Statistical analysis: the language of prediction
Posted by Nathan Vegdahl {nathan_vegdahl[111AT22]yahoo[11DOT221]com} on November 22, 1999 at 08:42:48:

Hello!
I am presently researching prediction of events. I am not specificly researching erathquake or volcano prediction, but prediction in general. What I have discovered is that prediction works best in terms of statistics.
What that means is that instead of saying that "this will happen" or "that will happen" you say "this has this much of a chance of happening and that is this much of a chance of happening".
Also, you do not take the event with the highest chance of happening and say that that is what will happen. Instead, you include all possibilities with non negligible chances of occuring, and you include them in your prediction.
That is the problem with wheather channels: they always say the thing with the highest percent chance of occuring and give it's percentage, but then they neglect to tell you the other possibilities and their chances of occuring.
Prediction is not a "for sure" thing. You must present possibilities, and do not misrepresent them as certainties.


Follow Ups:
     ● Re: Statistical analysis: the language of prediction - Pat In Petaluma  20:08:25 - 11/22/1999  (901445)  (0)