When I read Fooled by Randomness I found it useful. Not groundbreaking work, but it drew some nice analogies between statistical distributions and human's over-certainty.