Console

Predictions API

Live win probabilities. The match state behind them.

Get a match win probability for each player in ATP Tour men’s singles. Join forecasts to score revisions, inspect their history and bring them into your application through REST or WebSocket.

Get an API keySee the data

One forecast, with its context intact.

Inspect a recorded model response, the probability for each player and the match revision behind the calculation.

Recorded ATP Tour forecast · 6 Oct 2026

Novak Djokovic50.7%
Alex de Minaur49.3%

Recorded model output. Player order follows player_ids in the response.

Response excerpt
{
  "data": {
    "prediction_id": "prd_cb34aa59-8a6e-41aa-999f-66484e847aaf",
    "match_id": "mat_777f2e56-db02-431a-92f8-aa2e265ffa9b",
    "player_ids": [
      "ply_26f0d340-6ba3-4d48-8a14-22797f6c1361",
      "ply_48bc0dbb-3445-4f5e-ae3e-0b9affc7ebab"
    ],
    "probabilities": [
      0.5073358118534088,
      0.4926641881465912
    ],
    "point_id": "pnt_5f5a07ed-073c-4749-b4b4-a2929528395f",
    "point_sequence": 65,
    "point_revision": "1",
    "score_revision": "2",
    "state_revision": "151",
    "model_version": "v3.3-77feat-filled",
    "computed_at": "2026-10-06T12:08:30.661292+00:00",
    "quality": {
      "enrichment": "point_by_point",
      "features_used": 73,
      "features_total": 77,
      "qualification": "shadow"
    }
  },
  "meta": null,
  "page": null,
  "error": null
}

A forecast you can join, store and revisit.

Match identity, player order and score revision preserve what a prediction refers to. Computation time and model version describe how to compare it with another forecast.

match_id
The DFT match identity used to join points, players and the tournament edition.
player_ids / probabilities
The probabilities array follows player_ids order. Values range from 0 to 1 and refer to the same match state.
score_revision
Identifies the score revision used for the forecast, so updates remain aligned with the match state.
computed_at
UTC timestamp retained with the accepted forecast.
model_version
The model version retained with the prediction for comparisons and history.

From a live match to a useful update.

  1. Choose an ATP Tour match.

    Discover a supported match and keep the same match ID for scores, points and predictions.

    Read the Predictions API reference

  2. Keep the score attached.

    Read the current prediction with its score revision, computation time and model version.

    Explore Real-time data

  3. Read the prediction history.

    Page through retained forecasts or retrieve one immutable prediction by its ID.

    Read prediction history

Questions

Add predictions to your tennis application.

Create your account, use the API reference for prediction access and connect forecasts to the same match records your application already uses.

Get an API keyCompare plans
Read the docs