and K-values are not published.

Win probability formula: P(win) = f(Δ) where Δ is the rating difference between players — exact parameters are proprietary.

All stats are calculated from real PPA Tour match results. No cherry-picking, no backtesting tricks. This is our live model running in production.

Read our full methodology →