Posted by Roger Musson on April 21, 2001 at 07:53:32:
I'll try to answer several things at once: Canie: yes, about half the posts just came up as random bits of html. But you are still missing the point, it doesn't MATTER how you define a near miss; the important thing is that the pursuit of hits per se is a snare and delusion. What we are looking for is significance. If I predict a magnitude 4.0 earthquake in S America tomorrow, it doesn't matter whether it's a hit or a near hit or whatever. It's trivial. What is needed is not a measure of "hit rate" but a triviality evaluation. Roger: my method is published - Seis Res Lett, vol 68 no 6 pp 944-946. Where's yours? Who's peeking at whom? Yes, you do need multiple predictions to evaluate anything. As one of the Bond villains philosophises: "once is happenchance, twice is coincidence, three times is enemy action".
|