"Man vs. machine" misses the point

FLAGSTAFF, Arizona, January 7, 2004 -- A couple of months ago, Garry Kasparov, generally considered the world's greatest chessplayer, played a four-game match against a software program called X3D Fritz. Having lost a similar match against the program's predecessor a few years ago, Kasparov was eager to prove that man could still beat machine. Alas, he failed, managing only a 2-2 tie, with a win, a loss, and two draws against the server-grade multiprocessor system.

With Kasparov's second failure, it's pretty clear that we've reached the point some of us flesh-and-blood chessplayers have been dreading for two or three decades now. Humans can no longer beat computers at chess. Kasparov says he's going to try again -- but it's a losing battle. Computer hardware gets faster, and chess software gets better, over time -- they don't get worse.


The essential question of chess remains unanswered: Can White force a win, or can Black force a draw?


It's testimony to the complexity of chess that the game has stood up so well to decades of exhaustive analysis. Thousands of theoretical books and compilations of games are available for study. Electronic databases of grandmaster games allow players to keep up with the latest additions to this body of knowledge. Finally, the game has been subject to theoretical analysis by software programs so strong that no human can beat them. Yet the essential question of chess remains unanswered: Can White force a win, or can Black force a draw?

Let your modem do the driving

I haven't played a face-to-face game in more years than I'd care to relate, but over the past three or four years, I've been playing sporadically on the Internet Chess Club (www.chessclub.com). This is one of the coolest social changes spawned by the Internet phenomenon -- you can play chess anytime of day or night, against people from all over the world, just by logging in and announcing your availability for a game. (Conversely, chess-by-wire has become a serious problem for organizers of over-the-board tournaments -- a lot of people have given up playing OTB, because it's so much easier to stay home and play on the dot-com. But that's another story.)


Playing against a computer, once you're behind, you're toast.


These days, 50 bucks will buy a chess program that'll give Kasparov a good game -- which is to say, a program that'll flatten anyone else. Mac OS X even comes with a pretty good program built-in. I've played a few games against it -- as well as against other software programs -- but found it less than compelling. Part of it is that I'm a lousy chessplayer (for the record, my USCF rating is a pathetic 1156; on chessclub.com, my blitz rating is around 1275); as a result, I invariably make a couple of silly tactical mistakes in the early going, and after 12-15 moves, find that I'm in a position where I've got no chance to come back and win. That's largely because the typical chess program, although it may not be any great strategic wizard, almost never makes a tactical error. Give it a chance to win a pawn or a minor piece, and the program is not going to miss it.

Playing a human opponent of roughly the same strategic acumen, however, is much more fun, simply because humans make mistakes. Get behind against a computer, and you're not coming back -- but get behind against a person, and you've got reason for hope. If you hang in there, set a few traps, and play aggressively, your opponent might slip up.

Cheapo? What's wrong with a cheapo?

When playing blitz games against human opponents, I have a few rules. Never resign as long as you've got a queen. No matter how hopeless the situation, with a queen on the board, there's always the chance that you can steal a cheapo win, if your opponent lets his guard down. The corollary, of course, is: Keep your queen the heck on the board, if you're behind, and if you're ahead, force the other guy to trade! Also, when in doubt -- and especially, when behind -- threaten mate, any way you can. In a blitz game, the guy might miss it, or accidentally forget and move a piece that protects him against the mate threat. In blitz games, play as aggressively as possible -- aggressive play tends to cause mistakes on your opponent's part. (It's also a lot more fun than cautious play -- and fun, not artistry, is the main object of blitz chess.)

This is basic stuff -- but what I'm getting at is that these tactics don't work worth beans against a computer. That's because computers (or, rather, programs) are tactically superior to you and me. A program is not going to hang a queen, miss a mate threat, or let you pull off a cheapo despite a huge material disadvantage. That being the case, once you're behind, you're toast.


Chess isn't just a series of calculations. It's all about going nose-to-nose with your opponent.


The other -- and, as far as high-level chess is concerned, more important -- part of it is that chess is a physical sport. Non-players have no way of comprehending the number of calories you burn during a three- or four-hour chess game, or how drained you can be when getting up from the table. Knowing how to play the game isn't enough; you have to be physically fit enough to keep your brain dialed in the whole time. For a computer, this isn't a problem -- they just keep calculating at the same level, no matter how long the game lasts. This is why chessplayers tend to peak in their 30s or early 40s -- a player who's 50 or 60 years old still understands the game as well as the younger player, but at that age, stamina isn't what it used to be.

For the money and for the show

Garry Kasparov is needlessly debasing himself and his fellow mortals by playing these computer matches. One reason he does it is that the promoters -- in the most recent case, the company that makes X3D Fritz and other "virtual reality" games -- pay him a lot of money to play. Another reason he does it is simply to promote chess. ESPN-2's live coverage of November's match did a lot to increase the game's visibility, which badly needs a boost, having endured a long, slow decline in this country since the glory days of Fischer-Spassky in 1972.

But chess isn't merely a series of calculations, which is what computer programs reduce it to. It's about going nose-to-nose with a human opponent, and combining knowledge with stamina, gamesmanship, guts, and psychology to grind your opponent down. There's a famous quote from Bobby Fischer to the effect that the most satisfying moment in chess is not checkmate, but rather the moment at which you break your opponent's will. That being the case, what's the point of playing against hardware and software? They've got no will!

Resistance is futile

Even 30 years ago, when early software, running on room-sized computers, couldn't beat a fairly decent professional player, the handwriting was on the wall -- sooner or later, computers would be able to beat the best human players. But so what?

Garry Kasparov is a great man. He should be playing nationally-televised matches against the strongest flesh-and-blood opponents out there. It's a shame that promoters won't pay for that. Those of you who saw PBS's coverage of Fischer-Spassky will undoubtedly recall how much more compelling it was to watch two great minds square off (even second-hand, as no TV cameras were allowed in the competition hall), than to watch one great player shovelling against the tide of software technology. Kasparov vs. Fritz is better than nothing, but it's a far cry from a real chess match.

Copyright © 2004 Kafalas.com, LLC


Feedback?  Fire off a letter to the editor, and we'll post it on the letters page. Letters may be edited for clarity or length. Also, you may have to endure some friendly abuse from the editor, should he disagree with you.


Return to the home page