Kramnik Interview

Comments

>>Always nice to see the liberal use of my photographs from Bahrain.

They know you're the easiest going guy; that's why they don't even bother to give you credit :D
Posted by: ikalel at June 10, 2006 23:17

I cannot see the interview? why?
Posted by: Apple at June 11, 2006 02:19

I love his way back to form: He read a book about how he used to play, and decided to go back to that! Now if someone could write a book about me...

Q
Posted by: Quely at June 11, 2006 05:23

I found the following Q and A as a confirmation that Kramnik no longer is in any opposition to Kirsan or Fide:

Q: The contracts of your match with Topalov have been signed; the venue – Elista – and the dates have been decided. Will this match definitely close the schism that has been hanging over the world’s chess scene for the last years?

A: Yes, that is the whole point. This was supposed to be done already a couple of years ago, under the Prague agreements, but for some reason it didn’t happen. But I think that this match will definitely be good for chess. This situation is not good for everyone, not for me, not for Topalov, so I think that this is in the interest of everyone. I don’t know what will happen in the future, I don’t know if it will be useful or not, although I hope so. It gives a lot of opportunities to build a very strong road of chess.

End of quote. Yes, Kirsan and Fide is what we got, let's make the best out of it.
Posted by: Akselborg at June 11, 2006 07:09

There is no question in my mind that Kramnik is the better match player. Topalov should give him a good fight though. Like Leko did.
Posted by: Morrowind at June 11, 2006 07:54

Hey why did my comment that Kramnik is the better match player, and that Topalov should give him a good fight, like Leko did. Not post???
Posted by: Morrowind at June 11, 2006 07:56

oops!!
Posted by: Morrowind at June 11, 2006 07:57

After reading the interview and being honest with myself, I now have mixed emotions about the upcoming (hopefully) World Championship Match. I realize I've been hard on Kramnik when in fact I should have shown some empathy.

I know what it's like to become addicted to pain killers; in my case vicodin and what it can do to your level of play. I now believe that his treatment for that is proving successful and his previous form appears to be returning. I hope it will be a complete success and that the match proves to be a battle royal of opponents with contrasting styles of play.

I will continue to root for Topolov only because of his brilliant play as of late, but I see it as a much closer competition now and I am anticipating a great match.
Posted by: chesstraveler at June 11, 2006 12:37

Hm... I thought he lost 3 out of last 4 games with Topalov and drew one. What was he talking about beating him? Did he mean in Amber Monaco????
Posted by: Meke at June 11, 2006 13:52

Is Kramnik's treatment for his illness significantly expensive? He is still in treatment now and probably will be again in the future. He needs big paydays. Is this one of the reasons why he agreed to this 'reunification' match? Putting his title on the line in this FIDE match (and subsequently in an 8-player tournament, if he wins) will only degrade his title.
Posted by: ikalel at June 11, 2006 14:14

Kramnik has already been asked about this in an ACP interview. This is what he had to say:

Already in San Luis Veselin Topalov agreed to participate in the next FIDE world championship, which will take place in Mexico according to the same format – a double round-robin with eight participants. Having signed a contract for playing a match against Topalov, in case of victory you’ll have to replace him in this championship. Did you agree with this?

I think that a classical format of the world chess championship should be preserved in any case. In my opinion, the title won in a world championship match between two strongest chessplayers outweighs the title won in a tournament. I’m quite conservative in this regard. I would certainly preserve the world championship matches if it depended on me. However FIDE has different views and different plans. Well, I can play in the tournament...

Vladimir, I don’t speak about your principles and preferences, a law aspect of this case matters – having signed the contract you will have to play in the “Tournament of the eight” in case you defeat Topalov in the match.

For the moment it’s more important to settle all these title confusion, which started in 1993 and was not stopped in 2002. It is of such great importance that I made concessions for this. It’s okay! Do you want me to participate in the tournament? I will do this for the world to have a unified world champion!

http://chess-players.org/eng/news/viewarticle.html?id=523
Posted by: ikalel at June 11, 2006 14:53

ikalel, what title are you talking about? The was from 2000 has expired long ago... also, I think Kramink has only to gain from this match as he is hm.. let see, #7 or so at the moment?! I perfectly understand that other high caliber players are mad about him getting a shot at Topalov's crdible title so easily.
Posted by: Meke at June 12, 2006 11:35

Here we go again...
Posted by: acirce at June 12, 2006 11:55

Meke--

A very interesting, original, and insightful post. But don't worry about Kramnik's title.

