Author: martin fierz
Date: 04:12:02 09/11/04
Go up one level in this thread
On September 10, 2004 at 12:05:30, Peter Berger wrote: >On September 10, 2004 at 09:49:41, martin fierz wrote: > >>On September 10, 2004 at 09:37:18, Jorge Pichard wrote: >> >>[snip] >> >>>I'm not the only one who post matches in progress....... >> >>just one more thing: you are conducting ONE experiment: how does the opening >>book influence the match result of jonny-movei. you have posted about a zillion >>(ok, i'm exaggerating, i know) intermediate results, with cross-references all >>around. it is very hard to follow such a flood of posts. >>here's a constructive suggestion: wait until you have all matches finished and >>then publish the results: >> >>jonny - movei: >> >>book A: 16-4 >>book B: 11-9 >>book C: 5-15 >> >>that is about all information that is needed. one click, a 10-line-post, and >>that's it, people interested in this experiment will be grateful that they don't >>have to look up all the information in 10 different posts. >> >>i'm not saying that your experiment is useless in itself. the influence of >>opening books is an interesting question. but if you play too little games (and >>it doesn't matter whether that is 10 or 20...) it's just tossing coins... >> >>cheers >> martin > >I don't think this is true. It looks like a pretty safe conclusion with your >above results, all else being equal, that book A is superior to book C. I know >others have done programs to simulate these kind of questions in Monte Carlo >experiments, it would be interesting to know the level of certainty - I don't >think that more games are needed at all, given the above results. > >Peter those were just made-up numbers. jorge's results (which he did post in user-friendly form, thanks!) are much less clear - and therefore i stick to my opinion that 20 games are rather too little. cheers martin
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.