The Chess Mind

David writes, by email:

Dennis--

I'd be curious to hear your thoughts (either by reply, or preferably on your blog) about whether today's grandmasters have distinctive "styles of play" in the strong sense in which that term is often used to characterize the greats of the past. My curiosity was piqued by the following exchange on ChessCafe.com in Misha Savinov's interview of Levon Aronian:

MS: Do you have some of Larsen's traits?

LA: Yes, I like jumping with my pieces here and there and pushing the pawns. Normally it makes my position just rotten. But then I try to create some active play, increase tension, look for tricks. I enjoy crooked positions.

I found it refreshing to hear such a distinctive (self-)description of a player's style. We think of Tal as an intuitive attacker and lover of chaos; Petrosian as a master of prophylaxis and defense; Botvinnik as an iron strategist; and, yes, Larsen as an eccentric, ever-optimistic prober and unbalancer. But (regardless of the accuracy of those particular descriptions) my impression is that nowadays stylistic differences among top players are much less acute. Leko may be slightly more defensive, Shirov more aggressive; but the level of technique (and perhaps computer preparation?) are so high today that most top players seem to have an indistinct, or concrete and context-sensitive, or perhaps universal, "style." Their choice of openings seems to dictate the types of positions they play more than anything else. What tests might one devise to determine the differences in "playing style" among current top GMs? Do convenient labels come to mind to describe the style of, say, Adams, or Anand, or Bareev?

Anyway, curious to know your thoughts.

Sorry to see you were spammed by a Polgarophile.

--David

David, I think you're right that stylistic differences have grown increasingly less noticeable, thanks to deep opening preparation, the ease of assimilating information thanks to computer databases, and above all due to the increased knowledge and strength of elite players.

Thus with deep opening preparation, practically all GMs at least dabble in a large range of openings, looking to use specific theoretical ideas whenever possible. This cross-pollination and mass opportunism hides the most obvious way in which style can reveal itself.

Likewise, the increased level of knowledge and technical skill smooth out the stylistic edges. Player X might be, by inclination, a card-carrying member of Swashbucklers-R-Us, but given current levels of defensive technique, to say nothing of the ability to prepare (especially with the computer) against ultra-forcing approaches, player X will need to rein in his or her tendencies if she hopes to succeed. (Look at Alexei Federov's massive failures when he was trying to climb into the world elite. Perhaps he has the talent to make it there, perhaps not, but his regular use of the King's Gambit and Dragon Sicilian turned him into a full-point bye for the big boys.)

That said, I still think there are stylistic differences. Look at Kramnik's 1.e4 games the past several years. He won quite a number of nice games, even in sharp Sicilians, but the way he won them had a different feel from that of a Kasparov or Anand.

But what is style? In my view, it's not a simple property but a composite of many factors: one's openness to risk, whether one prefers trying to win over avoiding a loss, whether one prefers rigid pawn structures to flexible ones, preferences for static vs. dynamic positions, preferences for outdigging in familiar fields vs. locating new ones, aggressiveness vs. solidity, the degree to which one is willing to endure tension in a position, and so on.

Of course, any strong GM will opt for a certain pawn structure, even if it's outside his usual range of comfort and special expertise, if he believes it is the best choice in a given position. But I think these preferences, and no doubt many others, which reflect both psychological dispositions and areas of particular excellence, will come to the forefront whenever no obviously best plan or move or idea suggests itself.

Returning to your initial example, David, of Aronian's "crooked" style, it will be interesting to see if Aronian can maintain his elite rating once he's playing with the Linares crowd, and if he does, if that happens because he "rounds off" his style.

A final thought in the form of a question. Is "style" something positive, or is it a way of saying one plays some positions and handles some situations less badly than one handles other positions or situations? Suppose I'm a pretty effective attacker, able to beat even titled players regularly by engaging in head-hunting, but a mediocre "positional" (a nearly meaningless term, in my opinion) player and a lousy technician. As the term is used in amateur chess, at least, I would be described as a player of attacking style. Suppose that my positional and technical skills develop to the point where I'm just as effective in those kinds of situations as when I'm attacking. Let's further suppose that those are the only three elements of style, and they occur with approximately equal frequency in my games. Do I have a style any longer? (One might say I have a "universal" style, but I'm not sure that it really signifies anything.)

Greg [last name withheld to preserve the secrecy of his repertoire!] writes:

Dennis, I will be returning to tournament play in November after a very long hiatus from the game. I have been using Chessbase CD-Roms and a few books (Play the French, 3rd edition) on the French Defense, Chigorin QGD, and Colle System to try to prepare for this event. I like openings that don't require massive preparation, allow me to minimize my opponent's options, and are reasonably sound. I don't want to try to memorize all kinds of book lines and I'm wondering if you have any suggestions about the best way to learn any or all of these openings. My last published rating was 1652, and I've been working with Junior 9 to get back into practice.

Thanks. I enjoy your blog.

Greg

Greg: I'm not sure you can play the French well without some pretty serious work, but let's waive that and turn to your question: what's the best way to learn these (or any) openings?