The rules, valid since Steinitz's 1886 title, state that a world chess champion's title expires:

--If a champion's rating goes down by an unspecified amount,

--For an unspecified period of time.

But these rules also state that if the champion's rating goes back up to a certain unspecified level for an unspecified time his title retroactively "unexpires".

Thus, if Kramnik's excellent Olympiad is followed by a strong Dortmund and a victory against Topalov, he will have a clear championship title dating back to 2000.

If he doesn't keep playing well, however, we'll have to go back to whatever date Kramnik's title expired and fill in the champions from that date forward; you know, the way they've been doing it for 120 years.
Posted by: greg koster at June 12, 2006 13:27

All hail the 120 year tradition of 12-game matches and champions having to go through an eight-man round-robin to retain the title.
Posted by: Yuriy Kleyner at June 12, 2006 16:14

Yuriy--

It was a shame we ever departed from the 24-game standard. Those were the days. But there is at least a scrap of precedent for a shorter match:

Lasker-Schlecter 1910...ten games
Lasker-Janowski 1910...eleven games

I agree with you, Mig and many others that it's silly for candidates matches to lead into a round-robin WCC tournament. But....Kirsan abandoned his experiments with the knockouts and with the semi-rapid WCC time-control; so there's hope he'll abandon the match-into-tournament experiment as well.
Posted by: greg koster at June 12, 2006 21:56

>>ikalel, what title are you talking about? The was from 2000 has expired long ago...

Really? I think you've been visiting this site too much. Here, for some unknown reason, most people choose to ignore this: http://www.worldchesschampionship.com/. Be honest, did you forget about the 2004 WCh match or just choose not to pay any attention to it when you made your post?

Now that we're talking about this, I think Kasparov's statements, about Kramnik not being World Champion anymore because of his poor tournament results, are off the mark. Wch matches are a totally different from tournaments, of course. Obviously, Kasparov is still steaming about Kramnik's decision not to give him a rematch and he's going to use Kramnik's tournament results to diminish his title. But we all know that, apart from being ill, Kramnik was experimenting with different opening repertoires, which he has the right to do. Besides, at least Kramnik was playing and not taking year long breaks from chess, like did Lasker (and Botvinnik?).
Posted by: ikalel at June 12, 2006 21:59

Greg, very insightful, original and thoughtful posts from you as usual. Thanks for the info. I didnt know that titles could expire and "unexpire" as such. I thought they were for life. Especially if you manifestly were not qualified to play for the WC in the first place, and then steadily declined over the next few years.
Posted by: d at June 13, 2006 06:21

Since when did it matter if you qualified to play for the world championship or not? In what way had, say, Alekhine qualified to play Capablanca?If the champion chooses to put his title on the line against you, and you win, you're the champion: end of. When someone beats you, he's the champion. We had all this crap about poor tournament results and world number sevens and so on back in Petrosian's day; people bleating about how there should be a tournament champion, and so on.

You can win as many blitz playoffs as you like and as many tournaments against the likes of Kasimdzhanov and Judit Polgar as you like, but it doesn't make you the champion. Only beating the champion does that.

I mean, for God's sake, Judit Polgar was world number nine or something according to ratings. Does anyone seriously imagine she'd have any chance in a match with Kramnik?
Posted by: rdh at June 13, 2006 07:06

As long as you provide a reasonable opportunity to challengers and don't get defeated in a long match, you keep the title for life.

Thus my WCC chronology is: Fischer1972, Karpov1975, Kasparov1985, Kramnik2000-present. (And if Kramnik-Topalov is the last long match WCC, then the winner will, in my book, be the last Classical Champion.) What's your WCC chronology, d?
Posted by: greg koster at June 13, 2006 07:21

"As long as you provide a reasonable opportunity to challengers and don't get defeated in a long match, you keep the title for life."
Ah.... now I agree with you. Kramnik is the WC for life. As for my "chronology", I have a very subjective method of selection (unlike you), so it will be of no interest to you.
Posted by: d at June 13, 2006 09:15

That is not a scrap of precedent, that is evidence for the fact that the 120-year tradition is nowhere near as consistent and honorable as might appear. There is also a question of why somebody who claims to be a strict adherent to the historical tradition not get upset over such a drastic change in format as a biennial round-robin.
Posted by: Yuriy Kleyner at June 13, 2006 09:41

I doubt anybody would bet on Polgar against Kramnik today. I am pretty sure nobody bet on Euwe against Alekhine 70 years ago either. And that's why you can talk and predict and wager all you want. And then when the day is done, you play the match. And it might turn out that the person you thought had no chance wins. A GM not known for great results suddenly puts forth the best effort of his life and wins.
Posted by: Yuriy Kleyner at June 13, 2006 09:46

This reminds me of the recent discussion on another forum about Federer (in tennis), despite being ranked #1, not considered the better player over Nadal because of recent tournament results...

