NBA model. Efficiency, pace, rest.
Pace-adjusted offensive and defensive efficiency, rest and back-to-back fatigue, and market de-vig. The NBA model is currently in shadow build — it is being validated on historical data before any pick is published, and this page is honest about that.
NBA is in shadow build — no published picks yet
The NBA model is still being validated on historical seasons and is not yet firing published picks. We would rather show you nothing than publish a pick we have not proven out. When it clears validation and the season is live (October), the slate and graded reads will appear here and settle on the public ledger like every other sport.
Get notified when NBA goes live →The model rates every team by pace-adjusted offensive and defensive efficiency, then adjusts for schedule context — rest days, back-to-backs and travel — before converting a projected margin and total into a fair no-vig price to compare against the market. It runs in shadow today: producing reads that are graded internally so we can measure them honestly before anything goes public.
| Model input | Coverage | What it captures |
|---|---|---|
| Pace-adjusted efficiency | Rolling | Off/def rating per 100 |
| Rest & back-to-backs | Per game | Fatigue the market underrates |
| Pace / possessions | Modeled | Total and tempo context |
| Market edge (de-vig) | Shadow | Fair price vs the line |
| Validation status | In build | Not yet published |
Every pick is tagged by confidence so you always know what you are looking at — and only ◆◆◆ and up are published.
| Window | Record | Win % | Units |
|---|---|---|---|
| All sports · since Apr 15, 2026 | — | — | — |
NBA-specific settled results join the ledger once the model fires in-season. Every pick is settled publicly — wins and losses. See every pick settled publicly →