Author: Thorsten Czub
Date: 15:39:06 01/27/04
Go up one level in this thread
On January 27, 2004 at 16:30:44, Thomas Mayer wrote: >You know that this is wrong... no - why is this wrong. you play 30 or 40 moves within a tolerable amount of time. and then the game comes to the point where the result is more or less random because there is not enough time for the programs to hold the position in the stage it was before they had no time trouble. playing blitz games is like a tree in the wind. first the tree fights the wind. but then there comes a point (of no return) when the tree cannot fight the wind anymore and it crashes. blitz games are like this. and you can forget the 30-40 moves before. THERE you measured chess strength. but when it comes to time problems, you do not measure strength. this is the reason people like fischer, who are interested in strength, invented the fischer-bonus. to make the game data more senseful. >playing with no increment makes sometimes sense when the time control is big enough that the decision of the game is done BEFORE both sides have severe time problems. >and it tests time usage of the engine. and is this telling us anything about the strength when a chess engine moves faster than another ?? >It's just another way to test things. IMO its a senseless method. therefore i almost never used it. no matter which hardware i had, i normally gave them 40/40 for the first overview and 40/120 to get more reliable results. >Period. :-)) >And you should apology for claiming Sarah not to show all her results. But I >know, you will never do... you don't understand. SHE asked for ALL games in the postings BEFORE. my request for all of HER data was just a repetition of her own sentences. there is no need for any apology. because i never mistrusted her . she is not cheating, only her results are not true. >aha, so 3+2 fullfill this ??? LOL... 3+2 e.g. is much better. it gives the programs more time overall. and the games do not have so many weak moves towards the end of the games. you can laugh about this. but i do prefer game material that makes sense to replay. i can ENJOY senseful data much more than senseless data. why shall i replay games with the program that it would have never moved if it would have had more time. I don't want to gamble. I want to measure something. therefore i test the thing within a certain range. i am not interesting what happens OUT OF THIS RANGE. even when i play 40/40 games on my 400 mhz AMD machines the data makes more sense because they are not running against the time control as in blitz games. blitz games are IMO for people who like the situation that a clock runs out. it's a fun thing. action. but for measure purposes i would advise to use at least fischer+2" or more increment.
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.