Here's an interesting situation that may be all too likely to pop up under KI. What if the Candidates matches were cancelled?

What are the qualification standards for the biennial FIDE WC tourney if no candidates are produced? Rating invites again?
Posted by: cynical at June 13, 2006 12:17

d--
On the contrary..I'm quite interested in your WCC chronology. Is it fair to lay in the weeds attacking Kramnik's legitimacy without letting us know the players who you think has been legitimately holding the title for the past decade or two?

Yuriy--
FIDE having locked up all the top players, Kramnik had no choice but to play the Topalov match and hope that Kirsan changes his mind at some point and brings back the long match WCC. Having signed on to the FIDE program, it wouldn't be appropriate for Kramnik to publicly bash it.

Thanks for the Alekhine-Euwe reminder...
Posted by: greg koster at June 13, 2006 12:43

ikalel, I don't think that even Kramnik would come up with better excuses! Man, have you heard of the psychological term Einstellung Effect? Ok, it is translated as (mental) set effect... I advise you to have a look at the research there and then come back and reread your post. Fits perfectly! lol
Posted by: Meke at June 13, 2006 13:55

Yeah, Kramink is just like Federer... won everything but the French Open.
Posted by: Meke at June 13, 2006 13:57

I wasn't talking about Kramnik, but since you bring him up--

The very least he could do is not to give Kirstan's system a boost a month prior to election. He could have used the leverage of world title to pressure Kirstan to promise or endorse championship matches for the future, threatening to support Kok if he didn't. Instead, Kramnik, according to his interview to Sport-Express, took a total hands off approach to negotiation and told his team to accept any terms offered.

This is without even asking why all of the world GMs suddenly would agree to Kirstan's FIDE-exclusive contracts. Large part of the reason is that Kramnik provided and proposed little in the means of alternative world title cycle organization.
Posted by: Yuriy Kleyner at June 13, 2006 15:42

Yuri
I too wish Kramnik refused to go under FIDE. I don't think it would have changed the election by one vote though.

Moreover I don't think he would have been able to organize anythign good since so many top players already signed to play exclusively in FIDE WC cycles. That was the kick in the teeth other players gave Kramnik when he seemed to be unsuccessfully tryign to get a competing cycle going. I remember an interview with Lautier saying he did not have success getting sponsorship.(this woudl seem to imply he tried) I don't know if the oher top players gave him any support on this. Obviously if Lautier and Kramnik don't have the other top players in thier camp (or worse the other top players agree not to play in thier cycle) they will have a tough time.

What happened in the back rooms during this time I don't know. But Kramnik and Lautier did clearly indicate they would try to go it alone without FIDE. Then everyone turned thier back on them and signed up for SL. Even Leko and Svidler. I don't get it. I think Svidler and Kramnik are on good terms and maybe even friends. So I don't understand.

Fact is I really have no idea what happened. I just try to put pieces together based on SWAGs.
Posted by: Niceforkinmove at June 13, 2006 16:07

Kramnik didn't have any leverage. He got the match because of pressure from his Excellency , Mr. Zhukov. If you give both the players the option of getting out of the match, it is Topalov who will opt out because he doesn't need the match. If Kramnik doesn't get a match his title would have as much value as Fischer's or Susan Polgar's.

Posted by: peach at June 13, 2006 16:33

peach,are you really serious?
Posted by: ryan at June 13, 2006 17:26

Yes, I am really serious.
Posted by: peach at June 13, 2006 17:33

Chesstraveler: Kramnik is addicted to painkillers? Where did you get this info?
Posted by: RR at June 13, 2006 20:19

I'm not lying in the weeds Greg, I'm sunning myself on the beach and looking at the weeds and their inhabitants from outside.
Posted by: d at June 14, 2006 06:29

"As long as you provide a reasonable opportunity to challengers and don't get defeated in a long match, you keep the title for life."

Kasparov, #1, and not retired, proposed a rematch. Kramnik blatantly refused.

Current #2, Anand, didn't have an opportunity either (and do not get me started with "he refused blah blah" - of course not).

