If they're taking an average over the Earth's entire history, and the actual distribution of impacts over time is not uniform (as seems likely), that might expl

If they're taking an average over the Earth's entire history, and the actual distribution of impacts over time is not uniform (as seems likely), that might explain the discrepancy. Though in that case it seems odd that they would use that average to predict a probability going forward.