The Man vs Machine debate continues

10/20/2004 – A few days ago we published a resume by David Levy, president of the International Computer Games Association, on the recent man vs machine team championship in Bilbao, which ended in a bitter 3.5:8.5 loss to the humans. A number of letters were fired at us, some applauding and some vigorously contradicting David's conclusions.

ChessBase 14 Download ChessBase 14 Download

Everyone uses ChessBase, from the World Champion to the amateur next door. Start your personal success story with ChessBase 14 and enjoy your chess even more!


Along with the ChessBase 14 program you can access the Live Database of 8 million games, and receive three months of free ChesssBase Account Premium membership and all of our online apps! Have a look today!

More...

The number of letters that came in as a reaction to David Levy's articles was not very large. For the first time we are able to publish all letters that we received, roughly in the order in which they arrived in our news mailbox. They are a reaction to the following articles, which you should read if you haven't done so already:

Your opinion on Man vs Machine

Vincent Ong, Singapore
Thank you for your wonderful website. I visit it more often than all the other chess websites put together. It seems to me that the mistake that the GM team makes is to play the computers the same way that they play other GMs. At the level that computers are now playing, the only way for humans to beat computers is to play anti-computer chess. Even Kasparov had to do this when he really needed a win against Junior. I think the GM team would have done much better if they had included Boris Alterman, instead of, say, Ponomariov. Although Boris is rated much lower than Ruslan, it is his anti-computer skill that counts in this case. I found the games at Bilbao more interesting than the games at Brissago because some of the games at Brissago are merely draws at the end of the opening book.

Tony Wong, USA
H human vs computer in chess is boring. David Levy's quote "My own view is that the games in Bilbao are far more interesting for the chess public than are those played in Brissago". Who is this chess public he is talking about?? IMs, Fide Masters, National Master? Certainly he can't mean the average mass of weekend tournament players, because I'll bet good money that the average chess player might know what's going on in Brissago, but for sure they have no idea what went on at Bilbao. I just stumbled on to this story by accident while looking up info on the classical world championship match. Being a chess fan and no great player by any means, I find there is no interest in seeing my fellow human get the crap kicked out of him in a game of chess against a computer. I rather see the human drama of man vs man or man vs woman or man vs whiz kid. I want to know something about the persons playing. I want to relate to the persons playing. Hey, I like my computer, but I hate playing chess against it. Just as I like my car but would never run a race with it. Competing against machines suck. I'll take the imperfect human struggle any day.


David Levy, President of the International Computer Games Association

Leon Piasetski, Kyoto
I agree with David Levy about the importance of preparation, and also felt that Karjakin and, especially Ponomariov, seemed to drift into positions that were easy for computers to figure out and consequently gave the humans more headaches. However, I wonder if computer programs are getting an unfair advantage because of fast time controls. It seems unnecessary to put biological entities under excessive pressure! It's psychologically tough facing a non-human challenger that presumably plays 'perfect' chess. Why add the extra burden of time pressure which has almost no effect on the machines but is a critical issue for humans?

There are two main considerations, getting an advantage and maintaining at least equality. If the computer gets a winning advantage it's all over. For the human a winning advantage is just the start of a long process. At times critical positions arise where a second best move may reduce winning chances. The problem is compounded when the desire to continue trying for a win clouds judgment. Grandmasters may decide to play safe rather than find the best possible move because they don't have enough time to recognize and avoid all the traps. This is why so many recent games end in draws; no one wants to lose ignominiously.

Two hours for 40 moves might be an acceptable time control for human or computer events but it's definitely too fast in these mixed competitions. Perhaps two hours for the first 30 moves with one hour increments each 20 moves would allow grandmasters to demonstrate their resilience and consistency. I believe grandmasters can evaluate positions more accurately than computer programs and slower time controls will give them a chance to prove that they can do so consistently throughout a game.

Charles Wilson Columbia, SC, USA
First David Levy exalts the computers by making more than one statement to the effect of "Programs do not make such oversights", as if computers have no faults. Then he spends several sentences defending the poor performance of Junior by making the excuse it was "off form". Well, excuse me while I hurl. Levy is trying so hard to put a spin on here, he ought to work for President Bush.


