What data do I need to build a PrizePicks prop model?
You need three things: graded lines to train on, player game logs for recent form, and the current board to score, all from this API. The base rate to beat on standard NBA lines is 45.6% over across 45,126 graded lines.
| Market | Standard over rate | Standard rows | Goblin over rate | Demon over rate |
|---|---|---|---|---|
| Points | 43.9% | 726 | 64.1% | 20.0% |
| Pts+Rebs | 46.0% | 695 | 64.2% | 22.9% |
| Pts+Rebs+Asts | 44.4% | 723 | 62.5% | 21.4% |
| Rebounds | 47.9% | 422 | 69.9% | 19.3% |
| Pts+Asts | 46.2% | 563 | 63.6% | 21.9% |
| Rebs+Asts | 49.7% | 459 | 65.6% | 21.3% |
| Assists | 44.8% | 281 | 68.6% | 18.7% |
| 3-PT Made | 53.9% | 165 | 66.1% | 18.3% |
| Steals | 50.0% | 172 | 67.0% | 14.6% |
| Blks+Stls | 50.4% | 119 | 65.6% | 20.4% |
The training set flattens each graded line into opening_line, closing_line, tier, opponent, home_away and rest_days, alongside actual_value, result and the game’s realized minutes. Those last three fields describe what happened; keep them out of pre-game inputs. Game logs add prior-game form, and time-bounded team and opponent stats add the matchup.
Decide whether the model predicts the next box-score value or the chance of exceeding a specific recorded line. For an earlier decision, reconstruct that line at the cutoff instead of using a later closing_line. Split by date and game, keep voids separate, and do not train only on projections still on the current board.
Observed line history includes a partial NBA capture on 27 Oct 2026; broader recording starts 28 Oct 2026. Game logs can cover older seasons, as listed in meta.coverage, but those seasons are not a historical PrizePicks line archive. Free is an integration sample of up to 100 training rows; Developer supports the full available paginated training set and season-wide game logs.
One flat row per graded line
curl "https://prizepicksapi.com/api/v1/training-set?league=NBA&tier=standard&limit=1000" \
-H "Authorization: Bearer $PRIZEPICKS_API_KEY"Go further
Related questions
Independent data service, not affiliated with PrizePicks. Information only, not betting advice.