Here's my quick answer, which I suspect I've given before on my blogs (but couldn't find). First, pick one or more "heroes" for the opening in question. (Ex. Kasparov, Anand, or Gelfand in the Najdorf; Kramnik in the Berlin Defense; Short in the QGD for Black.) Create a database of their games, and play through them several times, quickly, to get an initial sense of how the opening and resulting middlegame tends to look when played by its leading practitioners. You'll notice certain move order finesses and typical middlegame plans, and that's not bad for starters!

Second, if you have a book on the opening in question, look through the main lines and model games. See what your opening should look like when everything goes right - and when everything goes right for your opponent, too. Pay attention to the key points emphasized by the author; less attention to fine points and memorization.

Third, start playing the line, especially in blitz. That will help solidify what you've seen, and make more personal and real the problems you need to solve. Compare what happens in your games with the heroes' games and with theoretical sources (ECO, NCO, an opening monograph, a computer book, etc.), not just to confirm your move order but the middlegame follow-up as well.

As you work up a track record of playing the line and examining your performance, it becomes time to start working on the details. Here, finally, it's appropriate to do a little memorizing of key lines and move-order finesses.

To recap: start with the big picture, with ideas. Develop some experience to see what the real problems are and to gain a feel for the line, and only then, after taking ownership of the line for yourself, should you start to worry about serious theory.

Hope that helps!

I've been avoiding mention of the recent stories about the arrest of Colorado master and chess instructor Robert Snyder (for sexual assault with minors), because if he's found not guilty (and is in reality not guilty), the wrongfully inflicted damage to his reputation is likely irreparable, and I don't wish to be a party to it.

However, that cat is out of the bag all over the world of chess blogs, and since I have a word or two of advice that might be of use, here goes.

First of all, if you're a chess teacher, REFUSE to be the sole chaperone of ANY kids you didn't bring into the world. I have turned down several opportunities to earn some extra income by driving kid x to a tournament, and I simply won't do it. (I'm referring to driving a male child. All this gets raised exponentially if we're talking about a female student.) If the child is a minor, then even he's big enough to physically damage me, I'm not going anywhere with him in the absence of another adult, preferably one of his parents. Everyone stays out of trouble that way, and if you're a good teacher, you'll make plenty of money anyway.

Second, for parents: follow this same rule. If you can't bring your kid to a tournament, then he doesn't go. Or if he or she does go, this only happens when there are multiple adult chaperones, who do not stay in kids' rooms. In fact, I'd say that unless there's a medical emergency, no adult should ever be by him- or herself in a kids' room. Further, unless these chaperones have been investigated by the school district, forget about it. A freelance guy like me should NEVER be the chaperone, even if I'm with my wife, another chessplayer or instructor, whatever, if it involves situations where I could be out of the public eye with a child. (Part of a group, sure, as long as the group has some sort of worthy accreditation and, again, no one is ever alone with the kids.)

All this seems like common sense to me, just as (male) pastors should never counsel women behind closed doors, teachers should never be in an office with students of the opposite sex behind closed doors, etc. Instructors (or parents, if they take off work to take their kids) might lose a little money, and kids might miss out on opportunities to play every now and then, but that's life. Chess teachers don't have to worry about false rumors, and parents will have a lot less to worry about with their kids.

In round 1 of the ongoing Eastern Open in Connecticut, my friend Brian Karen, a longtime high-2000/low-2100 player, achieved his first ever non-loss to an IM or GM in a tournament game, drawing GM Sergey Kudrin.

I'm happy for him, and the most interesting feature about the game is just how uninteresting it is! With White, Brian played a normal, solid opening, and Kudrin, a good, tournament-tested GM, was unable to achieve anything at all. Brian didn't have to play something crazy or obscure to trick Kudrin, didn't need some cheesy "Winning with Crappy Gambits" volume or video - nothing like that. On the other hand, he didn't go in for an absolute main line, but a solid second-tier variation. So one's choices aren't limited to becoming a theory zombie or going the trickster route; there is a middle way, especially with White against higher-rated players.

Take a look here and see for yourself!

Time for a brief foray outside the world of chess. (Those interested in chess alone can take a nap and wait for my next post.)

Intelligent Design Theory (IDT) has gained both fans and foes in the United States, and states, in brief, that it is possible to detect design in nature by employing a three-step filter. First, can some object or phenomenon be explained as the product of physical law? If so, then we stop: design cannot be inferred. If not, we proceed to step two: is it the product of chance? Again, a "yes" answer puts a stop to things, but a "no" allows us to conclude, or at least seriously consider, the possibility of design.

This so-called "design filter" is an idea of William Dembski's, and is well-illustrated by one of his favorite examples. It turns out that in elections, the first name listed on a ballot receives more votes, all things being equal (pathetic, isn't it?), so name placement it supposed to be determined by random processes. Take, however, a case in which one party winds up with its candidates listed first 40 out of 41 times in a given election.

