Your model assumes a constant effect in each iteration. Is this justified?
I would envisage a constant chance of recovery and an asymptotically declining estimate of recovery. It seems more realistic, but maybe it’s just me?
It’s a toy case, in reality the chance of recovery might be “0.2+0.3*estimate”, but the same general reasoning applies and the end result is still regret of rationality.
Your model assumes a constant effect in each iteration. Is this justified?
I would envisage a constant chance of recovery and an asymptotically declining estimate of recovery. It seems more realistic, but maybe it’s just me?
It’s a toy case, in reality the chance of recovery might be “0.2+0.3*estimate”, but the same general reasoning applies and the end result is still regret of rationality.