Win Stay, Lose Shift Or Imitation – Only The Choice Of Peers Counts
Win Stay, Lose Shift as well as imitation strategies for iterated games rely on an aspiration level. With both learning rules a move is repeated unless the pay-off fell short of the aspiration level. I investigate social adaptation mechanisms for the aspiration level and their impact on the efficiency of learning in a large population of agents that repeatedly play one round of a symmetric 2x2 game against randomly chosen opponents. It turns out that if the aspiration level is given by the last payoff of the current opponent the population receives the maximal symmetric payoff of the game in the long run. If the aspiration level is determined by independently chosen agents the outcome is related to the evolutionarly stable strategies. This holds for win stay, lose shift as well as for imitation based learning. These results suggest that the choice of peers can be crucial for the efficiency of learning.
When requesting a correction, please mention this item's handle: RePEc:vie:viennp:0109. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Paper Administrator)
If references are entirely missing, you can add them using this form.