Greengard's ChessNinja.com

Playing with Fire

| Permalink | 24 comments

If a mantra of the 60's was "don't trust anyone over 30," maybe we need our own version: Never invite anyone over 20." Just look at the challenge match between US champ Hikaru Nakamura and Sergey Karjakin in Cuernavaca, Mexico. (Nakamura won game five and the match yesterday.) With one game still to play, four out of five games have been decisive and the only draw was an exciting 60-mover that ended in a dead rook endgame. It's enough to help you forget this year's Linares and Dortmund draw-fests.

Of course there are reasons for this beyond fighting spirit. Most youngsters don't have the hyper-prepared openings that lead to so many drawish positions or the technique that means drawn endgames are drawn. But attitude gets most of the credit. "Black to play and draw" hasn't sunk in yet. Their innate optimism leads them to play to win inferior (or equal) positions instead of trying to draw them asap.

This doesn't need to be the latest log on the "how to solve the draw problem" fire. Draws are a natural part of the game. But I believe that a few years of legislation could lead to a needed mental sea change. Eliminate draw offers, see what happens. More extreme attempts can wait. Prohibiting move repetitions, as in Shogi, is a bit much. (Think of the effects: no more sacrifices for perpetual check.)


What an amazing slugging match! The last time I saw anything like it was Kasparov-Short 1993.
I can't help but feel a little bit sorry for Karjakin, though...

The quality of this match has been completely horrible. Instead of a "drawfest" we have a blunderfest worse than anything in, for example, Tripoli. Some might prefer this to real high-quality chess (everything is of course relative here) and I have no problem with that. Others like me will have to wait for Corus and Linares. I'd rather see Kramnik and Leko draw 20 times.

What? You have to post analysis when you pop off like that. The games have been very good, certainly no blunders as bad as Kramnik's against Leko in the Marshall. (Karjakin's rook swap in the game five endgame was probably the worst. Not sure if he thought it was drawn or just lost no matter what.) None of the games would be out of place at Corus and would only be out of place in Linares because they weren't 25 moves of theory, three original moves and a draw offer.

When players play to win they make mistakes, and that's at any level. Playing it safe and drawing a boring a position in 28 moves and saying it was high-level chess is garbage. No one need apologize for style, but saying you draw a lot because your chess is better and you see everything is silly. Aggressive players like Anand, Moro, and Kasparov can play to win without playing bad chess. This involves risk.

"You have to post analysis" Is that a rule here? I'm sorry then, I won't do that. Too much work and everyone can look at the games themselves. Sure, the games have also been "very good". Compared to mine, for example. It's all relative.

It's not a rule, just asking on what the comment is based on. We do aspire to a minimum level of rigor around here. From the GM-annnotated bulletin, Kasparov's comments to me that Nakamura has been playing very well, and from my own analysis of the games, I think the level has been quite good. Game one in particular was impressive and original in many ways. So when someone says it's a blunderfest I wonder how strong they are or where they've been getting their analysis.

I don't know (or care) how good the games have been. But I do know they have been very entertaining.

And yes, I agree the main reason is that they haven settled on those ultra-safe opening repertoires of the top players.

Hey, I find myself watching the games! I never felt like watching the Kramnik-Leko match (even though I tried!).

Good match!

"Prohibiting move repetitions, as in Shogi, is a bit much. (Think of the effects: no more sacrifices for perpetual check.)"

King and Pawn vs. King becomes a win in virtually all cases, and cascading effects on all other endgames. It would be a completely different game than chess; IMO a more boring one.

It's not the same as eliminating stalemate. From my understanding of the way it works in Shogi, it's the initiating side that must diverge from the repetition. That does away with sacrifices to create perpetual check, although I suppose in some cases the weaker side is forced to make a losing move instead of a repetition. This is one reason why Shogi games tend to be so brief, tactical, and dynamic. I wouldn't want that for chess, although experimenting with it couldn't be worse than, say, shuffle chess and other variants.

I had a long dinner conversation with AF4C honcho Erik Anderson in San Diego about potential rule changes to create more decisive results. I'm sympathetic with that desire, but I try to work against shortening the time controls and making changes to the game itself.

Actually this is a good topic for a separate post to solicit contributions to the "things that DON'T work" list. It's enough for chess fans to say "but we have to keep good chess," but sponsors often don't understand that rationale. With the spectacularly hard-fought 2005 US Championship we just saw I'm hoping anything radical is off the table for 2006.

"Certainly no blunders as bad as Kramnik's against Leko in the Marshall".
What?? You must be kidding! Oh, but maybe you saw immediately that 23.Qf2 was a decisive error because of 26...Qe3+ 27...Bxf3+! and 29...Ng4.

