Imagine I argue that, to sort an array of n integers, I must ask myself "is this permutation the sorted array?" for every n! permutations of the integers. Thus

Imagine I argue that, to sort an array of n integers, I must ask myself "is this permutation the sorted array?" for every n! permutations of the integers. Thus it is obvious that humans cannot correctly sort a list of 100 numbers, because even if I could correctly determine whether a given array of 100 numbers is sorted in a single second, it would take 9.3 × 10^157 seconds to sort such an array. This is clearly silly, because I ignore all the other algorithms for sorting such an array--I assume that the only available option is something comically inefficient. In just the same way, when Steve Keen argues that it's impossible that my purchases are rational because choosing rationally requires me to iterate through every possible bundle and select the best one, he is making big assumptions about the type of problem I face, the tools available, and the meaning of rationality. > And yet you - or someone, inevitably - is also going to claim that markets are somehow magically "efficient" precisely because they're capable of usefully approximating exactly this kind of optimisation. As others in this thread have explained, question of whether markets usefully approximate this kind of optimization is distinct from the question of whether individual people perform this optimization successfully. Deviations from efficient markets absolutely occur--the question is what my priors should be. The fact that, for example, it's extremely difficult to reliably identify and profit from financial market inefficiencies suggests that, in fact, markets really do behave astonishingly well, given all that we know about human irrationality.