Random Posts

Thursday, August 23, 2018

Did the Russians Really Cheat?

     It has long been speculated that the Soviets coerced Paul Keres into throwing games to Mikhail Botvinnik in the 1948 World Championship Match-Tournament, held in The Hague and Moscow. 
     Keres's native Estonia, had been annexed by the Soviet Union in 1940, invaded by Germany in World War II, and retaken by the Soviets. Keres found himself controlled by the Nazis during the war and was forced to participate in German tournaments and they used him for propaganda purposes. After the war, Keres may have escaped execution and he may have been playing for the Soviets under duress. Who knows the real story? 
     Reuben Fine, who declined to take part in the tournament, said,"I don't want to waste two months of my life watching Russians throw games to each other." All the Soviets needed to do was fix a few results to ensure that Botvinnik became champion. 
     In an article titled "The Russian Have Fixed World Chess" which appeared in Sports Illustrated in 1962, Bobby Fischer complained that the Russians prearranged draws against each other in order to conserve energy for play against him. He also complained that in their games against him the Russians audibly gave advice to his opponents. 
     There was enough suspicion of collusion that FIDE scrapped the old system in favor of a series of elimination matches. Years later after his defection Viktor Korchnoi accused Soviet players of cheating by ganging up on Westerners in tournaments and throwing key games when necessary. Or, was Korchnoi, who had no great love for the Soviets, just saying that to make them look bad? 
     I recently ran across a 2006 (revised in 2014) paper by Charles Moul and John V. C. Nye of Washington University in St. Louis titled Did the Soviets Collude? A Statistical Analysis of Championship Chess 1940-64. They examined whether the Russians intentionally drew against one another in order to focus their efforts on non-Soviet opponents in order to maximize the chance of a Russian winning. Read entire paper.
     They used statistical data from candidates' tournaments and USSR national tournaments for their study. The conclusion: “The likelihood that a Soviet player would have won every single Candidates tournament up to 1963 was less than one out of four under an assumption of no collusion, but was higher than three out of four when the possibility of draw collusion is factored in.” 
     Regarding Curacao 1962 in which Fischer accused the Soviets of colluding and later supported by Korchnoi, the authors opine that Fischer’s rating suggests that he was not the favorite “even if his rating/performances put him a few points shy of the leaders. Even if his true rating were higher (because Fischer had been improving rapidly in the year or two previous to the tournament) Fischer’s poor form/performance in the early rounds of Curacao meant that he had no realistic chance of winning. The Monte Carlo simulations do suggest, however, that he would have suffered from any collusion since his no-collusion probability of winning was 19 percent and it dropped to 3 percent with collusion. We did assume that all the Soviets were colluding and did not model the specific claim of Korchnoi that collusion was a private arrangement of Keres, Petrosian, and Geller.” In other words, Fischer wasn't yet strong enough to win at Curacao.
     They concluded that over the course of the Candidate tournaments the player most hurt by Russian collusion was Samuel Reshevsky. He was the favorite in the Zurich 1953 Candidates’ Tournament and his second place finish in the tournament was strong enough that even small collusive effects might have meant the difference between success and failure. 
      As the authors point out, all this is no surprise, but their paper provides “strong statistical evidence in support of this result.” 
     Back in 2005 Tim Krabbe disagreed with charges of collusion at Curacao, claiming there was no conspiracy at all. See his post number 299 HERE
     Let's take a look at the game in question, a 14-move draw between Keres and Petrosian that was played in round 25 (out of a total of 28). 
     It's significant because the game was one used by Fischer in Sports Illustrated and Life International to prove his point. His claim was that Petrosian agreed to a draw in a winning position. In the his book Curacao 1962, Jan Timman reached the same conclusion. I wanted to see if today's chess engines had anything to suggest, so I left Stockfish running on the final position for 90 minutes while I went out to cut the grass. Stockfish determined that black's advantage was about 1.25, probably enough to win at the GM level. 
     As a result of this game Keres and Petrosian remained in first place. Geller needed a win to stay in the hunt, but barely managed to get a draw in his game against Dr. Miroslav Filip.   Benko managed to convert his advantage against Korchnoi in a hard fought game. Fischer had the bye because Tal had withdrawn due to illness. 
     Timman's analysis of the final position was faulty, but that's not important because he didn't have a strong engine to check his analysis. Besides, I can imagine that writing a book where you have to analyze a hundred plus games doesn't leave a lot of time to spend on any one position. What is important is that his instinct was correct as was Robert J. Fischer's...black is winning, if not in Stockfish's world, at least he is in the real world. 
     My only question is if the Soviet players were colluding why wasn't Keres encouraged to lose this game? On the other hand Petrosian was pretty much assured of first place and at this stage of the tournament he didn't really need a win. Besides, it was his nature to accept a lot of draws and win only when he had to. 
     The real question is, did the Russians (as we called them in those days) really cheat? 

No comments:

Post a Comment