[Roger Hunter]

(02-13-2015, 11:20 AM)Island Chris Wrote: Wait a minute Roger. The way you worded that comes across reads as completely unfair. I have a proposal due Tuesday so have not gone through Duffy's recent posts, but if there really is a M6+ quake within a day or less of EVERY signal, then that would be pretty spectacular. That is because M6 quakes are not evenly distributed in time, and it is pretty common to go quite a few days with no M6+ quakes

The way to proceed is to plot a list of signals again a list of M6+ quakes on a graph. It would be pretty clear visually if there is something going on. If it looks like there is a relation qualitatively, it is probably pretty simple to get the statistical significance: no need to make it complicated, just 2 rows of numbers, I don't know how to do this, but Brian or you probably do.

I will read Duffy's posts eventually.

Chris

You're right and I apologize. I'm recuperating from cataract surgery so didn't look deeply enough.

The actual odds on finding a quake in a 2 day window chosen at random from that year are slightly less than 50% so 7/7 is better than expected. I'll get the significance next.

Later....

The odds on getting 7 hits in 7 tries with odds of 0.497 are 0.007 which passes my 99% test. This would make Duffy a winner under proper test conditions. Since those conditions were not met, I suggest a formal test is in order.

What's needed is a larger sample, posted in advance to rule out cherry picking. If that test passes, Duffy needs to be seeking grant money!

Roger