Basketball projections. Verified results. Know the coverage.
Find the saved forecast, see what happened, and understand what the comparison can tell you. This guide separates our verified game results from our separately saved pregame baselines.
A player-game pair needs both a saved pre-tip minutes and points forecast and a validated played final result. The same player can appear in several games; pairs are not a count of unique players.
Published basketball forecast and final-result coverage
Across the 1,421 included WNBA player-game pairs, the descriptive mean absolute miss is 4.6 minutes and 4.8 points. These are selected historical comparisons, not an estimate of future accuracy or sportsbook win rate. See the calculation and every source game.
Two kinds of saved forecast
Use the right page for the question
01 · Forecast versus final
How close was this game's forecast?
Game recaps pair an original pre-tip forecast with verified final minutes and points. They show actual minus projected for each included player, with the original capture time and source game.
Missing forecasts and players who did not play are excluded from the comparison. A recap's sample does not represent the complete slate.
These separate snapshots contain Research estimates based on recent games and a verified roster. They preserve the model, roster receipt and capture time.
They are not lineup-adjusted: roster membership does not establish a starting five or current availability. After tip, the page remains a dated baseline; it does not become a live projection or the forecast graded in a recap.
Published coverage is not the entire prediction ledger.
Only games and player comparisons that pass the evidence checks appear in these search pages. Missing records are not backfilled with a current forecast, an assumed lineup or a synthetic zero. A saved pre-tip forecast alone does not prove alignment with every injury or lineup revision.
The live board is the place to check reported lineups, current model status and sportsbook props. The archives answer what was saved and what happened; their results do not establish calibrated probabilities or returns from betting.