Actually, all it would do is multiply everything. Remember, you only count hits and misses. So in your 100 quake example if you had a 1 day window incremented by minutes you would have 1440 hits instead of 1 or 2 and the total would multiply by 1440 as well. You wouldn't bell up any; you'd just have a bunch of hits as the sliding windows passed over the clusters. The partial window problem at the end would be compounded as well, with the final window getting smaller and smaller.

I see no merit in this.

David is upset that we didn't take his suggestions. Tough.

Roger