Case closed.
Posted by: zarghev at June 14, 2006 06:50

Fairly dim post. If Gazza wanted a rematch, he could have made it contractual in 2000. He didn't. Anyway, Kramnik signed up to Prague. All Gazza had to do was beat Ponomariov and he'd have had his rematch. It wasn't Kramnik's fault that didn't happen. And - my memory's failing and I could be wrong, but didn't Anand have every chance to play the Dortmund event which Leko won to play Kramnik?

It's a mystery to me why Kramnik is quite so unpopular with the masses, especially the US and Eastern European masses. I suppose the former can't cope with draws (like football) whereas the latter all love the Mickey Mouse world championship events as it gives them a payday.

Posted by: rdh at June 14, 2006 08:20

Zarg--

Kasparov provided an arguably reasonable opportunity to Shirov, then Anand.

Kasparov having been defeated in 2000, the logical "handpicked" opponent for Kramnik would have been the other gentleman rated higher than himself: Anand. Dortmund 2002 was a reasonable opportunity for both Kasparov and Anand to compete for the title.
Posted by: greg koster at June 14, 2006 09:14

"Kasparov, #1, and not retired, proposed a rematch. Kramnik blatantly refused."

zarghev, you do know that Kasparov never came to Kramnik with any sort of financial offer for a second match with Kramnik? All he did was fire off statements in his interviews to the likeness of "Kramnik should give me a rematch." If Kasparov really wanted to play Kramnik again, he should have rounded up the money himself IMO. In any case it's a bit too much to say that he "blatantly refused" an offer, because Kasparov never really made one!

"Current #2, Anand, didn't have an opportunity either (and do not get me started with "he refused blah blah" - of course not)."

Sorry, but Anand has been given several chances to play for the world championship since 2000, and the fact is that he did refuse them. First there was the London 2000 match--if you remember correctly he was the original recipient of the offer, but declined it. Then there was Dortmund 2002 and Libya 2004, then a part of the Prague Agreement; Anand rejected both. Anand is my favorite of the current top players but he's had plenty enough "reasonable opportunities" to compete for Kramnik's world championship, as has Kasparov. Your title should not go defunct just because your rivals don't want to play chess to qualify for it.
Posted by: JWS at June 14, 2006 09:15

greg sez: "Kasparov having been defeated in 2000, the logical "handpicked" opponent for Kramnik would have been the other gentleman rated higher than himself: Anand."

ROFL.. What is the logic pray?

Kramnik lost a chance to popularise chess, and make himself into a legend by giving a rematch that the entire chess world (except Kramnik, Greg and his love cohorts) was slavering for. the world loves a grudge match. And people wonder why Kramnik's unpopular?
Posted by: d at June 14, 2006 10:03

d--

If, as you appear to argue, a census of slaverers should determine the WCC contender, then the billion-plus Indians slavering for an Anand match likely outweigh the folks slavering for a Kasparov rematch.

Slavering aside, why would it have been logical to handpick Anand rather than Kasparov for a match in, say, 2002? Because a Kramnik-Anand match would have been the best road to recognition of the world's best long-match player. Thus:

Kramnik-Anand2002
If Kramnik won, he'd have beaten both Kasparov and Anand in long matches. If Anand won, he'd have beaten Kramnik, who had beaten Kasparov.

Kramnik-Kasparov2002
A victory by either man wouldn't have done much in the way of proving long-match superiority over Anand.

Handpicking Anand or Kasparov in 2002 would have disappointed massive numbers of chessfans. So Kramnik chose the best solution: providing, in 2002, a reasonable opportunity to Kasparov, Anand, and the other top players to determine, not on the rating list, not in a popularity poll, but on the chessboard, the most deserving challenger.

Dortmund wasn't a perfect qualifier. But it was a reasonable opportunity for the world's best players to challenge for the title.
Posted by: greg koster at June 14, 2006 10:25

Niceforkinmove,

You think it's a coincidence Kirstan announced the match when he did? You think organizing a reunification of the title and setting up a world championship match would have been no boon to Kirstan or Kok? Ok. I disagree then. I offer the emphasis placed by Kirstan and Kok in their campaigns on past successful organization of championships and the fact that this is the issue that gets most people in chess community excited.

I have not seen much in terms of coherent ideas communicated by Kramnik or Lautier for a chess cycle of any sort (and merely saying "well, that would be swell" doesn't count). Sponsorship in such a situation does not come first, it comes second. For example, FIDE put forth the proposal for preliminary matches for round robin and now is seeking sponsors for that. They have also opened bids for things like World Cup and the final cycle itself, an approach similar to one previously used by Kasparov.

