Author: Roger D Davis
Date: 22:19:23 07/02/99
Go up one level in this thread
I don't dispute the points you made in your first paragraph at all, and I agree with everything you said in your second paragraph up to your conclusion "GMs will lose respect, chess will seem less amazing, and computer chess will dwindle to the few people who really love it." I guess I'm just so used to having my butt kicked that anyone who can squeak out an occasional draw against these beasts gets my respect. IMHO, in the future computers will play so much a part of our lives that we'll all know how powerful they are. They'll do many, many things for us, and they'll be so powerful that they'll do it mostly transparently in ways we can't even imagine. We'll take it for granted, but we'll know how powerful they are. And consequently, anyone that can squeak out a draw against one, or even come close, will be respected. I just don't see grandmasters losing respect just because they can be beaten consistently by a computer. Deep Blue beat Kasparov, but chess is stronger than ever, and it's not because the public knows all the rationalizations for his defeat that are typically discussed in these forums. I think chess will spread even more, because it will be fun to program, because programming resources will be widely available to almost anyone. I also think it will be more fun to play, not only because of various styles within "normal chess," but because chess variants will become widespread, too. Before we get to the "chess solved" stage, I can imagine we might even see "chess compilers," where you pick your extensions, search techniques, where you want your evaluation to be done, and so on, and then it's all compiled into an executable for you. Should be fun. I guess I think people have already gotten to the point where they believe that anything a human can do a computer can do better...someday. So when the grandmasters start losing at 40/120, well, it's a milestone for us here, but by the time that happens I don't think it will be a big deal in public mind, assuming it is even now. Roger >I think you miss my point. In the near future, even if programmers seriously >weaken their program with strange styles, they will _still_ be able to clean any >human's clock. And at that point, there is no point in having a GM tournament >you can take seriously if some of the participants are named things like >"KnightGrabbing Attacker" and "Morphy-Style". > >I am talking about the way the public _views_ chess, computer or otherwise. Once >computers dominate there will be no reason to have programs with styles except >for relaxation games at home. There will be no reason to have computers in human >tournaments at all. GMs will lose respect, chess will seem less amazing, and >computer chess will dwindle to the few people who really love it. > >James > >>You and I agree that it would be ludicrous to have grandmasters play a >>deliberately weakened program in order to avoid getting clobbered. I never >>claimed otherwise. >> >>However, while I find that different programs do have different styles of play, >>usually the style is created at the price of weakening the chess program. You >>mentioned a search ability tantamount to showing that chess is solvable. My >>claim is simply that this will allow chess programs to play in different styles >>in a way that at least do not as readily sacrifice the possibility of a draw, as >>they do now, and that these various styles could be very >>entertaining/frustrating depending on the ingenuity of the engine author. You >>can now get clobbered in the way you choose. This is possible now, but probably >>only if you're a master level player. A grandmaster player would see the holes >>that our current techniques create to create the style and exploit them. IMHO, >>the search tree is vast enough so support a variety of styles at a much higher >>level of play than we see now. >> >>I also do not believe that once computers dominate completely the game will >>become less interesting because the mystique of human talent will be gone. For >>me, that mystique has already been trasferred to the programs as they kick my >>ass over and over again. Beyond this, sith a search capacity that you mentioned, >>one tantamount to solving the game, it would be possible to computer exact >>scores for the moves possible from any particular position. Not only would that >>make highly competent styles possible, it would also make it possible to rank >>order the moves from better to worse with much greater reliability across >>programs than exists now, make the computer an ideal teacher. Also, the computer >>could rate your play much more precisely, or even classify the legal moves into >>the various styles of play it recognizes. >> >>My assumption is that these possibilities will prove interesting to engine >>authors, simply because human beings have always been interested in intellectual >>activities that have no practical benefit (add to this competition, and its >>starting to sound close to the definition of a game). Moreover, by that time, >>chess programming resources will be widely available, and college kids will >>develop them as a senior project, or the like. >> >>Right now we're in an era of strength. Strength is everything among the elite >>programmers. But we're gonna move into an era of features and fun, and chess and >>chess programming will expand. >> >>Just an opinion. >> >>Roger >> >> >> >> >> >> >> >> >> >> >> >>On July 02, 1999 at 18:32:18, James Robertson wrote: >> >>>On July 02, 1999 at 17:02:09, Roger D Davis wrote: >>> >>>>I don't think the field will shrink...I think it will expand because it will be >>>>possible for computers to play in genuinely different styles without making >>>>mistakes obvious even to master level players (didn't say grandmaster level >>>>players). You can imagine a style that destroys you tactically by following >>>>lines especially dangerous to humans, another one that presents you with shallow >>>>tactical threats but makes you run out of time, another one that just prolongs >>>>the game indefinitely without ever doing anything, and so on. >>>> >>>>Roger >>> >>>Computers are at that point now. I think it is not difficult to make programs >>>play in different styles; take for example CST. >>> >>>But in 5-10 years? In top level competitions? Computers will be so much better >>>there will be no contest. It will be ludicrous to have the 'greatest humans on >>>earth' playing a deliberately weakened baby chess program just so they have a >>>chance to not get clobbered. >>> >>>Face it. Chess GMs are considered to be above average intelligence, because most >>>people have tried chess and found it _hard_. There is a 'mystique' attached to a >>>good chess player; they can outwit a machine, they have unusual ability to grasp >>>spatial stuff, etc. etc. The moment a computer proves chess is 'solvable', much >>>of that 'geniousness' is gone, and the sport of chess will have been struck a >>>serious blow. >>> >>>This is perhaps bleak, but I do think it is the future. >>> >>>James >>> >>>> >>>> >>>>On July 02, 1999 at 14:28:47, James Robertson wrote: >>>> >>>>>On July 02, 1999 at 14:16:50, James T. Walker wrote: >>>>> >>>>>>Hello, >>>>>>I see the change in strategy by the top GM's in the Frankfurt Masters tournament >>>>>>as an admission that they can no longer compete with Fritz on the "up and up". >>>>>>Resorting to anticomputer strategy/tactics by the worlds best chess players is >>>>>>the first signal that the end is near. Computer domination is just around the >>>>>>corner. >>>>>>Jim Walker >>>>> >>>>>Of course it is. Once computers do dominate, the public will lose interest in >>>>>them, and I think the computer chess field will shrink to just a few of us >>>>>computer nerds. It is ironic that the people who want so hard for computers to >>>>>dominate as quickly as possible are actually just moving the date forward when >>>>>computers are unable to interest the average chess player. People used to holler >>>>>and cheer when there were airplane - auto races. No such thing now, and computer >>>>>chess will go the same way. :( >>>>> >>>>>James
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.