Author: Uri Blass
Date: 05:01:43 08/25/01
Go up one level in this thread
On August 25, 2001 at 07:46:20, Sune Fischer wrote: >On August 25, 2001 at 07:06:31, Uri Blass wrote: > >[snip] >> >>The competition is not to answer which program is the best but to decide about >>the world champion. > >Then what's the definition of a world champion? The definition of the world champion is the program that wins a tournament that is called the world championship. It does not have to be the best program and there were a lot of cases in the past when the winner was not the best. > >>The question which program is better is also not clear even on uniform platform >>because it may be dependent on the hardware or the time control and it is >>possible that at blitz on p200 program A is best and on tournament time control >>on dual P1000 program B is best. > >That is correct, a new time would mean a new category and thus a new tournament. >I see no reason that the blitz champion should also be a 2 hrs/40 moves >champion. >The time and hardware should be agreed upon well in advance so the programmers >can tweak it to those specific requirements. >I don't see a problem with that, as long as it is the same to all participants. >Otherwise the winner will most likely be the one that can afford the most >powerfull hardware. > >>If your program can use the cray when Deep Junior cannot and using the cray help >>your program to be better than Deep Junior on the best hardware that Junior can >>use than it means that you can claim that your program is better than Deep >>Junior because if I have >>1,000,000$ for buying a chess hardware and chess software and I want to buy the >>best program then I am going going to prefer your program+cray and not Deep >>Junior. > >And what if you only have a single CPU system and what to get the best program, >would you get Shredder or Junior? Which of those are the strongest on the single >platform? This is where is where I see the problem, it's how people will draw >the wrong conclusion; They will think that the winner (e.g. Shredder) is the >best program for a simple CPU system. No I believe that Deep fritz or Junior7 are better for a simple CPU system. They have forgotten or not heard that... >well it was running on faster hardware, and it didn't actually beat Fritz and >Junior, and..... >Conclusions become so hard to make. > >I think what people generally want to know is: what prog is the best on my home >computer (say an Athlon 1 GHz with 256 SDRAM). Then it is better if they use the ssdf results and not the world championship results. >A common platform like that would tell a lot more about the general strenght of >the programs, doing a comparison on completely different hardware is just simply >nonsense. playing only 9 rounds is also nonsense and there should be hundreds of rounds in order to know. Playing only 9 games for programs is the main reason for errors and not the difference in hardware. Another reasons for errors is the fact that some programs usually do better opening preperation for specific opponents that are not relevant for the customers. Uri
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.