That game was really just one big blunder from Kramnik's side. Extremely embarrassing. "I thought we had computerchecked it", lol. No denying that there are such moments even in the absolute top. Some of Kasparov's stupid endgame blunders lately come to mind as well. We're still just talking rare occurrences though.

The thing that makes "quality" hard to evaluate is the time situation in each case. A 2600 will play better with time than a 2700 without time, at least in most cases. Kramnik's blunder was horrid for being preparation and for being in the middlegame with plenty of time. It wasn't an error of thinking, but of not thinking.

Did *I* see it immediately? Of course not. But had I arrived to that position from either side I would have found it given time, just as Kramnik would have had he looked at the board at all. It's clear that such sacrifices are possible in the position and you don't have to be a GM to count beans once the idea is seen.

Regardless, why would we measure blunders based on what we would see instantly in a GM game? The mark ?? means a single move that changes the best-play result of the game. That's a relatively objective expression. A 1400 player can find blunders in GM games if he has enough time, not even counting computer analysis.

Fritz doesn't find the error because it's a pretty long combination, hardly what we normally understand as a blunder in chess (i.e. a silly move).
Of course, moving quickly without checking the computer analysis because of Leko's time pressure can also be called a blunder, but in a completely different sense (not a silly move, but a silly decision).

In Shogi, draws by move repetition (sennichi-te) are not illegal, and if the same position arises 4 times from the same sequence of moves, the game is declared drawn by repetition. I believe that such draws are quite rare in actual practice.

The only exception to the "move-repetition draw" rule described above is where all of moves by the player forcing the repeated positions are checks (i.e. in Shogi there is no such thing as perpetual check). If a player is careless enough to check the other player into the same position for a fourth time, he/she is actually loses due to making an illegal move.

There are also other ways that Shogi games may be drawn, e.g., when both sides have "entering kings." Nevertheless, it is definitely true that a draw is a very rare result in Shogi. The typical Shogi game consists of some opening moves where the pieces are being developed and the kings are "castling" (castling is a bit different in Shogi, but the idea is similar), some middlegame manoevering, and then a race between two mutually-unstoppable mating attacks. As a result, the normal question in a game of shogi is who can mate first, not whether either side will be able to mate at all. This results partly from the recycling of defenders captured from the opponent to refuel the attack, and partly from the rule against perpetual check (i.e., because it is virtually impossible to bail out during an attack). Draws in the similar chess variant Crazyhouse are also quite rare, and would be rarer still if perpetual check were illegal.

There is also a rule in Shogi that drawn games (rare as they are) don't count at all, and after a draw the two players must replay the game with colors switched so as to obtain a decisive result.

The recycling of pieces that occurs in Shogi cannot, of course, be duplicated in chess without radically changing the game. And disallowing perpetual check as in Shogi would not change the draw situation in chess much because only a small percentage of games end in perpetual check. The "replay drawn games" rule in Shogi is probably workable only due to rarity of the triggering event, and would not seem to be practical in a game such as chess where draws between strong and closely-matched players are relatively frequent and where such a rule would cause major scheduling problems.

So I am not sure what inspiration can be drawn from Shogi to help decrease the number of draws in chess. Shogi is simply, by its nature, a game highly likely to lead to a decisive result, and the same cannot be said of chess when played at the highest levels. Perhaps the rules in chess will have to be modified at some point future to avoid endless drawfests among top players. Kasparov's idea of using a different starting position every year is very interesting (although it would probably virtually eliminate the market for opening books). Perhaps another short-term solution to short draws is in the area of financial incentives for fighting chess. If half of the prize fund in every major tournament went toward rewarding fighting spirit instead of point totals, professional players would probably have sufficient incentive to avoid most short draws.

I don't think it's really necessary at present to change the game of chess to reduce the number of draws. Something like increased financial incentives for fighting chess would probably work quite well at the moment, as even the top players can still play fighting chess when they are sufficiently motivated. But in the long term, perhaps 20 or 30 years from now, competitive chess as we know it may well die a natural death. This will not necessarily be a great tragedy, as there are other games like Shogi which are not experiencing the same problems. Shogi is very "chess-like," uses similar mental skills as chess, is more complex and arguably more exciting than chess, and will probably not be killed off by excessive draws (or excessive opening knowledge)in the foreseeable future. It is really a matter of historical accident (and perhaps partly due to the use of traditional kanji sympbols on shogi pieces, which can be somewhat offputting for Westerners when they first try to play) that shogi is not more popular in the West.

