It's probably worth looking at optimism across multiple independent processes. As a proportion of human interaction, I would wager that processes which are per

It's probably worth looking at optimism across multiple independent processes. As a proportion of human interaction, I would wager that processes which are percievably stationary (or stationary enough), and that a human will interact with over and over again is actually relatively small. I would hypothesize that had this task been set up with 'moving parts', then optimists could rationalize away their expectations as 'this time it's different'. I think ultimately, optimism 'works' not because it influences the probability of events happening, but because it influences the distribution of 'payoffs'.