The other GMs became exclusive to FIDE when they signed on for San Luis in the summer of 2005. That gives Kramnik over 8 of post-Brissago months alone to put forth a proposal of some sort. Instead of trying to foolishly play chess with his arthritis he should have spent that time focusing on the business side of things, especially seeing as he just re-proved his credentials as champion.

Not being aware of your opponent's contract (Topalov not negotiate without FIDE's permission,) telling your team to accept any offers and saying that you don't really care about the outcome of negotiations is one sure fire way to end up with nothing you want. And this is without bringing Prague into it. Which, had Kramnik agreed to play the winner of Kasparov-Kazimdzhanov, would have obliged FIDE to create a candidates cycle for him.
Posted by: Yuriy Kleyner at June 14, 2006 11:31

i'm a sucker for punishment..

- If, as you appear to argue, a census of slaverers should determine the WCC contender, then the billion-plus Indians slavering for an Anand match likely outweigh the folks slavering for a Kasparov rematch.

No, actually its quite unlikely. The sporting aspect of an Anand-Kasparov match was less intriguing.

- Slavering aside, why would it have been logical to handpick Anand rather than Kasparov for a match in, say, 2002? Because a Kramnik-Anand match would have been the best road to recognition of the world's best long-match player.

Your logic is laughable. How do you know that handpicking Anand or Kasparov in 2002 would have disappointed massive amounts of fans? Did you conduct a poll? My impression is exactly the opposite. And no, I didnt conduct a poll either. Just my impression from message boards, statements from GMs and other relevant people etc.

Regardless of the parity of it, your statement that Kramnik chose Dortmund out of lofty ideals to maintain the integrity of the WCC is even more ludicrous. Why does your cynicism so evident and mean when talking about anybody other than Kramnik competely reverse itself when talking about your blue-eyed boy? Kramnik did what was best for numero uno, self. Pity he didnt see the larger picture.

- Dortmund wasn't a perfect qualifier. But it was a reasonable opportunity for the world's best players to challenge for the title.

OK. Greg has spoken, so it must be true.
Posted by: d at June 14, 2006 11:53

"...it was a reasonable opportunity for the world's beat to challenge for the title."

...and both Kasparov and Anand refused to participate. Moreover, the suggestion that a "grudge match" could have made Kramnik into a "legend" is hilarious. An acceptably grudgy match actually took place, and Kramnik defeated the greatest player ever without a loss. That's legendary chess. Each player who has crossed Rubicon and climbed the mountain to wear the wreath is a chess legend, and Kramnik is in the clear line from Steinitz. But there's no haven from naysaying ponces in any case: if Kramnik had granted and won a rematch there would be posters calling him out for refusing to grant a second rematch, all their fatuous trilling encouraged and enabled by this blog.
Posted by: Clubfoot at June 14, 2006 17:07

Much the way yours is.
Posted by: Mig at June 14, 2006 17:09

Riveting comeback, Migalo, but you forget the truth: I'm one of a number on this blog whose posts you discourage, vilify and often delete - that is, pro-Kramnik.
Posted by: Clubfoot at June 14, 2006 18:09

For someone who has never posted content, not even by accident, that's quite a high opinion of yourself. Bragging about what a troll ass you are is quite the topper.
Posted by: Mig at June 14, 2006 18:15

ponce: pimp

d--
You don't think a great player from the world's second most populous nation has fans? Do these fans count even if they aren't GMs and don't have internet access?

In providing the world's top players a reasonable shot at the championship were Kramnik's inner motives altruistic or selfish? I don't know enough about Kramnik to judge. But whatever his motives, he did right.
Posted by: greg koster at June 14, 2006 19:01

Every post is content, Migalo, even the occasional hack job from Net solicitors. And if calling oneself "pro-Kramnik" is expressing a high self-opinion, then rock and roll!! Thanks for acknowledging that Kramnik fans see past your unsubtle Kaspy-kissing sour grapestorm of the past five years.

What a drag - perhaps you've been sleeping with the fishes too long, because there's no fight left in you. It took just one post before you broke out the name-calling. -sigh-
Posted by: Clubfoot at June 15, 2006 01:05

greg,

"You don't think a great player from the world's second most populous nation has fans? Do these fans count even if they aren't GMs and don't have internet access?"

