The chess opponent, with whom the young lady had to deal in 1956 in the US Los Alamos, was almost two meters tall and quite intimidating. MANIAC, as he was called, had a tremendous amount of energy and was able to perform complex calculations to optimize thermonuclear weapons. He did not give the young woman on the chessboard a chance.

What happened in the mid-fifties in the US lab has made history. The victory of MANIAC was the first success of a machine against a man in chess at all. Scientists had alienated the giant, who had been in Los Alamos since 1951, and written a code to allow the computer to play chess. The match against the subject, who only learned the rules weeks before, only took place on a small 6x6 board. And it was a revolution.

Getty Images

Computer MANIAC in Los Alamos

The history of computer chess is full of anecdotes about how brilliant spirits wanted to teach machines the fascinating game early on. One of the most impressive achievements was the English math genius Alan Turing, who wrote a chess program in 1936 with his "paper machine" before any computers were available. The machine needed a human CPU, Turing himself, who had to laboriously calculate the features of his program with paper and pen. And yet: it worked.

95,000 letters to the editor at ZDF

My occupation with computer chess began half a century later. I was a young science journalist back then and had read an article about the progress that had been made in this area. After a little research, I went to my boss at ZDF. "Did you know that computers can play chess now? " I asked the legendary science presenter Hoimar von Ditfurth. I never forgot his answer: " I where!" For half an hour, I explained how it worked, then he said, "Write down everything, we'll do a 43-minute documentary."

Javascript is required for the display.

A few months later we had the Scottish International Master David Levy play CHESS 4.8 in our studio in Hamburg, the strongest chess program that existed back then. The trains were carried by a robotic arm, which was controlled by a mainframe in Minneapolis via transmission line. The game was very exciting and ended in a draw.

The documentation was a great success. The SPIEGEL reported in detail about it (and organized a short time later a second match against World Cup challenger Viktor Korchnoi). After the ZDF broadcast, the station received over 95,000 letters to the editor. All readers wanted the notation of the game, which had been commented on by Levy, Helmut Pfleger and the computer. All this is described in this report, which also contains the replayable game and the comments.

Kasparov 32, chess computer 0

In June 1985 then Garry Kasparow, the 22-year-old World Cup challenger, was invited by SPIEGEL. In Hamburg, the Russian should compete against 32 of the strongest chess computers that were then to buy in a simultaneous. Kasparov won the simul 32: 0 and proved that the human mind was far superior to the computers.

Kasparov was also at my home and we discussed for a long time about computers and the development of a chess database. I was not a programmer, but luckily I found Matthias Wüllenweber, a physics student who had already programmed a rudimentary chess database. We showed Kasparov the program and founded the company ChessBase at his urging. Today, ChessBase is a classic database that allows players to quickly search a very large number of games and search for new openings and new ideas and weaknesses of their opponents.

photo gallery


8 pictures

History of chess programs: Fritz and the paper machine

Fritz, more than just a nickname

In 1990, we had an idea: Would not it be nice not only to give the chess players access to a large amount of data, but also to add a "chess engine" to the program, a module that users could use to analyze? At that time, the first chess programs for IBM PCs had appeared, and we were able to get a Dutch programmer, Frans Morsch, to develop for us an engine that could be called up in ChessBase to analyze with their help.

imago sports photo

World Champion Kasparov against chess program Fritz (1994)

After that, it was only logical to develop an independent chess program, and this was completed in 1991. We baptized it "Fritz" for a reason: Germany had recently been reunited, and for the first time after World War II the Germans were indeed proud of their country. Fritz is the nickname that the British had given the Germans in the wars. Also, the name was better than something like "Grandmaster Chess," an original consideration.

On December 28, 1992, I took the current beta version of Fritz 3 to Cologne, where world champion Garry Kasparov was preparing for a competition. We installed the program on his HP OmniBook laptop and he played more than 30 flash games as if intoxicated. Fritz won four of them. It was a historic moment: for the first time in his life, Kasparov had lost a game against a computer, albeit with a shorter reflection and under informal conditions.

Deep Blue 2 creates the sensation

Kasparov played four years later against a machine that had been developed with millions of dollars by IBM. Kasparov won the competition against Deep Blue in Philadelphia quite comfortably. The worldwide attention that the match provoked prompted the computer giants to challenge Kasparov for a rematch. That was in 1997 in New York, and Kasparov had this time to do with an improved version of Deep Blue . He lost the competition, but not because the machine was stronger than him, but because he was psychologically inferior to his opponent.

ASSOCIATED PRESS

World Champion Kasparov, left, against Deep Blue (1996): "Psychologically inferior"

In my opinion, the computer had a game strength of almost 2700 Elo points, while the World Champion had more than 2800 Elo. But Deep Blue had a library of millions of opening moves he could access during the game. Kasparov tried to counter this with unconventional variants, but at least did not fit his style. He knew for a fact that playing against a machine was a lot different than chess competitions with human opponents: one mistake and one loses. During the game you are under constant pressure all the time. That has cramped his style.

Where are we now?

In 2002, Fritz played a match against World Champion Vladimir Kramnik and reached a 4: 4 draw. In the rematch in 2006, which went over six games, Fritz won 4: 2. A little later, after the Chess Engines skipped the 3000-point mark on the Elo scale, competitions were abandoned man-on-machine (the strongest human players come to around 2800 Elo points). Today, top engines such as Fritz, Stockfish, Komodo or Houdini are approaching the mark of 3500 Elo points. That is, if a human plays against these programs, it would be as if Usain Bolt is racing against a Ferrari. Just unfair.

But where are we now? In 2017, Google's daughter Deep Mind introduced an artificial intelligence called Alpha Zero , which was unlike any previous traditional chess engine that has been handcrafted over the years by top programmers. Alpha Zero had just explained the rules of the game and showed how the characters draw. After that, the program, which ran on massive hardware, played millions of games against itself, winning a profound chess strategy all by itself. After a few days of computational work, the program had reached superhuman strength and was able to win against one of the best engines in the world. It became the strongest unit ever to play chess.