But the situation where P(simple hypothesis) > P(complex hypothesis) is not equivalent to the claim P(one hypothesis from the set of simple hypothesis) > P(one from the set of complex hypothesis).
This is because the set of complex hypothesis is infinitely larger that the set of simpler ones, and even if each simple hypothesis is more probable, the total probability mass of the set of complex hypothesis is larger.
It means that any situation is more likely to have complex explanation than simpler one. This could be called anti-Occam prior. Not sure if it is related to what EY meant.
But the situation where P(simple hypothesis) > P(complex hypothesis) is not equivalent to the claim
P(one hypothesis from the set of simple hypothesis) > P(one from the set of complex hypothesis).
This is because the set of complex hypothesis is infinitely larger that the set of simpler ones, and even if each simple hypothesis is more probable, the total probability mass of the set of complex hypothesis is larger.
It means that any situation is more likely to have complex explanation than simpler one. This could be called anti-Occam prior. Not sure if it is related to what EY meant.