Author: Robert Hyatt
Date: 08:51:51 09/06/05
Go up one level in this thread
On September 05, 2005 at 10:58:22, Uri Blass wrote: >On September 05, 2005 at 10:40:28, Rolf Tueschen wrote: > >>Significance means without reasonable doubts, Uri. ;) > >I understand but not significant results by that definition still may be >significant for people or in other words important for people. > >I think that information that does not give significant results for >the question which program is better may be still important for the customers >and it is better than no information. > >Uri That's a _dangerous_ attitude. As it is just a short step from saying "disinformation is better than no information" and that certainly is not true... I "little information" can lead to conclusions that are wrong. And someone without the necessary experience can draw the wrong conclusion. For example, you walk into a casino, sit down at a blackjack table (my favorite game) without knowing how to count cards. That is a _losing_ game for you, but you can win or lose most any amount short-term because of statistical variance. Long-term, you are going to lose, period. But if you tell someone with no experience, "Hey, I studied these basic strategy cards, played and won $500 in two hours" then off they go. And they get busted. Had you waited until you had played a hundred hours (or more) before reporting your results, you would not have been reporting that you had won, because you would not have won. But short-term, standard deviation allows strange results that can be misleading to the uninitiated...
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.