Haha.. The thing is, I dont underestimate Indian fans, sharing some genetic makeup with them. Yes, Anand has a multitudinous fan base, but why do you think they support Anand just because he's Indian? They support Anand because he's a brilliant Chess player, while being Indian. Doesnt mean they dont support anybody else. You can support Anand and be enthusiastic about other players. It's this condescending, know-all, sly and insinuating way you have about you that gets my goat and makes me post, even though I know there's no light at the end of the tunnel. Feel free to have the last word.
Posted by: d at June 15, 2006 04:06

Hahaha, you are a piece of work Clubfoot! I can't say I am especialyl pro Kasparov, if at all, but the guy was working for that rematch very hard and had the results to back it up. I am not against Kramnik either but what was he doing at the same time - playing 5 games per year? And then reading his interviews you just realise in what kind of 'reality' that guys live... (well, a similar conclusion could be drawn for Kasparov)
Posted by: Meke at June 15, 2006 05:57

d,

You share some genetic makeup with Indian fans? How human of you. I share 99.99% or more of my genetic makeup with Anand and with Ruslan.
Posted by: gg at June 15, 2006 06:40

Don't we all share 90-some percent of our genetic makeup with chimps?
Posted by: greg koster at June 15, 2006 07:36

Handpicking Anand or Kasparov for a 2002 WC match with Kramnik wouldn't have disappointed me. Then we would have gotten a match between Anand and Kramnik or Kasparov and Kramnik. What would have annoyed me would have been a tourney-style qualifier that threw in some vaguely-top-10 scrub and cheated everybody out of a match between the real top players...

Oh, wait...

OK, So Leko isn't a scrub, is more-than-vaguely top 10, and moreover was playing the best chess of his life in 2003-4 (which is better chess than several "World Champions" I could name have ever or will ever play). That said, I bet you could have acquired a better sponsor than Dannemann for Anand-Kramnik, and certainly for Kasparov-Kramnik II. Not that I have anything against Dannemann, but I'd never heard of them before 2004, and the match wasn't heavily promoted.

Let me be the first to say that I don't give a damn about fairness or qualifiers, I'm perfectly comfortable with handpicked (read: Sponsor-Friendly) challengers, as long as they are the -right- handpicked challengers.

For instance, if you think of Kramnik as undisputed champion, then he handpicked Topalov, which is eminently reasonable given that VT has been burning the rest of the 2700 club alive for two years now and is #1 in the world. Likewise, if you think VT is undisputed champion, VK's pedigree is also very high.

Radjabov? Not so much. He's not quite on the same level. Still, it's an "extra" match, in addition to the overall cycle - and I have no problems with the Champ agreeing to play -more- title defenses if people come up with the money on their own...
Posted by: gmc at June 15, 2006 09:31

Kasparov "working for that rematch very hard"? He didn't even bother to play in the qualifier to get it.
Posted by: acirce at June 15, 2006 10:58

gmc--

The Dortmund2002 format had several problems: homeboy Lutz, and the shortness of the group tournaments and the final matches. But if San Luis2005 had used the Dortmund2002 format we might have had some nice four-game matches: Topalov-Morozevich, Anand-Svidler, Topalov-Anand.

Following London2000 I'd have looked forward to Kramnik-Kasparov or Kramnik-Anand (in that order) more than Kramnik-Leko.

But we'll have to agree to disagree as to whether the identity of the challenger or the fairness of the process is more important.
Posted by: greg koster at June 15, 2006 12:15

It depends on whether the idea of fairness is "giving the opportunity to a large number of challengers" or "choosing the strongest challenger (or, if you prefer, having a system that results in selection of strongest challneger") for the title". Dortmund gave several GMs a chance to win a match with Kramnik. The question is whether the system in place was likely to result in the best GM winning.
Posted by: Yuriy Kleyner at June 15, 2006 17:49

"The question is whether the system in place was likely to result in the best GM winning." More so than the San Luis format, for what that's worth.
Posted by: acirce at June 15, 2006 18:22

Agreed, Yuriy,

--In the "Group 2" tournament-of-four, Adams drew five games, lost one to Leko, and went home. If Adams wins that game, Leko, the ultimate winner, would have gone home. In a longer format who knows, Adams might have proven to be strongest of them all.

The alternative to Dortmund was handpicking.
But if you had handpicked Kasparov or Anand, you'd have had to slam the door on the other man--which doesn't seem right either.

With FIDE in opposition and the world's #1 badmouthing them, however, we should probably be satisfied that they were able to pull off any halfway-decent qualifier and WCC.

Posted by: greg koster at June 15, 2006 18:57

This whole thing is so deja vu. It was decided in the sixties - the best player isn't the one with the highest rating; it's the one the others can't beat in a match.

