Brian;

You're right, of course.

i'm trying to get around Amit's idea that if he's no better than chance, 5 out of 10 people should do better. Even with large samples, it seldom works out that way.

Roger