Quote:
Originally Posted by Matt L
Kuan, the distribution will be anything but "normal." I think you mean "equiprobable," if that's even a word.
Also, what you mean to say (I'm sure) is not that the spot will be hit once in every 35 turns, but that in the long run, the spot will be hit 1/35 of the time.
In theory, you will eventually win against any nonzero odds with probability one. But of course, not only will you not likely live long enough to achieve this, but the lottery in question probably won't either.
Let me add that probability one does not mean the same thing as "it will happen." It means that the probability measure of it "not happening" is zero. These are close, but definitely not the same thing.
|
I love this stuff!
I spend a lot of time trying to partial-out the statistical errors in spatial analysis of combined events. I do the stuff on the computer and then go out and look at what the analysis means when applied to what's real and on the ground. There are just so damned many factors that are inconsequential most of the time but under extraordinary circumstances, become significant, and some unpredictable (and unknown) number of them are contributors. Just stick them all under eta and be done? That works at the first approximation but when you start partialing-out the contributing factors life gets insanely complex.
I am increasingly of the belief that the tools used are insufficient because the underlying theory is insufficient. Somewhere there is a 10 year-old whose mind in 10 years, will encompass the problem and develop a better theory of probability that works with pathological conditional probabilities. I hope I live to see the day. For me, it would be as important as seeing the first Moon walk.
B