> ## Documentation Index > Fetch the complete documentation index at: https://padelapi.org/docs/llms.txt > Use this file to discover all available pages before exploring further. # Player Form and Pair Synergy > How the Padel API measures whether a padel player is in form and whether a pair beats the sum of its parts, by comparing actual results against what the Elo ratings predicted. Win rate tells you what a player or a pair has done. It does not tell you whether that was **more or less than expected**. Winning 70% of your matches is poor form for the world number one and an outstanding run for a pair ranked 40th. What matters is the gap between the results and the strength behind them. The Padel API measures that gap with one method and publishes it through two endpoints: `GET /api/players/{id}/form`. How a player is doing lately against what was expected of them. `GET /api/pairs/{player1}-{player2}/synergy`. Whether two players together win more than their individual strength predicts. Same measurement, different subject. Both return the same core fields with the same meaning, so a client that reads one can read the other. Both endpoints [require a **Business** subscription](https://padelapi.org/billing). *** ## The measurement For every match, the API takes the [Elo rating](/docs/guides/padel-elo-rating) of all four players **as it stood right before that match** and turns the gap between both pairs into a [win probability](/docs/guides/padel-match-prediction). Averaging those probabilities over a set of matches gives the win rate the player or the pair was expected to reach. Comparing it with the win rate they actually reached gives the gap. Because the ratings are taken at the time of each match, the expectation only uses what was known before the ball was served. A player who has since climbed the rankings is not judged retroactively against their current rating. ### Shared fields | Field | Meaning | | - | - | | `expected_win_rate` | Win rate predicted from the ratings of the four players at the time of each match, as a percentage. | | `actual_win_rate` | Win rate actually achieved over the same matches, as a percentage. | | `delta` | `actual_win_rate` minus `expected_win_rate`, in percentage points. Positive means over-performing. | | `matches_played` | Matches behind the numbers above. | | `by_period` | The same four values over consecutive calendar periods, oldest first, each with its `from` and `to` dates. Periods without matches are left out. | | `trend` | Direction of `delta` across `by_period`: `rising`, `stable` or `declining`. | Both endpoints accept the same filters as the stats endpoints: `after_date`, `before_date`, `level` and `round`. ### How to read `delta` | `delta` | What it usually means | | - | - | | **Above +10** | Clearly over-performing. Winning matches the ratings said were close, or beating stronger pairs. | | **+3 to +10** | Slightly ahead of expectation. | | **-3 to +3** | Performing at their level. The ratings describe them well. | | **-10 to -3** | Slightly behind expectation. | | **Below -10** | Clearly under-performing: losing matches they were favoured to win. | `delta` is `null` when there are too few matches for the gap to mean anything. A handful of results is dominated by luck, so the API publishes the rates but withholds the gap. `matches_played` tells you how much evidence is behind any number you do get: a `delta` of +12 over 60 matches is a much stronger statement than the same value over 8. `trend` looks at how `delta` moves across `by_period`, giving more weight to the periods with more matches. It is `null` when too few periods have a `delta` to tell a direction. A good `delta` does not last forever by construction. Every win against expectation lifts the players' ratings, which raises what is expected of them next. Sustained over-performance is what a player or pair on the way up looks like; a `delta` that stays high for a long time is a genuinely remarkable run. Qualification matches count in both endpoints. Since the expectation already accounts for how strong the opponents are, beating a much weaker pair in qualifying barely moves the gap. *** ## Player form `GET /api/players/{id}/form` measures a recent window: the last 180 days by default, ending today. Set `after_date` and `before_date` to measure any other window. `by_period` is split by calendar month. ```bash theme={null} curl 'https://padelapi.org/api/players/65/form' \ -H 'Authorization: Bearer YOUR_API_TOKEN' ``` ```json theme={null} { "id": 65, "self": "/api/players/65/form", "from": "2026-04-01", "to": "2026-09-28", "expected_win_rate": 71.85, "actual_win_rate": 80.00, "delta": 8.15, "matches_played": 30, "by_period": [ { "from": "2026-04-01", "to": "2026-04-30", "expected_win_rate": 70.12, "actual_win_rate": 66.67, "delta": -3.45, "matches_played": 6 }, { "from": "2026-05-01", "to": "2026-05-31", "expected_win_rate": 71.30, "actual_win_rate": 71.43, "delta": 0.13, "matches_played": 7 }, { "from": "2026-06-01", "to": "2026-06-30", "expected_win_rate": 72.05, "actual_win_rate": 83.33, "delta": 11.28, "matches_played": 6 }, { "from": "2026-07-01", "to": "2026-07-31", "expected_win_rate": 72.61, "actual_win_rate": 100.00, "delta": null, "matches_played": 3 }, { "from": "2026-09-01", "to": "2026-09-30", "expected_win_rate": 73.20, "actual_win_rate": 87.50, "delta": 14.30, "matches_played": 8 } ], "trend": "rising", "windows": { "last_7_days": { "expected_win_rate": 73.94, "actual_win_rate": 100.00, "delta": null, "matches_played": 2 }, "last_30_days": { "expected_win_rate": 73.20, "actual_win_rate": 87.50, "delta": 14.30, "matches_played": 8 }, "last_90_days": { "expected_win_rate": 72.92, "actual_win_rate": 88.89, "delta": 15.97, "matches_played": 18 } }, "same_period_last_year": { "expected_win_rate": 68.40, "actual_win_rate": 65.52, "delta": -2.88, "matches_played": 29 }, "streak": { "result": "won", "matches": 6 }, "third_sets": { "won": 5, "lost": 2 }, "tiebreaks": { "won": 7, "lost": 4 }, "connections": { "player": "/api/players/65" } } ``` On top of the shared fields, form adds: | Field | Meaning | | - | - | | `from`, `to` | The window actually measured, so you know what the defaults resolved to. | | `windows` | Overlapping zooms of 7, 30 and 90 days ending on `to`, each with the shared core. They are not periods: they all end on the same day and each contains the shorter ones. Use them for "how is this player doing right now". | | `same_period_last_year` | The shared core over the same window one year earlier, to compare a player against themselves. | | `streak` | Current run of consecutive wins or losses up to `to`. `result` is `won` or `lost`, `matches` its length. | | `third_sets` | Deciding sets won and lost within the window. Retirements are left out, since the third set was never completed. | | `tiebreaks` | Sets settled 7-6 won and lost within the window. | Keep the two time axes apart. `by_period` is a series of consecutive, non-overlapping months: plot it, and read `trend` from it. `windows` is three nested zooms that all end today: show them side by side, never on the same line. Form is measured for the player, whoever they played with. A player who changed partners mid-season carries the same form across both partnerships, and each partner's rating is part of the expectation. *** ## Pair synergy `GET /api/pairs/{player1}-{player2}/synergy` measures every match the pair has played together by default, because a partnership is judged over its whole life, not the last few weeks. `by_period` is split by calendar quarter, so you can see a pair settling in or wearing out. ```bash theme={null} curl 'https://padelapi.org/api/pairs/65-66/synergy' \ -H 'Authorization: Bearer YOUR_API_TOKEN' ``` ```json theme={null} { "id": "65-66", "self": "/api/pairs/65-66/synergy", "expected_win_rate": 69.43, "actual_win_rate": 78.13, "delta": 8.70, "matches_played": 96, "by_period": [ { "from": "2025-01-01", "to": "2025-03-31", "expected_win_rate": 66.80, "actual_win_rate": 64.29, "delta": -2.51, "matches_played": 14 }, { "from": "2025-04-01", "to": "2025-06-30", "expected_win_rate": 67.95, "actual_win_rate": 76.47, "delta": 8.52, "matches_played": 17 }, { "from": "2025-07-01", "to": "2025-09-30", "expected_win_rate": 69.10, "actual_win_rate": 80.00, "delta": 10.90, "matches_played": 15 }, { "from": "2025-10-01", "to": "2025-12-31", "expected_win_rate": 70.24, "actual_win_rate": 81.25, "delta": 11.01, "matches_played": 16 }, { "from": "2026-01-01", "to": "2026-03-31", "expected_win_rate": 70.66, "actual_win_rate": 83.33, "delta": 12.67, "matches_played": 12 }, { "from": "2026-04-01", "to": "2026-06-30", "expected_win_rate": 71.02, "actual_win_rate": 84.62, "delta": 13.60, "matches_played": 13 }, { "from": "2026-07-01", "to": "2026-09-30", "expected_win_rate": 71.50, "actual_win_rate": 77.78, "delta": 6.28, "matches_played": 9 } ], "trend": "rising", "connections": { "pair": "/api/pairs/65-66" } } ``` The expectation for a pair is built from each player's own rating, and those ratings come from every match they play, with this partner or any other. So a positive `delta` says the two of them together win more than their individual strength predicts: the partnership adds something. A negative one says they are worth less together than apart. The endpoint returns `404` when the two players have never played together. *** ## Next steps How the rating behind the expectation is built and updated. How a rating gap becomes a calibrated win probability. Career win rate, titles and set and game tallies, to put next to form and synergy. Profile of a pair, with links to its matches, stats and synergy. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.