Fig. 1: The leading eight norms with quantitative assessment.
From: Quantitative assessment can stabilize indirect reciprocity under imperfect information

a We consider the leading eight norms28,33. Each norm consists of an assessment rule that determines how an observer updates a donor’s reputation, and an action rule that governs players’ behavior when they are chosen to be the donor. The assessment rule takes the context of an observed action into account: how an observer judges a donor depends on the donors’ and recipient’s reputation. Similarly, the action rule uses the current donor’s and recipient’s reputation to decide whether the donor should cooperate with the recipient. In contrast to the original baseline model, we now interpret a positive assessment of an action as an increment of +1 to the donor’s reputation score, and a negative assessment as a decrement of -1. b The observing Player Z assesses Player X’s action of cooperating with a bad player as bad, such that he decrements X’s previous score by one. c, When it is Player Z’s turn to be the donor, he translates his and Player X’s reputation score into a binary label of “good” or “bad’’. Since both his and Player X’s score are above the threshold S, he judges both himself and Player X as good, and cooperates. d, e In the baseline model using binary assessment, the same starting scenario ends differently: Player Z changes his view of Player X from good to bad after Player X cooperates with Player Y, and therefore defects against Player X.