Bilbao, Rounds 6 & 7: The Magnus Monster is Back
After the first cycle in Sao Paulo, Magnus Carlsen looked like an also-ran - lucky to be in clear third place, "only" five points behind the leader, Fabiano Caruana. Two rounds into the second cycle, in Bilbao, he and Caruana are tied for first, and he's not far from catching Garry Kasparov's all-time rating record. (Prescinding from any possible rating inflation, anyway.)
All the non-Carlsen games in the second cycle of the "Final Masters" have been drawn (just three more Viswanathan Anand will keep his title!), but the Norwegian terror is again working his magic. In round 6 he had the white pieces against Caruana, and this was pretty close to a must-win game for him. He rose to the occasion, and in his usual style: a semi-theoretical opening (the King's Indian Attack vs. the French, in this case), using the middlegame to induce some enemy weaknesses, which leads to persistent pressure in the ending followed by the opponent eventually cracking. Check, check and check. Caruana maintained material equality for a long time, but the bishop ending they finally reached was hopeless. Curiously, it was a sort of mirror image of Fischer-Keres from Zurich 1959. There the players had light squared bishops, and Fischer had the only two pawns, on the f- and h-file. In Carlsen-Caruana, they ultimately reached an ending with dark squared bishops and white c- and a-pawns! In both cases the only real task was to prevent the defender from sacrificing his bishop for the bishop-pawn (the f-pawn in Fischer-Keres; the c-pawn in Carlsen-Caruana), and the winners coped in both games.
Then today, Carlsen had White again; this time against rating tail-ender Francisco Vallejo Pons. Once again Carlsen left the beaten tracks of opening theory most expeditiously, meeting today's French Defense with 4.exd5 against the Winawer. That isn't a completely theory-free zone, but Carlsen avoided most of that theory as well. The difference between this game and yesterday's is that Vallejo didn't make it to an endgame. Carlsen maintained pressure against Black's kingside, and eventually broke through and won.
Here are the pairings for round 8, with scores (remember, 3-1-0 scoring) in parenthesis. There are three rounds to go:
Karjakin (5) - Carlsen (12)
Anand (7) - Caruana (12)
Vallejo (4) - Aronian (9)
Reader Comments (7)
I have heard Carlsen compared to many players, but after his win over Caruana, I thought he most resembled Capablanca. What do you think?
[DM: Yes, I think he's in the broadly Capablanca-Smyslov-Karpov "family".]
A good writeup!
[DM: Thanks!]
I was very impressed by both those games. If you play well, Carlsen will ever so slowly get his pieces on better squares and win. If you play ever so slightly less than well, you're in trouble. Impressive comeback.
Rating inflation theoretically does not exist. The distribution of ratings is the same now as it's ever been. Players are getting stronger. Please see IM (and U. at Buffalo Professor) Ken Regan's paper on the topic: http://www.cse.buffalo.edu/~regan/publications.html#chess, ``Understanding Distributions of Chess Performances". A PDF is available at the site. In other words, if Carlsen breaks Kasparov's record, I think we can all hail him as the strongest human ever to live.
[DM: Long time no see, Howard. I know Ken (he is a regular commentator on this blog) and know his IPR paper, and have mentioned it on here many times. I have tremendous respect for him and his work, but still have some doubts; in particular, I'm not sure he has fully taken the depth of contemporary preparation and opening knowledge into account. Put differently, today's players have a big head start from Houdini & friends, and I'm not sure that his model fully captures the extent of that head start.]
Hi Dennis,
thanks for your report on Bilbao and on Carlsen´s ideas against the French. He is one of the good guys - never shy to try out something unusual. Quite courageous!
It´s a pleasure to read through your articles.
Bye, best wishes
Tiger-Oli
[DM: Thanks!]
Hi Dennis!
Your text on Carlsen was definitely inspiring - after playing through his game against Pons, I just could not stop and wrote a little article for our blog in Germany. If you want to take a look - it´s about the end of the French Defence, brought about by Carlsen´s brilliant move 4.e4xd5. :-)
http://www.schach-welt.de/BLOG/Blog/Aurevoir,Franz%C3%B6sischeVerteidigung
Bye, thanks, and have a good time!
Olaf
Long time no see, as well, Dennis. In my absence, I must have missed your references to Ken Regan's work. Anyway, Figure 1 in the paper I referenced does not really refer to any engines. It merely deals with population statistics and represents a model for how many players rated over 2203 should exist in 2012, based on some set initial conditions.
From the paper: "The theory of the growth of a population under limiting factors has been successful in other subjects, especially in biology. Since the work of Verhulst [10] it has been widely verified that in an environment with limited resources the growth of animals (for instance tigers on an island) can be well described by a logistic function ...
[math deleted]
We demonstrate that this classic model also describes the growth of the total number of chess players in time with a high degree of fit."
Regan does, however, write a lot about using engines to objectively rate chess players. He compares these objective ratings over time. But I believe Figure 1 of this paper is cited merely as independent support for Regan's argument, and it's this third-party work that I'm citing as an argument against rating inflation.
Here's another way to think about it. Let's just say skill is distributed on some bell-like shaped curve. Maybe not a true bell curve, but there's a peak or two near the center and tails on the edges. As you add more samples to any bell-like curve, the tails will creep out farther to the edge, based on simply larger numbers.
Hg