scooter@genie.UUCP (Scooter Morris) (10/25/85)
Hi! I am forwarding this for a friend, please reply via mail to ucbvax!ucsfcgl!genie!adl I have been avidly following the posted results of the ACM tournament; can anyone tell me if there are any publications devoted to computer chess so I can stay up on developments? Chess Life barely pays it lip service (except in their advertising!). By the way, the tournament was covered in Time magazine (p.88, Oct.28). Thanks. Scooter Morris Genentech, Inc.
bl@hplabsb.UUCP (Bruce T. Lowerre) (12/14/85)
In case you haven't seen it, there was an article in the October 28 issue of Time about the latest computer chess machine, Hitech. This machine was developed at Carnegie-Mellon University by Hans Berliner (a computer science professor who is also a grand master level player and former world correspondence champion) and Carl Ebeling who developed the special purpose chip that is used. The machine is currently rated at about 2250 which is master level but Hans feels its actual rating is about 2500. The machine will be entered in more turnaments to determine its actual rating and to find the level of player that can consistently beat it. In the mean time, its weaknesses are being patched. As Levy says in the article, "In the past, chess players came to laugh. Next year they will be coming to watch. Soon they'll be coming to learn."
davet@oakhill.UUCP (Dave Trissel) (12/19/85)
In article <3199@hplabsb.UUCP> bl@hplabsb.UUCP (Bruce T. Lowerre) writes: >In case you haven't seen it, there was an article in the October >28 issue of Time about the latest computer chess machine, Hitech. > >The machine is currently rated at about 2250 which is master level >but Hans feels its actual rating is about 2500. The machine will >be entered in more turnaments to determine its actual rating and >to find the level of player that can consistently beat it. In >the mean time, its weaknesses are being patched. > I find the 2500 claim a little hard to believe. As the ratings get higher the more chess knowlege is required to be embedded in the program. (Just examine how large the opening libraries have grown from year to year in the best programs.) And, after leaving opening book the more chess knowlege embedded in the evaluation algorithms the slower which means fewer moves examined which leads to weak tactical play blundering. One obvious method of attack is to so specialize the program that it has a well defined goal decided at the time of exit from the opening book. But this would take a substantial effort and the results would still be inflexible. Anyone know any details of Hans system? Skeptically - Dave Trissel {ihnp4,seismo}!ut-sally!im4u!oakhill!davet
eli@cvl.UUCP (Eli Liang) (12/20/85)
In article <623@oakhill.UUCP> davet@oakhill.UUCP (Dave Trissel) writes: >In article <3199@hplabsb.UUCP> bl@hplabsb.UUCP (Bruce T. Lowerre) writes: >>In case you haven't seen it, there was an article in the October >>28 issue of Time about the latest computer chess machine, Hitech. >> >>The machine is currently rated at about 2250 which is master level >>but Hans feels its actual rating is about 2500. The machine will >>be entered in more turnaments to determine its actual rating and >>to find the level of player that can consistently beat it. In >>the mean time, its weaknesses are being patched. >> > >I find the 2500 claim a little hard to believe. As the ratings get higher >the more chess knowlege is required to be embedded in the program. (Just >examine how large the opening libraries have grown from year to year in the >best programs.) And, after leaving opening book the more chess knowlege >embedded in the evaluation algorithms the slower which means fewer moves >examined which leads to weak tactical play blundering. > >One obvious method of attack is to so specialize the program that it has >a well defined goal decided at the time of exit from the opening book. >But this would take a substantial effort and the results would still be >inflexible. Anyone know any details of Hans system? > >Skeptically - Dave Trissel {ihnp4,seismo}!ut-sally!im4u!oakhill!davet On the other hand, who would know better than a rated GM... Think Hitech beats Berliner at games? -eli -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Eli Liang --- University of Maryland Computer Vision Lab, (301) 454-4526 ARPA: eli@cvl, eli@lemuria, eli@mit-mc, eli@mit-prep CSNET: eli@cvl UUCP: {seismo,allegra,brl-bmd}!umcp-cs!cvl!eli
patrick@mcc-db2.UUCP (Patrick McGehearty) (12/20/85)
> Anyone know any details of Hans system? > > Skeptically - Dave Trissel {ihnp4,seismo}!ut-sally!im4u!oakhill!davet HiTech is discussed in the Winter 1985 issue of Carnegie-Mellon Magazine (an alumni mag). The following quotes are out of context... ability to consider 175,000 positions per second Carl Ebeling has designed a "smart" VLSI "move generator" that accounts for much of the performance of HiTech. 64 custom chips are used in HiTech/ one per square Each chip "identifies all moves that can bring any piece to its square and evaluates them based on the computer's knowledge of fundamentals of chess strategy" ...a sophisticated analysis program developed by Berliner decides how things stand in any board position before starting a search... In summary, HiTech is a smart, brute force program Having know Hans personally, I would expect him not to exaggerate with respect to chess performance. After all, the program has beat two masters and drawn to a third. Best performance ever by a machine against human players.
