Hi all;

One of my programs evaluates the significance of a group of predictions with different predictions. It uses the individual probabilities to reward hits and punish misses. The final answer is the number of standard deviations above or below chance.

I ran the great quakes thru this program even though they all have the same probability to see if JOB had enough hits to count.

He does, by a wide margin. This doesn't make sense to me. 195 hits out of 3364 windows shouldn't be significant. But a hit gets 3.5 while a miss is only -.03 and the final answer is VERY high.

Roger