Amit;

Ok, I reran your evaluation after updating the quake list and making sure predictions and quakes covered the same date range. The NEIC catalog only goes to August of 2012 for some reason so you had 175 testable predictions and only 31 hits.

Then I wrote a random prediction program, telling it to create 175 predictions covering the identical date range and assigning mags with the same distribution you used.

Finally I told it to run repeatedly, checking the number of hits each time and stopping when it got more hits than you did.

It took only 10 tries to get 35 hits.

Both you and the random selections are significantly below chance but the random selection did beat you.

If you object to 10 tries, it's the same as asking 10 people to make 175 predictions and finding one of them beats you.

Roger