Re: How to determine conf interval about expected number based on probability?
From: Me (hlpme2004_at_hotmail.com)
Date: 08/30/04
- Previous message: Mark Von Tress: "Re: Finding Percentiles WITHOUT Sorting"
- In reply to: Richard Ulrich: "Re: How to determine conf interval about expected number based on probability?"
- Messages sorted by: [ date ] [ thread ]
Date: 30 Aug 2004 16:08:31 -0700
Richard Ulrich <Rich.Ulrich@comcast.net> wrote:
> Once you get a probability, you can place a CI
> on the theoretical outcomes based on the binomial
> or the Poisson.
Can you be more specific?
Consider that we have the probabilities p[i] for all
possible outcomes [1].
Sticking with strictly random behavior (without regard
for human idiosyncracies), how would we determine the
CI for each outcome for a given confidence level (aka
confidence coefficient)?
-----
[1] Consider the following game. Each player chooses
5 of 47 numbers. They match their choice against
5 of 47 numbers selected randomly at a drawing.
The outcomes are:
Match Probability
5 of 5 1 / 1,533,939 = p[5]
4 of 5 210 / 1,533,939 = p[4]
3 of 5 8,610 / 1,533,939 = p[3]
2 of 5 114,800 / 1,533,939 = p[2]
1 of 5 559,650 / 1,533,939 = p[1]
0 of 5 850,668 / 1,533,939 = p[0]
- Previous message: Mark Von Tress: "Re: Finding Percentiles WITHOUT Sorting"
- In reply to: Richard Ulrich: "Re: How to determine conf interval about expected number based on probability?"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
|