The two Ukraine super-talents Sergey Karjakin and Ruslan Ponomariov after the last round

Igor Stein New York
I am tired of humans whining and constantly complaining of computers having "advantages". Instead the carbon units should learn how to play better chess. And stop calling chess "an art form" and accusing computers of ruining the "beauty" of chess. Computers just playing better chess – plain and simple. I know it's hard to swallow. But that's exactly what we should do if we want to make any progress in this game.

Alexander Koskoros, Corfu, Greece
Great idea, great show. I am really looking forward to the next such event.

Chrilly Donninger, Altmelon, Austria [author of Hydra]
David Levy's description of the games is too human based. E.g. in the game Hydra-Ponomariov, the black position might be in a human-human encounter favourable. But I was absolutely sure Hydra wins the game. Once a position looks like this, it is only a question of time till the human runs into a right hook of the programm.

The GMs (their manager) complained, that they can not prepare specifically against the programms. For this one would need a regular human-computer tournament cycle. At the moment it simply does not pay to do some computer homework. It came never to their mind that we are trying to establish such a cycle. Especially it is over their mental horizon, that one has first to invest in a new market before one can get the money back. The Hydra team would have given them access to the machine to do some preparation. We are interested in a close fight.

GM Topalov tried at least during the games his best to defend his honour. They other two GMs frankly said that they do not care about their losses. Obviously they were just there to cash in some money. Their behaviour is even from this point of view stupid, because it was the first and last time they will get this chance. I will at least advise the Hydra-sponsor to spend only money for opponents who take their job serious.

Two boxers doing the Ali-Shuffle in their corner is not interesting. But a fight where one of the boxers looks for his first chance to be "knocked-out" is also not very exciting.


In a German computer chess forum Chrilly Donninger accused Ruslan Ponomariov of deviously behaviour. Apparently the FIDE ex-world champions trying to catch a glimpse of Hydra's analysis on the computer screen. Donninger kept his notebook semi-closed to prevent this, as documented in this picture by Nadja Woisin.

Kurt Utzinger Wetzikon, Switzerland
I agree with most of David Levy's conclusions. He made some very interesting comments. He however forgot to mention one thing: humans have in most of the games chosen the wrong opening or at least bad lines in so far as these openings or lines were full of complications. This does not work vs computers. It is still my opinion that by using a very boring style and the "do-nothing-but-do-it-well-strategy" the computers would hardly ever win a game, even against much lower rated players. And this would the GM's give the opportunity to sometimes win a game as their chess understanding is of course much better.

Carl Burgess, Liverpool, England
In spite of your hype, computer events are not that popular. I did not see a single article or news report on the matches in Abu Dhabi or Bilbao. The computer matches were even ignored by most chess sites. I believe that TWIC did not even publish the games. If you go to Google News and search for "Chess Hydra" or "Bilbao Fritz" you will find just six articles. Two are in a chess column in the Philippines, four are by www.chessbase.com. So we can only conclude that ChessBase is driving the PR behind these events, all by itself. This might make commercial sense, but do not draw the conclusion that they are more popular or important than regular human chess tournaments. They are not, there is no comparison.

John Potter Dallas, Texas
Frederic, who cares? I don't play chess because humans can beat computers or because they can't. Computers help me enjoy the game more by allowing me to study the game more easily, or to play online instead of having to go to a club, and that is good for me and for the game, also. I don't stay up late at night worrying about whether a computer can be a human at chess. Again, who cares?

Roscelin Compiègne
If the Bilbao 2004 event shows one thing, it is that grandmasters should stop "playing chess" against computers. It just doesn't make any sense. Computers do not "play", they just compute very fast. And they do it only because humans make them do so. Does it really make any sense to have humans competing against machines driven by other humans? Maurice Greene is not trying to win a 100 meters against someone driving a Ferrari, is he? Should we feel sorry for humankind because Greene would loose this kind of stupid "game"? Grandmasters in Bilbao were not beaten by computers, they were beaten by some men using computers. On the Playchess server this is considered a rule offence. I don't see any compelling reason to consider it any different in Bilbao. Give the Grandmasters a computer too, and we'll see what's what...


