Value Added

What it measures

A star who plays twelve games and a star who plays eighty are not equally valuable; a rate metric cannot tell them apart.

How to read it

Higher is better. Zero means a freely available player would have produced the same.

Why it exists

The impact index is a rate. It is a within-season z-score composite scaled to 0-100, and it describes how good a player was per minute while deliberately saying nothing about how much they played. That is the right shape for judging quality and the wrong shape for judging a season. Value Added closes that gap by multiplying quality above a replacement baseline by playing time, so availability counts for something. This is not WAR. It is not comparable to Basketball-Reference VORP, to FiveThirtyEight WAR, or to any published wins-above-replacement figure, and the scale is not chosen to resemble them. Two of its choices are judgment rather than derivation - replacement level sits at the 15th percentile of players with at least 15 games and 10 minutes per game, and playing time is normalised against a 2,000-minute starter season. Both are defensible and both are arguable; moving either shifts every number on the scale. They are stated here rather than buried so that disagreeing with the metric means disagreeing with a stated choice. Because the impact index is z-scored within each season, Value Added is era-adjusted by construction: a 2004 season and a 2024 season are each measured against their own league.

Limits

• Not WAR. Not comparable to Basketball-Reference VORP or FiveThirtyEight WAR.
• Replacement level (15th percentile) and the 2000-minute baseline are chosen, not fitted.
• Inherits every limitation of the impact index it is built from.
• Box-score derived, so it cannot see defensive positioning or off-ball value.
• The observed league spread is narrow; small differences between players are not meaningful.