he makes the classic error

For what it's worth, I interpreted his "there is a 26% probability this occurs by chance" exactly as "if there's no real difference, there's a 26% probability of getting this sort of result by chance alone" or equivalently "conditional on the null hypothesis Pr(something at least this good) = 26%". I'd expect that someone who was making the classic error would have said "there is a 26% probability this occurred by chance".