The most points you can win and still lose
Everyone who watches tennis has seen it: the player who won more points loses the match. It happens because points only count through games, and games only count through sets — each one resets to zero. That raises an exact question with nothing to estimate: how far can it go?
Best of three, tiebreak at 6-6 in every set
63.1%
of the points — 106 of 168 — and still lose, 6-7 6-0 6-7.
How the scoreline is built
Every game you win, you win to love. Every game you lose, you lose 2-4, and every tiebreak you lose goes 5-7. The set you are allowed to win, you win 6-0 without dropping a point — it banks 24 points at no cost. Then you lose the sets that decide the match as narrowly as the scoring allows. That is the whole trick: win decisively where it does not matter, lose by the smallest possible margin where it does.
Deuce makes it worse
The instinct is that dragging the games you lose to deuce banks you extra points. It does raise your share of that game: losing a game 2-4 gives you 33% of its points, and taking it through one deuce to 3-5 gives you 37.5%. But you are already above 63% overall, so every extra deuce point pours volume into a bucket scoring below your average and drags the total down. The ceiling takes every lost game in the minimum four points.
Tiebreaks work the same way. Losing one 5-7 is better than losing it 6-8, for the same reason.
The format moves the ceiling
- Best of five goes higher — 64.8%, 171 of 264. More sets means more places to bank a love set, so a longer match leaves more room to lose while winning.
- A match tiebreak raises it — 64.0%. Replacing a whole third set with one short tiebreak means the decider is a small part of the match, so losing it narrowly costs less of the total.
- No-ad scoring raises it most — 66.0%, 66 of 100, very nearly two points in every three. At 3-3 no-ad plays a single deciding point, so a game can be lost 3-4: three of the seven points, the best lost game any format allows. It is the same logic as deuce, run the other way — what you want in a game you lose is as many points as possible for as few points played.
Of the six formats this site models, that one has the highest ceiling.
What this does not tell you
This is a ceiling, not a rate. It says what the scoring system permits, not how often real matches come anywhere near it — and how often the player with more points loses is a question about real players, answerable only from match records, not from the rules. Nothing on this page needs a probability at all: it is counting, which is why it is exact.
The ceiling is exact within each format as stated. Every game, tiebreak and set shape was searched; a different tiebreak length or set length would move it, which is exactly what the formats page is about.