On 62,744 ATP and WTA matches, the overall (all-surface) ELO rating predicts better than the surface-specific rating on both tours, and a weight sweep finds the optimal mix at 0.7 overall plus 0.3 surface, almost exactly the split FiveThirtyEight published years ago. Our bot now uses that measured blend.
Most serious tennis models keep separate ELO ratings per surface, on the assumption that a rating mixing surfaces must be worse. Our bot's ELO probability used the surface-specific rating from the start. A run of losing bets led us to test that assumption.
In Kovalchik's comparison of 11 published tennis forecasting models on a full ATP season, ELO-based methods were among the most accurate [1]. Angelini, Candila and De Angelis extended ELO with margin-of-victory weighting and report Brier scores around 0.212 on tour-level tennis [2]. FiveThirtyEight's tennis ELO blended the two ratings rather than choosing, with roughly 0.71 weight on the overall rating and 0.29 on the surface rating [3]. We measured the question on our own data, both tours, down through Challenger and ITF, where samples are thinner and surface noise should matter most.
Every match since January 2025 where both players had at least 10 rated matches on the match surface: 62,744 matches. Two win probabilities per match from the standard ELO formula, pre-match ratings only, scored by Brier score (lower is better; 0.25 is a coin flip).
| Tour | Matches | Surface ELO | Overall ELO | Better |
|---|---|---|---|---|
| ATP | 36,958 | 0.2129 | 0.2095 | overall |
| WTA | 25,786 | 0.2091 | 0.2055 | overall |
The pooled rating wins on both tours, in every cut we looked at. Both figures sit in the range of published tour-level ELO benchmarks [2], a sanity check rather than a like-for-like comparison, since samples and eras differ.
We swept the blend weight, w × overall + (1−w) × surface, for w from 0 to 1 (the pool grows to 72,112 matches here because the sweep run included late-window rows).
| Tour | w=0.0 (surface) | w=0.3 | w=0.5 | w=0.7 | w=1.0 (overall) |
|---|---|---|---|---|---|
| ATP | 0.2114 | 0.2087 | 0.2076 | 0.2072 | 0.2079 |
| WTA | 0.2080 | 0.2049 | 0.2037 | 0.2034 | 0.2043 |
The optimum is w = 0.7 on both tours, better than pure overall too. FiveThirtyEight's 0.71/0.29 [3] replicates on a dataset with far more Challenger and ITF tennis. Surface information is worth about 30% of the rating.
Surface splits divide the same match history. A player with 300 career matches might have 40 on clay, and a rating built from 40 matches is a noisier estimate of current strength than one built from 300. Surface identity is real but modest and slow to change; small-sample noise is large.
We also tested whether ratings earned at ITF or Challenger level are inflated, so that players stepping up underperform their ELO.
| Tour | Bucket | n | Expected | Actual | Gap |
|---|---|---|---|---|---|
| ATP | stepping up | 7,936 | 42.1% | 42.1% | +0.0pp |
| ATP | stepping down | 4,066 | 64.2% | 64.3% | +0.2pp |
| WTA | stepping up | 4,733 | 42.9% | 42.1% | −0.9pp |
| WTA | stepping down | 3,264 | 60.7% | 62.1% | +1.4pp |
No effect on the ATP side. On the WTA side, a small tilt of about one point against players stepping up: worth noting, too small to justify separate rating pools per level.
On the morning of the study we switched the bot's ELO probability to the pure overall rating, per the first comparison. By evening the sweep gave a better answer, and the input became the 0.7 overall + 0.3 surface blend, falling back to the pooled rating alone when a player has fewer than 10 matches on the surface.
The per-surface ratings remain visible in our analysis, because a 150-point gap between a player's clay and hard ratings still says something about a matchup. Surface also enters the point-level serve and return simulation, which relies on the result that tennis points are close enough to independent and identically distributed for such models to work well [4].
One rating system: the K-factors, starting ratings and update rules are ours, and a differently tuned surface ELO (for example with margin weighting [2]) could close some of the gap. One period: matches from January 2025 onward. The 10-match surface threshold biases the sample toward matches where the surface rating should be at its best, and it still loses. The sweep resolution is 0.1, so the true optimum is near 0.7, not exactly at it. Our data leans harder on Challenger and ITF tennis than most published studies, which limits comparability with tour-only benchmarks [1], [2].
Method: our own ELO system (K-factor by match count, level-aware starting ratings, updated per match, main tour through Challenger and ITF). Pre-match ratings only, no leakage from the match being predicted. Surface comparison requires 10+ rated matches on the surface for both players. Blend applied at the rating level before the standard expectation formula. Transition buckets use the average level of a player's last 10 matches vs the current match level, minimum 5 recent matches. Brier scores are means over qualifying matches; database snapshot through July 2026.
References:
[1] S. A. Kovalchik, "Searching for the GOAT of tennis win prediction," Journal of Quantitative Analysis in Sports, vol. 12, no. 3, pp. 127–138, 2016.
[2] G. Angelini, V. Candila, and L. De Angelis, "Weighted Elo rating for tennis match predictions," European Journal of Operational Research, vol. 297, no. 1, pp. 120–132, 2022.
[3] FiveThirtyEight, "How we're forecasting the 2016 US Open," fivethirtyeight.com, 2016.
[4] F. J. G. M. Klaassen and J. R. Magnus, "Are points in tennis independent and identically distributed? Evidence from a dynamic binary panel data model," Journal of the American Statistical Association, vol. 96, no. 454, pp. 500–509, 2001.
No. On 62,744 ATP and WTA matches since 2025, the overall (all-surface) ELO rating predicted match winners better than the surface-specific rating on both tours (Brier 0.2095 vs 0.2129 on ATP, 0.2055 vs 0.2091 on WTA). Surface splits divide a finite match history into smaller, noisier samples, and below a few hundred matches per surface the added noise outweighs the surface signal.
About 0.7 overall + 0.3 surface. Sweeping the blend weight from 0 to 1 on 72,112 matches, w=0.7 minimized the Brier score on both the ATP and WTA sides, beating pure overall and pure surface ratings. This independently replicates the 0.71/0.29 split FiveThirtyEight published for its tennis forecasts.
Barely. On the ATP side there is no effect at all (players stepping up win exactly as often as their rating predicts). On the WTA side there is a small real tilt of about one percentage point below expectation for step-up players and about 1.4 points above for players stepping down. That is worth a note in analysis, not separate rating systems per level.
See today's picks: published before the match, graded in public →