Guide

How hit-rate windows work — and why the window you pick matters.

A hit rate is only ever a hit rate over some window of games. Prop Scanner shows last-3, last-5, and last-10 windows side by side because each tells a different story, and none of them is automatically correct.

Open NBA Scanner

What each window shows

L3 reacts fastest to a player's current role, minutes, or health. L5 balances recency with enough games to smooth out a single outlier. L10 is the most stable read of a player's underlying level, but can lag behind a real change in role or workload.

L3 (last 3 games): The fastest-reacting window. Best for catching a recent role change — a new starting spot, an expanded shot count, a return from a minutes restriction — before the larger windows catch up. Also the noisiest: three games is a small sample, and one blowout or foul-trouble night can swing it heavily.
L5 (last 5 games): Prop Scanner's default recency weighting leans most heavily on L5. It's close enough to current form to matter, but has enough games to dampen single-game noise.
L10 (last 10 games): The most stable of the three, and the best read on a player's underlying baseline. Weight it more when a player's role has been unchanged for a while; weight it less right after a trade, injury return, or lineup shakeup.

Reading windows together, not alone

The three windows agreeing is a stronger signal than any one of them alone. When they diverge sharply, that's usually a sign something changed recently, and it's worth checking the game logs and current role context before trusting the number.

All three agree: The strongest read — the player has been consistent regardless of which window you pick.
L3 diverges from L5/L10: Something changed recently. Check for a role change, injury, or new starting role before trusting the L3 number on its own.
L10 diverges from L3/L5: The longer-run baseline includes games that may no longer reflect the player's current situation — a prior role, an injury absence, or a different team context.