Michael;

It's the time taken up by the prediction window. If you make one prediction per month, each a week long that's 12 weeks or 84 days out of the year so the odds are 84/365 = 0.23 or 23%.

So you can expect to be right 23% of the time by chance alone. The problem is with clusters. If you hit one your success goes way up in terms of the number of quakes captured. But that's misleading. And if you miss a cluster the reverse effect makes you look bad.

Roger