I'd be interested to know which world champions didn't play chess as well as Leko was in 2003-4! Unless we're counting the likes of Kasimdhzanov, of course (who actually earned my considerable respect by refusing to call himself World Champion. I wonder to what extent, if at all, that admirable self-restraint contributed to ending the MM events?).
Posted by: rdh at June 15, 2006 19:33

Just because it was decided in the 60's doesn't mean it is valid today.
Posted by: peach at June 15, 2006 22:20

Corus 2001: Kasparov finishes first, Anand second
Linares 2001: Kasparov finishes first, Anand doesn't play
Dortmund 2001: Kramnik wins, Anand finishes last, Kasparov doesn't play
Astana 2001: Kasparov wins, Anand doesn't play
Corus 2002: Bareev wins (Kasparov, Anand and Kramnik don't play)
Linares 2002: Kasparov wins, Anand finishes fourth

On the basis of results in 2001 or early 2002, it would have been ridiculous to hand-pick anybody but Kasparov.
Posted by: Yuriy Kleyner at June 16, 2006 09:36

You can win a tournament, but not a match, by drawing with the top dogs and smashing the bottom-enders.

Kasparov's dynamic style works well in tournaments but he had a tough go in the Karpov matches, and he looked like a fish out of water against Kramnik and IBM.

I'd like to see Jeff Sonas develop a match-predicting formula.

What if, for Kramnik, Kasparov, and Anand, you
--threw out all the draws
--counted every loss
--counted wins against only the top three

By this method, Kasparov2001-2002 would have a much stronger "handpicked-challenger" claim than Anand2001-2002 based on:
--Kasparov beat Kramnik at Astana2001.
--Anand lost numerous classical games over that period; Kasparov lost none. (I think)
Posted by: greg koster at June 16, 2006 12:23

Greg,

You said that it would be unfair not to hand-pick Anand. I assumed this was based on his results from the era.

There is a tendency in any competition to do better against the weaker players than against stronger players. If Kasparov did in fact win only against the bottom-feeders during this era you are looking at different tournament brackets than I am.

Against Karpov, Kasparov played during the second-greatest player of all time. And still managed to never lose a match to him, winning three times and drawing once. Want to hold it against Garry that the final score was never a blow-out? In subsequent years, Kasparov faced two other men, who won match tournaments that included all the strongest players in the world at the time. He won there as well, handily. I agree that Kasparov did not look comfortable against Kramnik. His performance was off, nowhere near as good as it has been against Kramnik elsewhere (4-3=27). As for playing against the computers, show me a GM who hasn't been out of water in matches with them since 1997. Heck, show me who else beat Deep Blue. Kasparov's loss in a match came down to one game, and under conditions which favored the machine (no rest days, short time control, in-match modifications).

I like your formula, but with one modification. Why limit yourself to those three? Generally speaking, I evaluate non-losses against 2700s only. It eliminates the Vallejo Ponses in this world while showing GMs overall record against people like Topalov, Morozevich and Leko. I would actually love to compire something like that, if people are interested.
Posted by: Yuriy Kleyner at June 16, 2006 13:10

Yuriy, I admire your willingness to back up your opinion with hard facts but I am not sure it helps when it comes to those people who have their strong anti-Kasparov opinions. Except that they usually suffer from a selective memory loss, they can always come up with ingenious reasons why the 'facts' are not valid or relevant. It is just amazing... I wish I could get them to do some experiments... I pay well!
Posted by: Meke at June 16, 2006 15:05

Yuriy--

If we're wondering whether "Grandmaster A" can beat Kramnik, it matters little that "Grandmaster A" can beat players less strong than Kramnik.

But if "Grandmaster A" lost to players not as strong as Kramnik, it's a good bet he could also lose to Kramnik.

Thus for a 2002 match-rating tool, count victories over anyone as good or "better" than Kramnik. And count losses to anyone.

Kasparov's success in his four Karpov matches was a legendary achievement, no argument there. I'd just point out that Kasparov's victory margins weren't as wide as the elo ratings would have predicted. And that a match-rating tool might be nice to have in the toolkit.

Match Year KK elo-difference score(est) actual
1985 -20 11.3 13
1986 +35 13.2 12.5
1987 +40 13.3 12
1990 +70 14.4 12.5

I'm suggesting that in a match, a Karpov/Kramnik-style player will generally out-perform his rating; a Kasparov/Topalov-style player will generally underperform.
Posted by: greg koster at June 16, 2006 15:21

Match year....KKelo-difference...est score...score
1985................-20............13.3......13
1986................+35............13.2......12.5
1987................+40............13.3......12
1990................+70............14.4......12.5
Posted by: greg koster at June 16, 2006 15:25

Dang...

