What are Shapley Run Credits (SRC)?

Shapley Run Credits (SRC) are a player’s fair share of credit for every run scored by that player’s team. The “Shapley” label is used because SRC uses the Shapley Value from cooperative game theory for its calculation. The Shapley Value formula is a method for dividing credit among individuals involved in a collaborative team production that satisfies several mathematical fairness conditions.

The interpretation of SRC is very simple: a player’s SRC is the number of runs they deserve credit for creating. More precisely, it is the portion of the team’s runs for which that player deserves credit based on how their offensive contribution worked in collaboration with their teammates. Suppose a team scored 2 runs in an inning and that SRC divides the credit for those 2 runs such that player A is credited with 1.50 SRC, player B is credited with 0.50 SRC, and all other players are credited with 0 SRC. Then the Shapley Value calculation determined that player A deserves credit for 1.50 runs out of the 2 runs scored while player B deserves credit for only 0.50 runs of the 2 runs.

An important property of SRC is that it fully allocates credit for runs so that the sum of SRC for the players on a team exactly equals the total runs scored by that team. If the team scored 2 runs, then the players’ SRC add to 2. If the team scored 5 runs, then the players’ SRC add to 5.

Another property of SRC is that it is a counting statistic that can be added across innings, games, series, and seasons while retaining the same interpretation. If player A has more SRC in a series than player B, then player A contributed more to scoring in that series than player B.

The most distinctive property of SRC is that it allocates credit based on how well the player worked with their teammates to score runs. That is, SRC assigns credit based on the actual collaboration between teammates so that, for example, a double might result in more SRC in one inning than in another based on the different offensive contributions of the other teammates in those two innings. This feature makes SRC fundamentally different from other statistics that attempt to remove the player from the collaborative context. For this reason, SRC is a backward-looking statistic that reveals new insight into how to assign credit for what actually happened in a game, not a predictive statistic.