Accuracy and retaliation in repeated games with imperfect private monitoring: Experiments
通过实验研究不完美私人监控下重复囚徒困境中的策略选择，发现多数人采用慷慨的以牙还牙策略，但实际报复行为与理论预测不符，低准确度者报复强度低于预测，高准确度者则高于或等于预测。
We experimentally examine repeated prisoner's dilemma with random termination, in which monitoring is imperfect and private. Our estimation indicates that a significant proportion of the subjects follows generous tit-for-tat strategies, which are stochastic extensions of tit-for-tat. However, the observed retaliating policies are inconsistent with the generous tit-for-tat equilibrium behavior. Showing inconsistent behavior, subjects with low accuracy do not tend to retaliate more than those with high accuracy. Furthermore, subjects with low accuracy tend to retaliate considerably with lesser strength than that predicted by the equilibrium theory, while subjects with high accuracy tend to retaliate with more strength than that predicted by the equilibrium theory, or with strength almost equivalent to it.