Across nearly 50,000 tour rematches the head-to-head leader wins 59% of the time, but that edge is almost entirely because better players build better records. When the head-to-head favourite is the betting underdog, they win just 36.0%, virtually identical to the 36.7% the odds already imply, so backing the 'he owns him' angle against the price adds no edge.
Head-to-head records are the first stat fans and tipsters reach for. We rebuilt the head-to-head history of every ATP and WTA match in our database and tested it on 49,798 rematches where the two players had met before and one led the series.
The H2H leader won 59.1% of rematches. Most of that is the better player winning: a player usually leads a head-to-head because they are stronger. The useful question is whether H2H adds anything beyond the betting price.
Kovalchik's comparison of eleven published tennis forecasting approaches on a full ATP season found the plainest ones, regression on rankings and an Elo rating, performed best [1]. Overall ability carries the prediction; pair-specific history adds little.
On the 4,881 rematches with market odds, split by whether the H2H leader was also the market favourite:
| Situation | What happens |
|---|---|
| H2H leader is the market favourite (they agree) | Favourite wins 70.0% (n=3,337) |
| H2H leader is the market underdog (they disagree) | See below (n=1,544) |
The cases that matter are the disagreements, where the head-to-head points one way and the price the other:
| H2H leader (as the underdog) | Win rate |
|---|---|
| What they actually won | 36.0% |
| What their odds already implied | 36.7% |
| Head-to-head's edge over the price | −0.7 points |
As the underdog, the H2H leader won 36.0% against the 36.7% their price implied. The market already includes the head-to-head.
Leaders up 3-0 or more in the disagreement cases went 37.5% against 41.0% implied, a −3.5-point underperformance. The sample is 48 matches, so this is suggestive only, but there is no sign that a dominant record hides value.
H2H helps pick winners and does not help beat prices. When it disagrees with the price, the H2H underdog wins what the odds say. Our model treats head-to-head as a weak input the market has mostly priced.
Method: 306,141 ATP/WTA singles matches (all levels) processed chronologically to reconstruct each pair's head-to-head at match time; 49,798 rematches with a prior meeting and a clear H2H leader; 4,881 of those with recorded market odds, split into agreement/disagreement vs the favourite. "Implied" win probability is the no-vig odds (each player's 1/odds normalised so the pair sums to 100%). Small-sample caveat noted for the dominant-H2H subset (n=48). Re-verified 2026-08-07 on the grown database: 53,307 rematches, leader wins 59.4%; leader-as-underdog 36.1% actual vs 36.8% implied.
References:
[1] S. Kovalchik, "Searching for the GOAT of tennis win prediction," Journal of Quantitative Analysis in Sports, vol. 12, no. 3, pp. 127–138, 2016.
For guessing the likely winner, yes: the head-to-head leader wins about 59% of rematches. For betting, no: the odds already price the head-to-head in, so backing the H2H leader when they are the underdog wins only about 36%, the same as the price implies.
Not on that basis alone. When the head-to-head leader is the betting underdog, they win roughly 36%, essentially exactly what their odds already predict. The 'he owns him' angle has no measurable edge over the market price.
In our data, even 3-0-or-better leaders did not beat their market price (they slightly underperformed it), though that sample was small. The market appears to already account for lopsided head-to-heads.
See today's picks: published before the match, graded in public →