Who will win Wijk aan Zee?

So who will win already?

Before the start of this tournament we asked you to predict the three top finishers in the Corus 2006 A Group. For the best prediction there will be a prize, probably a copy of Fritz 9 signed by the three winners he or she had selected.

We received over 2500 replies, which we had to evaluate mechanically. Unfortunatly, in spite of our admonitions, there were hundreds of names wrongly spelt (how can you misspell 'Anand'?), which proved confusing to our evaluation module. Finally we asked for the exact score the three winner would be predicted to receive. This will be used as a tiebreak if there are many correct predictions. Here again many simply ignored that. We have evaluated their bare predictions in the following table, but they will be heavily handicapped in the tiebreaks.

As you can see from the above our visitors, in a majority, picked the winners according to the seedings in this event. Topalov, Anand, Aronian, that is what the ratings say and that is what people expect. The first player people tend to evaluate below his FIDE rating is poor Boris Gelfand, and there are quite a few people out there who expect Kamsky and Karjakin to play above their nominal rating. And apparently nobody expects one of the Dutch grandmasters to take any of the top places.

Well, that is what we mortals think. How about a professional? One of the replies we received was from Jeff Sonas, a statistician who has done extensive work on chess. Jeff sent us his opinion in the form of an article with a full breakdown of each individual player and his expected score.

Statistical analysis of Corus 2006

By Jeff Sonas

It seems pretty clear that the two favorites at Corus are Viswanathan Anand and Veselin Topalov, although it’s not clear which player is indeed more likely to win. My Chessmetrics ratings point to Anand as the most likely winner, although Topalov did have more significantly success during the year 2005. In any event, it seems very likely that there will be no runaway winner at Corus.

In the table below, I have listed the FIDE ratings for all 14 players (including their world rank on the rating list), as well as their latest Chessmetrics ratings, and a new kind of rating, which I am calling a “performance measure”. It completely ignores anything that happened prior to 2005, and only looks at players’ results across the single year 2005. Roughly speaking, it’s a performance rating calculated across the whole year, although there are some improvements upon the traditional performance rating calculation. I have adjusted all performance measures in this table so that they are directly comparable against the FIDE and Chessmetrics rating lists. You will be reading a lot more about this performance measure in an article I have coming out in a few days on ChessBase; I strongly believe that it should be used as a primary means for determining world championship candidates, rather than what FIDE actually used.

But by any measure you want, Anand and Topalov are 35-40 points stronger than anyone else in the tournament and have to be considered the favorites, with Levon Aronian probably next in line after those two players. One big question concerns former world championship candidate Gata Kamsky, who still has a high FIDE rating thanks partially to his great results a decade ago, but whose results since coming out of retirement suggest that his level of play is more like #100 in the world rather than the #25 suggested by his FIDE rating. Nevertheless, he had great results at the recent FIDE World Cup, and did qualify for the FIDE Candidates matches. It will be fascinating to see how well Kamsky can do against elite opposition at Corus.

Jeff Sonas is a statistical chess analyst who has written dozens of articles since 1999 for several chess websites. He has invented a new rating system and used it to generate 150 years of historical chess ratings for thousands of players. You can explore these ratings on his Chessmetrics website. Jeff is also Chief Architect for Ninaza, providing web-based medical software for clinical trials. Previous articles: Championship Chessmetrics Analysis The Sonas Rating Formula – Better than Elo? How strong are the top chess programs?