derek@uwvax.UUCP (Derek Zahn) (01/02/86)
After seeing the impressive HiTech results, and thinking a bit about possible ramifications, I have a question for you net- people. It seems clear that chess computers are getting better and better; to the point where I will bet that within 5 years there will be one that will play at true Grandmaster strength, and may be unbeatable by a human being. I was wondering what effect you think this will have upon the game of Chess. There are computers now that play better than I probably ever will, and I am not sure how I feel about that. To begin with, though, I don't think that computers belong in human chess tournaments. Special tournaments set up to test a computer's playing strength are fine, but I don't think that they belong in "opens" (the last one I played in was open to computers, although none entered). My reason for this is simple. By analogy, a robot could probably be built that could run faster than a human being, but that does not mean that we allow them in marathons. Tournaments and competitions are for humans vs humans. If computers started taking top prizes in our tournaments, I think it would delete greatly from the excitement of tournament play. I am not a chess-computerphobe; I have one at home and think it is a valuable training aid, but I have little desire to fight it for points at a tournament. Any opinions? derek
bl@hplabsb.UUCP (01/02/86)
> In article <623@oakhill.UUCP> davet@oakhill.UUCP (Dave Trissel) writes: > >In article <3199@hplabsb.UUCP> bl@hplabsb.UUCP (Bruce T. Lowerre) writes: > >>In case you haven't seen it, there was an article in the October > >>28 issue of Time about the latest computer chess machine, Hitech. > >> > >>The machine is currently rated at about 2250 which is master level > >>but Hans feels its actual rating is about 2500. The machine will > >>be entered in more turnaments to determine its actual rating and > >>to find the level of player that can consistently beat it. In > >>the mean time, its weaknesses are being patched. > >> > > > >I find the 2500 claim a little hard to believe. As the ratings get higher > >the more chess knowlege is required to be embedded in the program. (Just > >examine how large the opening libraries have grown from year to year in the > >best programs.) And, after leaving opening book the more chess knowlege > >embedded in the evaluation algorithms the slower which means fewer moves > >examined which leads to weak tactical play blundering. > > > >One obvious method of attack is to so specialize the program that it has > >a well defined goal decided at the time of exit from the opening book. > >But this would take a substantial effort and the results would still be > >inflexible. Anyone know any details of Hans system? > > > >Skeptically - Dave Trissel {ihnp4,seismo}!ut-sally!im4u!oakhill!davet The evaluation of positions after each move could be expensive if one were to include a large amount of chess knowledge in the evaluation. However, Hitech, like its predecessor Tech, uses a very simple evaluation metric, material gain or loss. However, with this type of evaluation, the program would be willing to play P-KR4 just as much as P-K4 for the opening move. Thus, the "chess knowledge" is put in to sort, or prioritise, the moves at the top of the search tree and the search is used to evaluate the tactical feasibility of the ordered set of moves. The kind of "chess knowledge" used depends on the position, i.e., opening book, middle game, king safety, end game, etc. > > > On the other hand, who would know better than a rated GM... Think > Hitech beats Berliner at games? I asked Hans how he plays against Hitech. Since he knows how the machine plays, he has an advantage. However, he says he played one game with it and drew.