Ponomariov kibitzing in the Topalov-Junior game

Owen Francis, Wales
I would be interested in David Levy's opinion on the following. Taking account human failings such as tiredness and temporary lapses, it seems to me that the top GMs are doing pretty well against these incredible calculating machines. In fact it appears that the top human players (e.g. Gary Kasparov) and the machines are reaching more or less the same level of performance.

Why is this? Why for example aren't the top programs thrashing Kasparov every time, or alternatively why haven't the top programs levelled off their performance at say a 2400 rating or whatever.

Is it in fact because the top humans are now playing chess just about as well as chess can be played by any "intelligence"? If so no amount of computing power or fancy program will be able to level off at a very much higher standard of play.

If the above is true it could imply that super-intelligent beings on another planet won't in fact be able to play chess much better than Gary. Going further, and in another context, scientists often speculate about whether our understanding of the universe will be limited by our intelligence. If some humans do play chess as well as it is possible to play, then perhaps humans (or at least some of us) are sufficiently intelligent to grasp all that could be grasped about the universe by any intelligence!

Alfred Acaling, Negros Occidental, Philippines
These matches are exciting but one sided. It would be better if the world's top five would play against them to redress the balance and to make sure of competition. In the U.S. they used to play man vs machines but they stopped doing this in the mid 90's as it is clear that it would be no contest. Good Luck in your future matches!


Sergey Karjakin being interviews after a game for a TV production on the match

Emerson Tan, General Santos City, Philippines
What to do when computers will beat all humans? When computers beat all humans, the interest will wane for computers versus humans. It will be hard to get sponsors. The only way to keep it interesting is to change the rules in favor of humans.

One rule is to allow the humans to move the pieces on the board so as to avoid visualization blunders like what happened to Kramnik vs Deep Fritz and Kasparov vs Deep Junior. It will also be lest tiring and less lack of confidence for humans if they can move the pieces several moves deep until they are confident of their position. It is not interesting if humans lose due to visualization blunders, under perform due to tiredness or take the draw because of lack of confidence thinking that he will blunder along the way against an opponent that will not blunder even in time pressure. Imagine if humans will be allowed to move the pieces, they will have confidence that they will enter complications and play with the computers head on, the position will become more dynamic that we will have more decisive games. We will have more quality games to replay.

With the proposal, it will be the calculating ability of the computer versus the judgment and intuition of the humans, not the calculating ability of computers versus human’s tendency to blunder, get tired or lose confidence. It's much more satisfying to see a computer win by outplaying a human rather than winning due to human blunder.

Another good thing about humans being allowed to move the pieces is that it will be more interesting for the TV because you can see what the Grandmasters are thinking. It's much easier to sponsor if it's fitted for TV.

Tomasz Pintal, Stalowa Wola, Poland
I would like to tell you, that this event was extremely popular in my opinion. Not only popular, but very interesting for overall publicity (like David Levy said). I am waiting for such event like this in the nearest future and organizers should give chances to best players on the world (e.g. Anand, Kasparov, Kramnik, Leko, Topalov) and most of us will be looking how strong these programs really are.


Passers by watching the games in the window of the playing site

Krishna Menon, Kerala, India
It looks as if only a team consisting of Kasparov, Anand and Kramnik have any chance against these machines today. A match should be arranged between Anand and Hydra to test the monster against one of the best humans. Kramnik and Kasparov have recently had their chances against machines. I hope this can be brought about.

Fran Rivera, Costa Rica
Whenever a human loses against a machine, the analyst say "oh, the human is not prepared for play against this program". The humans players also say "the machine doesn't understand chess, those moves are ugly". But they (the humans) cannot win. No, today the computers are stronger than humans. The world chess champion are not Ks, they are machines. The King is dead, long live the King.

Links


Discussion and Feedback Join the public discussion or submit your feedback to the editors


Discuss

Rules for reader comments

 
 

Not registered yet? Register