Ara;

> Not sure if I am getting this right. After 10,000 repetitions, "chance" did no better than 40. Don only got 33. So how is it that "chance" lost? Are you assigning "chance" the median score of 30 and saying that Don's score is higher than that? If so, how do you avoid the temptation to say that Don's score is only slightly higher "by chance"?

No, you're not getting it right. I made 10,000 repetitions of 226 random numbers, comparing them to Don's probabilities and counting the number of times the random number was smaller than Don's. Those counts made up a histogram and the largest number was 40 Don's 33 is 3 sigmas from center.

> Did you repeat each of 226 predictions 10,000 times?! That would be 2,260,000 repetitions. No wonder Chris is getting dizzy :)

Yep. Took several seconds. Want to check them by hand?

> Then again the term "random prediction" is a little misleading -- probably you are assigning Don's predictions to different time-windows, randomly. Not exactly starting from scratch and making a "random prediction."

Yes, it's really a different process. Sloppy language.

> Well, I'm on the sidelines, waiting to hear what Dr Jones' brother thinks.

Me too. I'm hoping it's valid because it fills a need.

Roger