The phrase promises something without ever saying how much. A player has lost the first set, the commentator says they are down but not out, and everyone nods. Nobody asks what the second half of that is worth.
It is worth 16.52%.
That figure comes from 116,382 completed best-of-three singles matches played between January 2023 and September 2026. In every one of them a player lost the opening set. Roughly one in six went on to win anyway.
So the idiom is honest. It is also almost empty, because the one number it implies is the one number that barely matters.
The same words, three times the chance
What decides the odds is how the first set was lost, and the spread is wide enough to make the phrase misleading.
Lose the first set in a tiebreak, 7-6, and the comeback rate is 21.92% across 14,265 matches. Lose it 6-0 and the rate is 7.18% across 5,833 matches. Those are the same four words of consolation covering a chance that differs by a factor of three.
The ladder in between runs in the order you would guess. A 6-4 first set leaves 19.41% over 25,546 matches. 7-5 leaves 18.89% over 10,252. Then 6-3 at 17.39% over 25,970, 6-2 at 13.3% over 19,739, and 6-1 at 11.06% over 14,269.
Read down that list and the idiom stops being one statement. It is seven, and the speaker picked none of them.
What the phrase quietly gets wrong
“Down but not out” carries a suggestion that is not in the data. It implies the way back is a matter of character, of refusing to be finished.
If that were true, conditions would move the number. They barely do. The comeback rate is 17.2% on grass over 3,593 matches. On clay it is 16.79% over 47,944, and on hard 16.47% over 61,751. Under three quarters of a percentage point covers all three.
What does move it is how much tennis is left. In best-of-five matches on the men’s tour the rate rises to 23.98% across 1,597 matches. A player who loses the first set there still has three sets to win, not two.
So the thing that rescues you is not resolve. It is remaining format, and the margin you have already handed over.
Why a phrase like this survives
Idioms of this shape last because they cannot be checked in the moment. By the time anyone could settle whether the player was out, the match is finished and the phrase has done its job.
They also last because they are useful. A commentator with four seconds has no room for a spread of seven percentages. “Down but not out” is a fair compression of a real fact. The chance genuinely is not zero. At 6-0 down it is 7.18%, which is small and is not nothing.
The trouble starts when the compression is mistaken for the measurement. Someone hears the phrase and forms a sense that half of these matches turn around. Then the results surprise them, and the results were never unlikely.
A better version of the same sentiment
The fix is not to retire the phrase. It is to attach the score to it.
“Down but not out at 7-6” means something close to one in five. “Down but not out at 6-0” means something closer to one in fourteen. Both are true, both are short enough to say out loud, and neither leaves the listener guessing which of the seven numbers was meant.
That is the general test for any idiom that smuggles in a probability. Ask what the phrase would have to know to be right, then ask whether the person saying it knew that. Here the answer is the first-set score, which is on the screen the whole time and is left out anyway.
The figures above come from the Live Tennis API, a JSON service for developers. It covers live scores, fixtures, players and match events, with point-by-point history behind them. The data is published openly under CC BY 4.0, free to reuse with attribution.
Retirements and walkovers are excluded throughout. Every state is counted once per match and read from the point-by-point record, with nothing estimated. That is the only reason the first-set breakdown can be reported at all.
