Across 209,000 ATP and WTA matches since 2023, the rested player beats one who played yesterday just 52.9% of the time. A three-set win the day before costs about 4.5 points net of opponent quality, on hard and clay alike. Our later 218,890-match study refines this into a gradient in games played.
Update 2026-10-09: the flat −4.5 point rule below was superseded on 2026-08-12. Re-measured as a residual against ELO on 218,890 player-matches, next-day fatigue is a smooth gradient in games played: +3.42 points after a win of 16 games or fewer, −4.36 after 31 or more on one day of rest, and the three-set versus two-set gap is −3.36 points rather than −4.5. Details: the fatigue-gradient study. The raw win rates below stand.
"He played a long one yesterday" is the most common fatigue claim in tennis previews. We tested it on every ATP and WTA singles match since 2023, main tour, Challenger and ITF: 209,006 matches.
When one player had competed the day before and the other was fresh, the rested player won 52.9%. That is close to a coin flip, and draw and scheduling quirks could account for most of it. Playing yesterday is not, by itself, a real handicap.
Among players who competed the day before, split by the length of that match:
| Yesterday's match | Win rate today |
|---|---|
| Won in straight sets | 51.6% |
| Survived a three-setter | 44.7% |
A seven-point gap.
A player who scrapes through three sets is often the weaker one in a tight bracket, so part of the 44.7% is quality, not legs. We controlled for it two ways.
Each player against their own baseline. Win rate after a three-setter against the same player's usual rate: −7.5 points, and 75% of 1,704 players did worse after a three-setter.
Same round, opponent strength removed. Winning a three-setter means advancing to a tougher next opponent. Restricting to players who won yesterday puts both groups in the same round (still a ~7-point gap), and scoring each result against its rating-expected win probability removes the next opponent's strength:
| Surface | Extra-set fatigue cost |
|---|---|
| Hard | −4.4 points |
| Clay | −4.6 points |
The confound-stripped cost of a three-set win the day before is about −4.5 points, a little under half the raw figure.
The one peer-reviewed study we know of on this question, 20,320 Grand Slam matches from 1992 to 2011 [1], found the same direction: sets spent in the previous round cut the probability of winning the next one, with women affected from one extra set and men from two. Our sample is ten times larger and covers best-of-three tennis at every level.
Clay does not punish more: −4.4 on hard and −4.6 on clay are statistically identical. ATP and WTA both show a 6–7 point raw gap. Grass has too few back-to-backs to judge. The penalty is as strong on the main tours as on Challenger and ITF, and it holds with Grand Slam best-of-five matches removed (routine 3–0 wins otherwise pass as battles).
On the roughly 20,000 matches where we have odds, the same split against the no-vig price:
| Yesterday's match | Won next day | Implied by odds |
|---|---|---|
| Won in straight sets | 51.2% | 51.0% |
| Survived a three-setter | 45.2% | 46.2% |
The market prices it. Straight-set winners land almost exactly on their implied number; the three-set group sits about a point under, inside the noise at this sample and gone once you pay the margin. As a standalone bet there is nothing here. The use is defensive: it shows when "he's tired" is already in the price.
Ignore "he played yesterday" when the previous match was routine. After a genuine three-set battle the effect is real but worth a few points, not a flipped coin, and it is no larger on clay.
Method: 209,006 ATP and WTA singles matches across all levels (main tour, Challenger and ITF), Jan 2023–present. Fatigue coefficient isolated three ways: each player against their own baseline; restricted to yesterday's winners so round-progression is held constant; and adjusted for next-opponent strength via rating-expected win probability. The market-pricing check uses the ~20,000 matches from 2026 with recorded odds; "implied" is the no-vig win probability from those prices. Raw win rates re-verified 2026-08-07 on the grown sample (216,081 matches): rested-vs-played-yesterday 52.7%, straight-set winners 51.8%, three-set winners 45.2%.
References:
[1] D. R. Goossens, J. Kempeneers, R. H. Koning, and F. C. R. Spieksma, "Winning in straight sets helps in Grand Slam tennis," International Journal of Performance Analysis in Sport, vol. 15, no. 3, pp. 1007–1021, 2015.
Not on its own. On the roughly 20,000 matches where we have odds, straight-set winners won 51.2% the next day against 51.0% implied by the price, and three-set-war winners won 45.2% against 46.2% implied. The market already prices the hangover. Its real use is defensive: it tells you when the 'he's tired' angle is already in the number, so you don't overpay for it.
Partly, and that was the biggest confound to remove. Restricting to players who won yesterday puts both groups in the same round, and scoring each match against its rating-expected win probability nets out opponent strength directly. That cuts the raw seven-point gap to about four or five points, but it doesn't disappear. Tougher next opponents explain roughly a third of the raw number. The rest looks like fatigue.
See today's picks: published before the match, graded in public →