"It's not the same as eliminating stalemate. From my understanding of the way it works in Shogi, it's the initiating side that must diverge from the repetition."

Right, and the implications of that are just as far-reaching as eliminating stalemate. Consider this elementary ending: White, King e5, Pawn d5; Black, Ke7. White to play _wins_. 1. e6+ Kd7 2. Kd5 Kd8 3. Ke5 and now Black can't play 3...Kd7 because it repeats the position after the 1st move, and so he loses the opposition and the game. (Or 1...Kd8 2. Kd5 Kd7 3. Ke5 and White wins the same way.) A non-repetition rule completely changes all endgame theory.

"With the spectacularly hard-fought 2005 US Championship we just saw I'm hoping anything radical is off the table for 2006."

I agree completely. I suspect that the biggest difference was that the main prizes weren't split (if I'm reading the rules page right). That creates an incentive for the player with worse tiebreaks to fight hard for the win. It probably helped that Mr. Anderson was known to be serious about rewarding fighting chess, too.

I believe that any successful solution to the quick-draw problem will involve incentives to fight hard, rather than punishments for drawing.


Oops! Meant 1.d6+ in my little example, of course. Anyway, just consider me as violently agreeing that "Prohibiting move repetitions, as in Shogi, is a bit much." I worry that people who _do_ support such stuff aren't thinking it through very well, though.

I am starting to think that a Kasparov-Anand rematch would be really more relevant. Of course Garry would never allow it as Anand is in terrific form. We should all remember that these times are not that unusual for chess. World champions have a history of avoiding their closest rivals. In the last 30-40 years we have been spoiled in seeing matches between the two best players in the world. But Lasker, Capa,Alekhine, and Botvinnik, and Bobby (he is still "champion" after all) all did such evading on a regular basis. It's just a fad though(combined with insecurity on Kramnik's part)... I hope.

Mig, If it has been posted somewhere I apologize but I have seen no explanation of the Larsen prize in the US Championship. How was it decided? Did it produce the desired effect? Too lazy myself to do stats tho cursory exam leads me to believe what I have said before: pay players to win and they will! Of course, when you're young, or especially when you're young and trying to make a name for yourself you play like a daredevil - as you should.

DP: What are you smoking? Look at Anand's score against Kasparov - if one of those two is afraid of the other, it must be Anand.

And who did Botvinnik evade on regular basis? He played the best challengers available for over 20 years.

Mig: " A 1400 player can find blunders in GM games if he has enough time, not even counting computer analysis." - I doubt that. Unless he really takes his time - like 6 months during which he gets to 1800 or so.

I have to disagree with Russianbear. I think Anand has moved to another level in the past couple of years, and you haven't seen many games between Anand and Kasparov during that time, have you? I don't think you can just look at old results for validation. Would Anand beat Kasparov right now? I don't know, but I think it would be a far greater match than their last one.

While more time normally leads to higher quality, the quality of the game itself is independent of time. For example a five minute game between Kasparov and Kramnik will always have a higher quality than a game between two amateurs without any time restraints.
In chess, quality has to do with complexity. In that sense, Karjakin's 39.Rf7 in the fourth game (which is instantly met by Bb5) is certainly a bigger blunder than Kramnik's 23.Qf2.

I have to say that I enjoyed both the fighting spirit of the games in Cuernavaca and the high quality (albeit not the length, alas!) of the games in Brissago.

When judging the quality of games, I think it's appropriate to look at the external context. Kramnik played 23.Qf2 out of his preparation. In a World Championship match, that type of mistake is not acceptable. It deserves a '??'. Had Karjakin played the same move against Nakamura, we would not (and should not) judge it as harshly.

What challengers did Capa evade as WC?

Hey mig,

after the blasting you gave acirce earlier in this thread for his comments about the quality of the Nakamura Karjakin match, i found the following para in your report at chessbase.com puzzling:

"Nakamura said the level of the games was well below his best efforts from the US Championship. Considering Karjakin's devastating form in other events this year, the result, and the spotty quality of the games, is something of a surprise"

AS far as Kramniks blunder it seems to me that many GMs may study a position with a computer and then memorize the line. The computers don't seem to catch the in this line as quickly as you would think. Its interesting because the blunder really had little to do with chess playing. Personally I would feel much better if I were kramnik and made that kind of mistake than if I just started miscalculating postions.

Twitter Updates

    Follow me on Twitter



    About this Entry

    This page contains a single entry by Mig published on December 14, 2004 4:47 AM.

    Official Bash Kasparov Thread was the previous entry in this blog.

    Krasenkow Cardiogram is the next entry in this blog.

    Find recent content on the main index or look in the archives to find all content.