Guide
What the confidence score is, and isn't.
Prop Scanner's confidence score compresses recent form, matchup context, and market signals into a single 0-100 number. It's built to save research time, not to replace it — treat it as a starting point for review, not a final answer.
Open NBA ScannerWhat feeds into the score
The score blends several signals rather than any one stat in isolation. Recent hit rate carries the most weight, but matchup context, market edge, and role-stability signals like minutes trends all shift the number up or down.
- Recent hit rate
- The primary input — how often the player has cleared this type of line recently, across the L3/L5/L10 windows.
- Market edge
- The gap between the model's own projection and the book's posted line. A wider gap moves the score, but a very large edge is treated more cautiously — it often means the book has information the model doesn't.
- Matchup and role context
- Opponent strength, pace, home/away, back-to-backs, and recent minutes trends all factor in as secondary adjustments.
How to use it
A high score is a reason to look closer, not a reason to stop looking. It doesn't account for late-breaking news, a lineup change an hour before tip-off, or context specific to that night's matchup. Cross-check the score against the player's game logs and any injury or role news before treating it as settled.
- High score
- Signals are aligned in one direction. Worth a closer look at the underlying game logs, not an automatic conclusion.
- Low or mid score
- Signals are mixed or working against the line. Doesn't mean the outcome is unlikely — it means the model doesn't see a clean edge.
- What it doesn't capture
- Same-day news — a late scratch, a minutes restriction, a lineup change — arrives after the score is calculated. Always check current status before treating any score as final.