If forced to interact a given number of times, beliefs about other agents seem not to depend on the order of observations but on the number of each kind of observation.
Experiment Description.
Repeat for a number of trials.
2 agents play a tournament, but before each tourney:
one agent's first moves are pre-programmed
the other agent gets to act in between these observations.
plot (A's beliefs about B and B's about A) vs K vs Rounds vs Observation sequence.
plot fitness as well.
left column is A's beliefs about B, right is other way around.
B is the preprogrammed agent.
Left graph does not give A a chance to react to B's initial actions.
If forced to interact a given number of times, beliefs about other agents seem not to depend on the order of observations but on the number of each kind of observation.
Experiment Description. Repeat for a number of trials. 2 agents play a tournament, but before each tourney:
plot (A's beliefs about B and B's about A) vs K vs Rounds vs Observation sequence. plot fitness as well.