Tennis Predictions

The ELO model surfaces picks only where it disagrees with the closing market by ≥5% on WTA grass-court matches. This is not a proven edge. The +6.25% backtest that originally justified this filter was re-run and reconciled in July 2026 against two other studies: overall Elo-edge ROI is significantly negative at every threshold, and the WTA-grass slice is statistically insignificant under both methodologies tested (+7.46% CI [−1.89, +17.23]; +3.5–4.1% CI [−4.9, +13.0]) — consistent with a selection artifact rather than a real edge. The filter stays only to accumulate a live closing-line record, which is the forward test; expect it to confirm the null. Other surfaces and tours do not beat the Pinnacle line and are not surfaced.