Imagine you are investigating a bug you find under rocks. You have a prediction about the subvarieties this species of bug comes in. But these bugs are really h

Imagine you are investigating a bug you find under rocks. You have a prediction about the subvarieties this species of bug comes in. But these bugs are really hard to find. You have to turn over 10^21 rocks to find just 1,959 bugs. And you find unexpected ratios of subvarieties! The statistical significance of this unexpected result is based not on the number of rocks you turn over but on the probability your theory assigns to the distribution discovered in your sample of 1,959.