trb@haddock.UUCP (01/03/86)
> I was wondering what effect you think this [computer chess players] will > have upon the game of Chess. There are computers now that play better > than I probably ever will, and I am not sure how I feel about that. I like the fact that computers are working on the game of chess. I don't see any problem with computers "solving" the game of chess. Even if a computer crunched for an eon and spat out "white always wins," perhaps accompanied by reams of data, what effect should that have on human chess players? None. I own a computer that, under fast time controls, almost always thrashes me. I'm about a 1650, and the computer is USCF rated a shade over 2000. (I think it has a 4MHz Z80 in it, talk about humiliating. Then again, if I put as much effort into chess study as went into my chess box, my rating would be over 2000 too.) It beats me 90% of the time at about a minute a move. I don't see the difference between that and a better computer beating a better human chess player. I still play chess. LONG before machines play perfect chess, they're merely going to play better chess then humans will. There are many endeavors at which machines excel. Machines can travel faster than humans, lift more than humans, cut down trees faster, work out math problems faster, etc. Humans still compete at all these endeavors, but at some point, it became ridiculous for humans to compete with machines in them. And some day, so it will be with chess. But for now, computers and humans are worthy competitors for each other. Andrew Tannenbaum Interactive Boston, MA 617-247-1155
bl@hplabsb.UUCP (Bruce T. Lowerre) (01/11/86)
> To begin with, though, I don't think that computers belong in > human chess tournaments. Special tournaments set up to test a > computer's playing strength are fine, but I don't think that > they belong in "opens" (the last one I played in was open to > computers, although none entered). If a tournament is "open", then it's open to any player. If you want to have "human only" tournaments, then OK. If the object of an open tournament is to find the strongest player, then so what if the winner is human, machine, or martian. > My reason for this is > simple. By analogy, a robot could probably be built that could > run faster than a human being, but that does not mean that > we allow them in marathons. Tournaments and competitions are > for humans vs humans. Says who? > If computers started taking top prizes in > our tournaments, I think it would delete greatly from the > excitement of tournament play. I think it would be just the opposite. > I am not a chess-computerphobe; > I have one at home and think it is a valuable training aid, > but I have little desire to fight it for points at a tournament. > > Any opinions? Yes. I find this whole thing very amusing. I was waiting to see if Hans would respond to this. Back in 1971 when Hans and I were graduate students at CMU, the ACM held its first computer chess tournament. The quality of the moves that the programs made ranged from weak at best to suicidal and illegal! This prompted Levi to make his famous bet that no machine could beat him within x (I forgot the exact number) years. Some foolhearty computer science AI types took hum up on his bet. He collected. Today, we see evidence that someday a machine may possibly be the world champion. This is a significant breakthrough in AI!! For a long time, chess was believed to be an intellectual activity that was beyond the ability of a machine. I'm happy and proud that Hans has been part of this; he is finally realizing his dream.
derek@uwvax.UUCP (Derek Zahn) (01/12/86)
> If a tournament is "open", then it's open to any player. If you > want to have "human only" tournaments, then OK. If the object of > an open tournament is to find the strongest player, then so what if > the winner is human, machine, or martian. I don't think that the object of a tournament is to find the strongest player. That seems a pretty narrow view. > > ... Tournaments and competitions are for humans vs humans. > > Says who? Obviously, this is a matter of opinion. In the open tournament I referred to, a surprisingly large number of people went out of their way to request NOT to be paired with a computer. The TD will not allow computers next year for this reason. > > If computers started taking top prizes in > > our tournaments, I think it would delete greatly from the > > excitement of tournament play. > > I think it would be just the opposite. Really? I have a hard time believing that the average (read: non-computer- science-nut) tournament player would be thrilled by the spectacle of somebody's machine taking top-prize from Masters struggling to make a living (or at least part of one) from chess. > Today, we see evidence that someday a machine may possibly be the world > champion. This is a significant breakthrough in AI!! For a long time, > chess was believed to be an intellectual activity that was beyond the > ability of a machine. I'm happy and proud that Hans has been part of > this; he is finally realizing his dream. I realize the marvelous progress that Hans+Co. have made in the field of AI, and as a computer scientist, can appreciate the accomplishment and all of the blood, sweat, and tears that have gone into it so far. There is plenty of which to be proud. However, I doubt that a machine will ever "be the world champion". It may BEAT the world champion, but will never be it. The idea is ludicrous, just as computers being world bridge, checkers, etc champions is silly, even though there must be computers that could kick the stuffing out of humans in these games. I guess that there is really nothing to worry about, though. As soon as computers get better than the world champion, they will most likely be banned from most tournaments. Until then, it will be fun to watch the progress, and I will not play a computer willingly at a tournament; if I want to get creamed by a machine, I have one in my living room on 24-hour call. derek
victor@klipper.UUCP (L. Victor Allis) (01/12/86)
>> By analogy, a robot could probably be built that could >> run faster than a human being, but that does not mean that >> we allow them in marathons. Tournaments and competitions are >> for humans vs humans. > >Says who? My dog would have liked to beat Carl Lewis in Los Angelos at the 100 and 200 meters, know he hears it is not for humans only. L. Victor Allis Free University of Amsterdam The Netherlands
wjm@teddy.UUCP (01/13/86)
In article <504@uwvax.UUCP> derek@uwvax.UUCP (Derek Zahn) writes: >However, I doubt that a machine will ever "be the world champion". It may >BEAT the world champion, but will never be it. The idea is ludicrous, just >as computers being world bridge, checkers, etc champions is silly, even though >there must be computers that could kick the stuffing out of humans in >these games. > I hate to disappoint you but I think bridge may be much harder for a computer than chess. It requires alot more judgement and psychology. > >derek bill