Author: Ed Schröder
Date: 00:55:24 07/12/00
Go up one level in this thread
On July 12, 2000 at 02:31:05, blass uri wrote: >On July 12, 2000 at 01:13:21, Ed Schröder wrote: > >>On July 11, 2000 at 23:39:47, Christophe Theron wrote: >> >>>On July 10, 2000 at 18:55:11, Bruce Moreland wrote: >>> >>>>On July 10, 2000 at 14:56:31, Christophe Theron wrote: >>>> >>>>>On July 10, 2000 at 14:15:39, Terry Ripple wrote: >>>>> >>>>>>I know Hiarcs7.32 is one of the most knowledge based programs, but what about >>>>>>the famous Shredder4, Rebel Century and Junior6? Where do they average on >>>>>>knowledge in comparison? >>>>>> >>>>>>Best regards, >>>>>>terry >>>>> >>>>> >>>>>For me the answer is easy. I know others will disagree... >>>>> >>>>>The program that knows the most about chess, and has the most relevant >>>>>knowledge, is the one that stands the highest in the rating lists. >>>>> >>>>> >>>>> Christophe >>>> >>>>I don't think this is a very good definition, although nobody ever defines it >>>>when they ask which program has the most. I think that what people mean when >>>>they say "knowledge" is that a program with more knowledge plays more like a >>>>strong human and less like a typical computer, since everyone would agree that a >>>>GM human typifies "knowledge". >>>> >>>>People want "knowledge" in a program because they think they can learn from >>>>seeing it expressed. They want knowledge because they want to ask questions of >>>>something knowledgeable and get answers. >>>> >>>>This has little to do with which robot whacks the other robots. Computers can >>>>play chess in a vacuum, if desired. It's an interesting thing to do, and lots >>>>of people are interested in doing it. But other people are interested in >>>>interacting with the program themselves. >>>> >>>>bruce >>> >>> >>> >>>I did not say "computer rating lists" but just "rating lists". >>> >>>My definition is the most explicit and the closest to what a mathematical >>>definition could be that I have ever heard. >>> >>>Can you give a better definition yourself ? >>> >>>Who is going to argue that the program that has the best knowledge about chess >>>is the program that wins more games than the other ones ??? >>> >>>What other way of measurement are you thinking about ? >>> >>>If we were talking about humans, wouldn't you agree that the player who has the >>>best knowledge about chess is the one that wins more games? >>> >>> >>> >>> Christophe >> >>I think that comp-comp is about ply-depth (the program with the deeper depth >>generally wins). > >I do not think that the deeper depth is defined. >Programs use different extensions rules and different pruning rules. I was speaking in general. In general this is a rule of thumb (ever been since day one of computer chess). >> In human-comp playing style, strategic understanding, and >>having the initiative are the main items and that ply-depth comes after that. > >programs can see strategic good moves by deeper ply-depth. Of course. Only that the order is different. To be more precise: COMP-COMP: 1) depth (60%) 2) playing style (20%) 3) initiative (15%) 4) strategic (5%) HUMAN-COMP 1) playing style (30%) 2) initiative (25%) 3) strategic (25%) 4) ply-depth (20%) Now you can argue about the given percentages but for me this picture is true. IMO. For instance in comp-comp it does not matter that both programs do not understand a strategic position, they both fail. In human-comp it is a matter of life or dead (winning or losing). Ed >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.