Under the hood: the models & the results board
The live edge: game state vs a slower line
5 min
The moneyline model works before tip-off. Once the game is live, a different engine looks for in-play value — the "live edge" — on totals (over/under).
The mechanic: pace extrapolation
The idea is simple arithmetic. Take the points scored so far, divide by the minutes played to get a scoring pace, then project that pace across the minutes remaining:
projected total ≈ current points × (total minutes ÷ minutes elapsed)
If that projection sits far enough away from the bookmaker's live over/under line — outside the model's confidence band — FinalSkore flags an edge: the market line looks slow to react to how the game is actually flowing.
Why tracked live picks are Q4-only
Early in a game the projection is jumpy: a hot first quarter extrapolates to an absurd final total. The more minutes played, the more reliable the pace estimate becomes. So while the app displays live reads from the second quarter on, only fourth-quarter picks are counted on the results board — that's the deliberate quality gate. Pre-Q4 reads carry a transparency note telling you they aren't tracked.
How to use it
- A live edge is a temporary mismatch between a fast-moving game and a slower-moving line — not a lock.
- Linear pace assumes the game keeps flowing as it has; fouls, garbage time and blowouts break that assumption late.
- Treat the Q4 picks as the serious signal and the earlier reads as context.
The live edge is pure extrapolation from the pace so far. It sharpens as the clock runs down — and it can still be wrong when the game's rhythm suddenly changes.