Active Player Modeling in the Iterated Prisoner’s Dilemma

Table 2

Player types.

Player type	Description
AllC	(i) Always cooperate
CCD	(i) Repeat , and
TFT	(i) Mimic opponent’s previous action; (ii) The first action is cooperation
NoisyTFT	(i) Almost the same as TFT; (ii) Action changes with a 10% probability
TF2T	(i) after two opponents consecutive cooperation; (ii) after two opponents consecutive defections
ATFT	(i) Adaptive TFT (, ); (ii) If , then or otherwise
ADP	(i) Test , , , , and , , , , when game starts; (ii) Choose an action based on payoff
Major	(i) Follow opponent’s major actions during the entire game; (ii) The first action is cooperation
Major 5	(i) Similar to Major; (ii) Considers only the five previous actions
FS	(i) Defect with probability p (initial value 0.5); (ii) is probability for opponent defection
Pavlovian	(i) Cooperate if my action was the same as that of my opponent in the last turn; (ii) Otherwise, defect
FTR3	(i) Use a , and pattern to recognize opponent’s strategy; (ii) If opponent’s response is the same as the pattern, cooperate