Padel Win Probability — what the ratings expect
Put two team ratings in, or two levels, and read the odds both ways: how often each side wins if the same match were played a hundred times, and how the games in a set are expected to split. It is an expectation, not a prediction — tonight still has to be played.
A team's rating is the average of its two players. On this scale 1000 is the usual starting point and every 400 points is roughly a 9-in-10 favorite.
Levels convert at 0.1 of a level ≈ 40 rating points, so half a level is about 200 points. That is a working approximation for club scales, not an official conversion — no federation publishes one.
How the odds are worked out
One formula does the whole page. The chance Team A wins is 1 ÷ (1 + 10 ^ ((Rb − Ra) ÷ 400)) — the Elo expectation, where only the gap between the two team ratings matters and the absolute numbers do not. A 100-point gap is 64%, 200 points is 76% and 400 points is 91%. Where you give four player ratings instead of two team ratings, each side's rating is the mean of its two players first.
Levels are converted to rating points before the same formula runs: one whole level is treated as 400 points, so 0.1 of a level is about 40. That conversion is ours. No federation and no booking platform publishes a level-to-Elo table, and there is nothing official to convert against — it is a working approximation that makes a half-level gap behave like a 200-point one.
The games line is a second, separate estimate: expected games for the favoured side = 6 × (0.5 + (probability − 0.5) × 0.8), to one decimal, with the other side taking the remainder of six. The 0.8 is a damping factor and it is also ours. Sets are streaky and margins compress, so a side holding 61% of the win odds is expected to take about 56% of the games rather than 61% of them — without the damping the games line would read as a scoreline prediction, which it is not.
Two honest limits. The displayed percentages are clamped to 1–99%, so no reading on this page says a match is settled. And a team average cannot know how two players fit together: the formula treats a 1200 with a 800 exactly like two 1000s, which anybody who has partnered badly knows is wrong. Ratings run 400–2600 and levels 0–7; a number outside those is read as a typo and rejected rather than computed.
Where this comes from
- FIDE Rating Regulations (rating difference and scoring-probability tables) Fédération Internationale des Échecs, FIDE Handbook B.02
- The Glicko system — the Elo expectation written out as a logistic curve on 400 points Mark E. Glickman, Harvard University
Rivals is not affiliated with any padel federation and nothing here is coached, reviewed or official. Where a number is our own judgement rather than a published figure, the note above says so and cites nothing. Method last checked 21 August 2026.
What a rating gap can and cannot tell you
A rating gap is a prediction, not a verdict. The formula behind every Elo-style system turns the difference between two ratings into one number: how often the stronger side wins if the same match is played many times. A hundred points is 64%, two hundred is 76%, four hundred is 91% — and none of those numbers know anything about tonight.
Doubles adds a caveat worth stating out loud. Averaging two players into one team rating ignores who is serving, who owns the net and how a pair plays together: two 1000s and a 1200 with an 800 are the same number on paper and very different opponents on court. The percentage is the ratings' expectation, not a scouting report — and if you want to see what a result does to those ratings afterwards, that is the Elo calculator.
Margins compress, too. Even a heavy favorite rarely wins to love, so the expected-games line damps the gap: a side with 61% is expected to take roughly three and a half games of six, not four and a half. Sets are streaky, and losing 6–4 to a stronger pair is closer to par than most people assume.
Questions
How is win probability calculated from ratings?
What does a 100-point rating gap mean?
How accurate is this for padel doubles?
Why bother knowing what the ratings expect?
More tools
Padel Elo Calculator
Free padel Elo calculator for doubles. Enter four ratings, a K-factor and the games score to get each player's new rating, the points moved and the win odds.
Fair Pairs Generator
Free fair pairs generator for padel. Four names in, all three possible 2v2 splits out — ranked by how even they are, with a random pick when levels match.
Padel Level Converter
Convert a padel level between the 0–7 scale, MATCHi 1–10, an LTA-style rating and club letters, with the caveat each earns. No official conversion exists.
Read next
The same math that makes Rivals ratings explainable
Rivals runs this expectation on every match your crew logs: the odds before the first serve, the rating move after the last point, and a plain-words reason for both — kept for the whole season.
Free for a crew of four. Nothing inside is metered.