Felix Salmon (via Andrew Gelman):

Many if not most of my opinions are wrong (although of course I have no idea which they are), and … many of the most interesting and useful things I write come out of my being wrong rather than being right. This is not, as Wilkinson noted to Cowen, an easy intellectual stance to hold: he calls it “a weird violation of the actual computational constraints of the human mind”. But I think it’s undoubtedly worth working on.

This makes me feel better about the surprise and then ensuing guilt I experienced the first time one of my published results replicated.