“Quand la théorie des jeux vous donne des indices sur les comportements du quotidiens…
Voici une stratégie optimale pour une interaction, elle ressemble à oeil pour oeil dent pour dent (qui est une des stratégie optimale, mais qui a deux problèmes : il y a des erreurs de mésinterpretations & une fois qu’une “dent” est déclenchée, une spirale se déclenche) :
“Start cooperation and continue to do so until one of the four tests below fails.
1. First impression: A defection on the first move is unacceptable. Revert for tit-for-tat.
2. Short term: Two defections in any three turns is unacceptable. Revert for tit-for-tat.
3. Medium term: Three defections out of the last twenty periods is unacceptable. Revert for tit-for-tat.
4. Long term: Five defections out of the last one hundred periods is unacceptable. Revert for tit-for-tat.
On the first violation return to cooperation after twenty periods of the tit-for-tat echo. But put the other side on probation. Reduce the number of defections allowed in the medium- and long-term tests by one.
If the other side does not violate the probation for fifty periods, then strike the record clean and return to the original standards. If the other side violates the probation, resort to tic-for-tat forever.
The exact rules for first, short term, medium term and long term impressions will depend on the probabilities of error or misperception, the importance you place on future gains and current losses, and so on.”
”