Loading page…
What does ‘best right now’ mean?
A current-quality estimate needs more than a single box-score season.
UB Value turns observed per-minute impact into a season-level value score by accounting for playing-time exposure. It deliberately stops short of predicting what a player will do next.
The question
How should a ranking separate per-minute impact from the season value created through playing time?
Context
A per-minute metric and a season-value metric answer different basketball questions. UB Impact describes observed quality per minute; UB Value also accounts for how much a player was on the floor.
Data
- Eligible NBA player-season UB Impact from the published player-statistics pipeline.
- Games played and minutes per game for playing-time exposure.
- The season population used to estimate replacement level.
Method
The analysis follows the method path shown below; each step is a boundary on what the note can establish.
- 01Replacement level is the 15th percentile of eligible UB Impact values.
- 02Impact above replacement is scaled and multiplied by minutes played relative to a 2,000-minute baseline.
- 03The snapshot writer ranks only finite value_score values and keeps UB Impact separate.
What the method establishes
- The resulting value_score is the canonical rank key for the Beyond the Box 100.
- The method does not produce a future-performance estimate, a wins total, WAR, or a confidence interval.
- Players without a value_score remain unavailable for the Overall ranking.
Interpretation
A higher UB Value means more modeled season value under this exposure-weighted method. It does not mean the player is guaranteed to remain there, or that the number isolates every part of basketball value.
Limitations
- The current implementation is built from descriptive impact features; tracking, deterrence, and every lineup interaction are not represented.
- No historical walk-forward validation is published, so the model remains experimental.
- The value scale is not comparable to WAR, wins, or salary dollars.