Background information
Elden Ring: Shadow of the Erdtree – an expansion that’s better than some full-price games
by Philipp Rüegg
When the best chess player was no longer a man
11.5.2022
Translation: machine translated
25 years ago, machines took over the world. But only in chess. Mankind need not feel offended or threatened by this.
Twenty-five years ago today, on May 11, 1997, the IBM chess computer "Deep Blue" won the final and decisive game in a competition against the then world champion Garri Kasparov. Deep Blue thus became the first chess computer to win a match under tournament conditions against a world chess champion. Deep Blue won two games, Kasparov one; the other three ended in a draw.
An earlier version of Deep Blue had already played against Kasparov in 1996 and also won one game, but lost the match by 2:4 points.
The match was a big event in the media. The result could be interpreted as a science fiction scenario come true: Machines outstripped humans in intelligence. Or maybe not. For the competition also prompted reflection on the question of what actually constitutes intelligence.
Is a chess computer intelligent?
The basic procedure of a conventional chess computer is simple: it tries out possibilities and evaluates them with the help of a huge database. It does this much faster, more reliably and more widely than any human. This not really intelligent principle leads to strong moves, if the computer is fast enough.
So it was less about intelligence than about computing power. As late as 1985, Kasparov played against 32 chess computers simultaneously and defeated them all. "For me, that was the Golden Age. The machines were weak, my hair strong," Kasparov later joked in a TED Talk.
For the 1997 competition, Deep Blue was massively upgraded. With its 480 built-in chess processors, it offered extreme computing power for the time. It could calculate up to 200 million positions per second.
However, Deep Blue's software was not so trivial that it only played through possibilities. It was further developed for the 1997 competition. The programmers even made interventions between games to further optimize the system. This made the computer unpredictable for Kasparov. Deep Blue, on the other hand, had access to all the games Kasparov had ever played, so it knew his opponent inside out.
Source: Stan Honda, Keystone
People get tired, have a bad day, or get rattled. Kasparov won the first game and gave up the second. Chess analysts pointed out after the game that Kasparov probably could have forced a draw by playing a perpetual check. This obviously affected the world champion, who subsequently fell short of his potential. In order to be less predictable for Deep Blue, he tried an opening in the last and decisive game that he usually never played. A high-risk strategy that didn't work out. Kasparov was forced into an impossible position and resigned after only 19 moves.
Machines can do better - so what?
Despite all these objections, one thing is clear: when it comes to a purely logical game like chess, computers are superior to humans. Even before the 1997 competition, it was foreseeable that it would only be a matter of time before humans no longer stood a chance against machines.
Today, this is a matter of course. Even simple programs, such as the software of a fitness machine, are at least equal to humans. Chess streamer Levy Rozman, after all an international champion with an ELO score of over 2400 at times, dryly remarks at the end of his Deep Blue analysis, "I played to a draw against a treadmill."
With the use of self-learning systems, modern chess software is becoming even stronger. AlphaZero taught itself to play chess in late 2017, and after a short time beat Stockfish, the strongest software at the time. AlphaZero calculates far fewer moves than a traditional chess program; but more purposefully.
There is no need for humanity to feel offended or humiliated because of this. After all, humans have made all this possible.
The triumph of the machine has done no harm to mankind or to chess. Today, more people than ever play chess. And we are also a long way from HAL-9000-style horror scenarios. Chess computers just play chess better, but they won't take over world domination.
Header image: Image: Keystone
David Lee
Senior Editor
David.Lee@digitecgalaxus.chMy interest in IT and writing landed me in tech journalism early on (2000). I want to know how we can use technology without being used. Outside of the office, I’m a keen musician who makes up for lacking talent with excessive enthusiasm.
6 comments
later