Instead of computing EPV as a single black-box number, the value of any possession action can be decomposed into three separable components: (1) probability of attempting that action type (pass/carry/shot); (2) probability of success if attempted; (3) expected value of the state after a successful attempt. This decomposition enables attributing value changes to specific causes — a player may choose the right action type but execute poorly (component 1 right, component 2 wrong), or may attempt a high-success action that goes to a low-value location (component 2 right, component 3 wrong).
For any action: compute P(action type) from the distribution of historical action choices in that state; compute P(success | action type) from historical completion rates in that state; compute E[value | success] from the historical EPV of destinations for that action type. The product gives expected action value; comparing to what actually happened reveals whether value was added or lost in each component. This is the foundation of action-level attribution — who made good decisions, who executed well, and whose choices led to good outcomes.
A player making consistently poor decisions but getting lucky outcomes will be rated well by any outcome-based metric. When luck regresses, performance collapses "suddenly" — but decision quality was always poor. Decomposing value into decision, execution, and outcome quality catches this before regression.
Possession value decomposition reveals that the highest-EPV-delta action in a goal-scoring possession is often the 3rd or 4th action from the end, not the assist or the shot. A press-breaking carry that advances 30 yards often has a higher EPV delta than the final through ball, because it shifted the entire possession from low-value to high-value territory. Credit assignment that weights only the terminal actions (goal, assist, key pass) misses the player who actually created the scoring opportunity.