To evaluate whether this is a case of fraud by the election supervisor, let's use the design filter. Was the ordering of candidates the result of some lawlike process? Not according to the supervisor, who claimed to have followed the rules by using a randomizing process. Ok, we move on to chance. Is it reasonable to expect that this could have arisen by chance? It's not impossible that it did, but as the odds against its happening were less than 1 in 50 billion, the best explanation was design: the supervisor engaged in fraud.

Dembski and IDT advocates wish to apply this to the biological realm. An organism's having certain features isn't the result of biological law, but of random genetic mutations. So it's not law, but, as the "random" part of "random genetic mutations" would suggest, a likely candidate for chance. But how much chance is acceptable? We must note that the degree of chance will be a function of the odds of a single event's occuring multiplied by the number of trials (our "luck resources"). (Thus the odds of flipping heads ten times in a row with a fair coin is 1 in 1024; pretty unlikely in a single trial, but overwhelmingly likely to happen many times over if hundreds of people are flipping coins all day.)

Determining how likely or unlikely an organism's having some feature really is will be a matter of great dispute and is the job for scientists and mathematicians. What can be said, according to Dembski, is that if it turns out that the probability of that feature's arising by chance, taking all the "luck resources" into account, is beyond some arbitrary but extremely small figure (1 in 10**150), then we can legitimately infer design.

Compare this with Young Earth Creationism (YEC), according to which the Earth is about 10,000 years old, "kinds" don't evolve into other kinds (i.e., while there is a great variety among, say, dog breeds, one will never get from a dog to a non-dog), geological strata aren't the product of billions of years but a worldwide Noahic flood, and human beings didn't evolve from non-human ancestors.

Both Intelligent Design Theorists and Young Earth Creationists believe that the biological realm is the product of a designer, and most advocates of these views are Christians. Further, like proponents of YEC, some of IDT's most prominent figures critique what they see as overoptimistic claims by the friends of (naturalistic) evolution. But does that make these two views "essentially" the same? One sometimes reads that IDT is "really" Creationism (by which the author has in mind YEC), but is this really so?

I don't think so.

I offer here several considerations to indicate that IDT and YEC are clearly conceptually independent - the differences are not merely cosmetic, but substantive.

First, IDT, as I understand it, does not intrinsically reject evolution, taken as a descriptive hypothesis about the lineage of organisms on Earth. One could in principle accept the thesis of universal common descent and hold that there is good evidence for intelligent design - in fact, I believe that is Michael Behe's position (or at least was at the time of Darwin's Black Box). That X can be designed by S even if Y is the proximate efficient cause of X is something regularly recognized in the case of human design (take, for example, the manufacture of cars and computer chips), and some, perhaps many, perhaps most ID theorists would say the same of the world's design. The young earth creationist, on the other hand, would not consider evolution as even a descriptive thesis, and that's a huge difference between the two.

Second, it seems to me that YEC is more closely tied to science in its proclamations than IDT. I don't mean by this that it is more accurate than IDT or that its scientific claims are acceptable - I'm not making any evaluative claims at all. Rather, my point is that YEC makes claims of a more empirically immediate sort than IDT: dinosaurs and humans co-existed, a worldwide flood explains geological strata, we can make inferences about mutation rates by examining the biological diversity that has occurred since the time of the Noahic flood, etc.

IDT, on the other hand, is something more like a philosophical hypothesis, or perhaps a meta-scientific framework. It's hard to cash IDT out in terms of concrete empirical predictions, and I'm not sure that it should. What it does claim is that design can be detected under certain conditions, and then looks to discover where those conditions have occurred.

Third, while YEC posits the God of the Bible as the designer, the IDT's designer (a) need not be the biblical God, (b) need not be omnipotent, omniscient, etc., (c) need not have created the universe, and (d) need not have any particular interest in humanity. (One might take Hume and Plato as examples of thinkers able to conceive of designers that, while immensely powerful and impressive, fell short of the God of traditional Western monotheism.)

Fourth, while the epistemic origins of YEC are religious (irrelevant though that is, if it can deliver the goods), there is nothing inherently religious about the design filter.

Fifth, the truth conditions for the two theories are different. (1) IDT could be true while YEC is false, and (2) YEC could be true while IDT is false. It's very easy to see how (1) could be the case, and (2) could occur at least two different ways. First, while it could be that the world has been designed (by the God of the Bible), Dembski's design filter might be a conceptually flawed way of detecting design. Second, YEC might be true, but its truth cannot be detected by Dembski's design filter (perhaps because the probability against a chance hypothesis doesn't quite reach the upper bound).

In short, the two are distinct in method, in how they interact with evolution (taken descriptively, divorced from a naturalistic metaphysics), in their predictions, in their empirical claims (and the lack thereof), in their religious implications and in their truth conditions.

They are simply not the same thing.

(N.B. for possible commentators: I am not endorsing (nor rejecting) either view in the above, but am arguing that the two cannot reasonably be equated. Second, I gently remind my readers to maintain appropriate decorum, regardless of their feelings towards either view, lest you suffer the same fate of the recent Polgar hyperfan!)