EXPERIMENTAL · NOT SAVED
Preseason preview
A first look at our current models using owner-confirmed starters. Preseason estimates are unvalidated and may be less reliable as lineups and rotations change. For research and entertainment only.
Today and the next seven days of imported preseason games · all dates and tip times Eastern. These previews are not saved, graded, or added to training history. Free throws count as first points.
Previews remain visible until 15 minutes after scheduled tipoff, including across midnight. Starter edits still close at tipoff. These are not live predictions.
| Matchup / date / tip | Game status | First-points pick | Preview status | View predictions |
|---|
TODAY · EASTERN TIME
Today’s Best Picks
The highest available first-points player probability on today’s slate, with that game’s possession and selected-player method forecasts.
New to the lab? Understand our three forecasts and their results. First points include made free throws; historical replays and saved pregame predictions are different evidence.
Loading today’s slate…
TODAY’S SLATE
Today’s Games
Tip times shown in Eastern / Central / Pacific Time · daylight saving adjusts automatically · free throws count as first points| Matchup / date / tip | Game status | First-points pick | Forecast | View predictions |
|---|---|---|---|---|
| Loading today’s games… | ||||
WALK-FORWARD REPLAY
Past predictions & results
Choose a stored date to compare a historical model replay with what happened. Replays use saved data from earlier dates; they are not predictions saved before tipoff.
ACTIVE MODELS · COMPLETED PAST GAMES
Accuracy to date · 2026 season
Loading accuracy for the 2026 season…
2026 season means 2026–27. Correct picks ÷ graded picks. Today’s games are excluded (Eastern Time). Method accuracy is conditional on our selected player scoring first—not the chance of getting both player and method right. Ties and unresolved evidence are excluded.
Historical replays use only earlier-date game history but were reconstructed later. Saved pregame predictions take precedence; each game is counted once. These are different types of evidence, not a single prospective validation.
ACTIVE MODELS · COMPLETED PAST GAMES
Accuracy to date · All seasons
Loading accuracy across past games…
Correct picks ÷ graded picks. Today’s games are excluded (Eastern Time). Method accuracy is conditional on our selected player scoring first—not the chance of getting both player and method right. Ties and unresolved evidence are excluded.
Historical replays use only earlier-date game history but were reconstructed later. Saved pregame predictions take precedence; each game is counted once. These are different types of evidence, not a single prospective validation.
HISTORICAL SCHEDULE
Stored games by date
Finding stored dates…| Matchup | Final | Evidence | |
|---|---|---|---|
| Loading the historical slate… | |||
FIRST POINTS · FREE THROWS INCLUDED
Three opening predictions
Select any stored game.Earlier dates only. These previews are not saved pregame snapshots. Possession and method estimates have not yet been validated prospectively.
FIRST POINTS · FREE THROWS INCLUDED
First points ranking
The same first-points forecast as the three opening cards.
The complete ranking will appear here after a game is selected.
| Rank | Starter | Team | First points % |
|---|
The same first-points probabilities as Three opening predictions; made free throws count. The verified actual scorer is highlighted. Historical replays were reconstructed later, not saved before tipoff. Ties are not forced into picks and unknown results remain unresolved.
Saved pregame predictions & actual results
SAVED BEFORE TIP · FIRST POINTS INCLUDE FREE THROWS
Saved pregame forecasts
The original possession, scorer, and selected-player method forecasts stay fixed. Results below use the latest saved grading evidence, not the reconstructed historical replays.
Loading saved forecasts…
| Game / saved before tip | 1 · First possession | 2 · First points player | 3 · Selected player’s method |
|---|---|---|---|
| Loading saved forecasts… | |||
Method comparison & grading rules
Both methods predict how the same selected player will score, if that player scores first. Conditional accuracy is measured only when that player scores first and the method is known. A wrong player is a player + method miss, not a conditional-method miss.
| Method model | Conditional hits / picks | Conditional accuracy | Brier / log loss | Player + method hits / graded |
|---|---|---|---|---|
| Loading method comparison… | ||||
No history is backfilled into this forward test. Unknown evidence stays unresolved; void games are excluded. Ties are not forced into picks. An outside-ten scorer counts as a player and combined miss, but is excluded from player probability scores. Lower Brier and log loss are better; no model is promoted automatically.
Behind the predictions
Model performance, watchlists, methodology, and historical research.