The pricing model
Padel bookmakers quote the match winner and stop. Set betting, total games and handicaps are almost nowhere on the board — so there is nothing to collect, and we compute them instead.
Why padel can be priced from first principles
Padel inherits tennis scoring: points to 15/30/40, games to six, sets to two. That makes a match a Markov chain over points, and given how often each pair wins a point on its own serve, every market follows analytically — match winner, set betting, total games, handicaps. No bookmaker’s opinion is copied at any stage.
Golden point is not a detail
Premier Padel and WPT play a single deciding point at deuce instead of the advantage series. That measurably helps the returner — an advantage deuce lets the stronger pair grind out its edge, one point does not — so it compresses every price on the board. A model that priced padel with tennis maths would sell the favourite too cheaply.
A board we generated, right now
Live from the feed, no key needed. Each row carries the true probability, the price before margin, and the price we would offer.
Abbate/Rodriguez vs Alfonso/Libaak
| Market | Selection | True probability | Fair | Price |
|---|---|---|---|---|
| match_winner | team_a | 30.8% | 3.25 | 2.91 |
| match_winner | team_b | 69.2% | 1.44 | 1.38 |
| set_betting | 0-2 | 39.8% | 2.51 | 2.37 |
| set_betting | 1-2 | 29.4% | 3.40 | 3.18 |
| set_betting | 2-0 | 13.6% | 7.35 | 6.77 |
| set_betting | 2-1 | 17.2% | 5.83 | 5.38 |
| total_games | over_20.5 | 63.4% | 1.58 | 1.50 |
| total_games | over_21.5 | 58.5% | 1.71 | 1.61 |
| total_games | over_22.5 | 51.9% | 1.93 | 1.80 |
| total_games | over_23.5 | 47.8% | 2.09 | 1.95 |
| total_games | over_24.5 | 45.0% | 2.22 | 2.06 |
| total_games | under_20.5 | 36.6% | 2.73 | 2.49 |
| total_games | under_21.5 | 41.5% | 2.41 | 2.22 |
| total_games | under_22.5 | 48.1% | 2.08 | 1.94 |
| total_games | under_23.5 | 52.2% | 1.91 | 1.79 |
| total_games | under_24.5 | 55.0% | 1.82 | 1.71 |
| handicap | team_a_+1.5 | 36.8% | 2.71 | 2.48 |
| handicap | team_a_+2.5 | 43.1% | 2.32 | 2.14 |
| handicap | team_a_+3.5 | 50.7% | 1.97 | 1.84 |
| handicap | team_a_+4.5 | 63.9% | 1.56 | 1.49 |
| handicap | team_a_+5.5 | 73.4% | 1.36 | 1.30 |
| handicap | team_b_-1.5 | 63.2% | 1.58 | 1.50 |
| handicap | team_b_-2.5 | 56.9% | 1.76 | 1.66 |
| handicap | team_b_-3.5 | 49.3% | 2.03 | 1.89 |
| handicap | team_b_-4.5 | 36.1% | 2.77 | 2.52 |
| handicap | team_b_-5.5 | 26.6% | 3.77 | 3.31 |
Anchored on: market:match_winner · Margin: 7% · Generated: 2026-09-04 18:14 UTC
What happened when we tested it
The model was run against 4,928 finished matches — every one we hold with a winner and two complete sets, back to 2022. For each it was given the first set’s score and nothing else, then asked to predict the rest of the match, so everything it is scored on is out of sample. Each baseline was fitted on those same matches, which makes it the strongest version of "the model adds nothing".
| Predicted from the first set alone | Model | Baseline |
|---|---|---|
| Total games (mean absolute error) | 3.973 | 4.757 |
| Match winner (log-loss) | 0.3956 | 0.4068 |
| Straight sets (log-loss) | 0.5756 | 0.5862 |
lower is better
A result nobody fitted
Run with golden-point scoring the model predicts better than with advantage scoring — 0.3956 against 0.3977 on match winner, and 0.5756 against 0.5798 on straight sets. Nothing was tuned to produce that. It falls out of the chain being built the way padel is actually played, which is the strongest evidence here that the structure is right.
What this does not show
- Nearly five thousand matches is a real sample, but padel’s recorded history is short — this reaches back to 2022 because there is not much before it. These numbers will keep moving.
- The model predicts about half a game fewer per match than reality — 20.68 against 21.23 — consistently. Neither the serve level nor a prior on pair strength closes it, so what is missing is a mechanism rather than a parameter: most likely that it treats sets as independent when a losing pair regroups.
- It is a pre-match model. We do not price a match in progress, and the feed refuses to.
- Where no bookmaker has priced a match, the model can anchor on FIP rankings instead. That is the weaker signal, is quoted wider, and every row says which anchor produced it.
Want it against your own coverage?
The full feed is a Pro key. Every selection ships with the probability and the pre-margin price alongside ours, so you can load your own margin.