The Signal and the Noise: Why So Many Predictions Fail-but Some Don't

by Nate Silver

This is not to suggest that our priors always dominate the new evidence, however, or that Bayes’s theorem inherently produces counterintuitive results.