Author: Rick Bischoff
Date: 15:40:14 10/07/04
Go up one level in this thread
>>You set your criteria before hand. i.e., do you want to be >>95% confident that you have the right answer? 99% confident? Then, you decide >>on the proper test to use and the sample size. > >Sorry, you can't decide the confidence level before you do the statistical test. >It's output from the statistical test. No, that is faulty; you can't decide on the confidence INTERVAL (for cetain types of tests) until you do the test, but the level should be decided before the test. Your test, in a sense, was doing exactly this-- you said yourself, if the variable "t" is outside of the interval (-2,2) you can be 95% sure that the means were different. Anyway, the point is moot since the basic assumptions are flawed. 1. You drop draw scores-- you should be either rolling them into wins or losses or using a multinomial distribution model. 2. You did not state that which distribution you believed the number of wins - number of losses to follow. 2a. Your sample is not random, although if this continues to be a picking point, I will drop it as the rest of my claims are valid and can't be refuted. 3. Your transformation of variables and your test on the statistic "t" is not valid-- you might be assuming a normal model, but I can't know that since you didn't say so. Nevertheless, even if you did, I can show you that your transformations do not lead to a distribution that can be safetly approximated by the normal distribution.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.