1985..........-20..........11.3........13
Posted by: greg koster at June 16, 2006 15:34

These ratings suggest one significant underperformance, one significant overperformance, and two matches that are roughly as much of a draw as could be expected from the ratings. They also more significantly suggest (and this is true of most ratings systems I know) that elo rating is heavily based on results of previous matches and that Kasparov continued to beat Karpov in matches his rating went up relative to Anatoliy Yevgenyevich's even though the relative skills probably did not change that much.

That's without even talking about how much Kasparov would want to go for an extra win against a strong defensive player or against a strong player, period.

"If we're wondering whether "Grandmaster A" can beat Kramnik, it matters little that "Grandmaster A" can beat players less strong than Kramnik. "
We are trying to estimate the relative strength of GMs A through F. Barring a scenario in which one of the GMs is consistently unable to score a victory against Kramnik, that would be the best indication of one's strength as a chess player.
Grandmasters A through F are probably capable of beating Kramnik. Do you mean who would triumph in a match? If so, barring a scenario in which GM A seems to be unable to beat Kramnik in any game they ever played (such as Shirov-Kasparov), it is a matter of relative strength as a chess player. The top 3 GMs we are talking about only play each other like four times a year. Base the rating on that and you are basing it on what is most likely one decisive game. Additionally, yes, how well you play against a GM whose skill is close (though not equal to that of yours) is a significant indicator of your skill as a player. That's why in other sports, such as NBA and NFL, while we do give a heavier weight to how well you do against other playoff teams, overall results is what matters when you decide who makes the playoffs.
Posted by: Yuriy Kleyner at June 16, 2006 21:29

Meke,

I know. I am doing this for the casual fans who might otherwise think the facts back up some of the ridiculousness which is occasionally spouted on here.
Posted by: Yuriy Kleyner at June 16, 2006 21:30

Yuriy--

I'll leave it to better mathematical minds than my own to determine whether Kasparov underperformed his rating in the Karpov/Kramnik matches.

Chance plays a much, much higher role in team games, NBA and NFL results, than it does in chess.

A better analogy might be the individual, low-chance contests of weight-lifting, sprinting, pro boxing, or college wrestling, where the better player might beat all comers dozens, or even a hundred times in a row. In such sports, and in chess, beating the tar out of the #3, #5, or #6 guy doesn't mean you can beat Vasily Alexeyev, Edwin Moses, Mike Tyson, Dan Gable, or Kasparov at their best.

A casual fan, I always appreciate your efforts in combatting spouted ridiculousness.
Posted by: greg koster at June 16, 2006 23:34

If you want to see (from a Chessmetrics perspective) whether players underperformed or overperformed their rating in various top matches, you can go to http://db.chessmetrics.com/CM2/Summary.asp?Params=1840AASSS4S3S000000000000111000000000000010100#StrongestMatchAnchor and then click on the match you're interested in. Basically it says that in the five Kasparov-Karpov matches, Karpov slightly outperformed his rating in matches IV and V, Kasparov slightly outperformed his rating in match II, Karpov way outperformed his rating in match I, and match III was exactly what would be expected.
Posted by: Jeff Sonas at June 17, 2006 00:21

Thanks, Jeff--

So--
Match I....Karpov outperforms rating by 2.9 games
Match IV...Karpov outperforms rating by .9 games
Match V....Karpov outperforms rating by .9 games
Match II...Kasparov outperforms by .8 games
Match III...exactly as expected

Question on Kasparov-Kramnik:
Kramnik won with +2. He outperformed his Chessmetrics rating by +2.2. Does that mean a drawn match was predicted? Even though Kramnik was a 69-point rating underdog?
Posted by: greg koster at June 17, 2006 00:53

No, it's a half-point vs. full-point thing. In 15 games Kramnik would be expected to score 6/15 or 6.5/15, and he actually scored 8.5/15, so he scored 2 or 2.5 match points more than expected. Like let's say they had been predicted to be exactly even, and Kramnik had won with +2. That'd be only 1 full point more than expected (8.5 instead of 7.5), and so in that case we would say he outperformed his Chessmetrics rating by 1.0. It gets confusing because of all the +2's, but Kasparov was predicted to win with +2, and Kramnik actually scored +2, which was 2 full points better than expected for him.
Posted by: Jeff Sonas at June 17, 2006 01:02

Jeff--OK, I see.

Would it be fair to say that "undynamic" players: (e.g. Petrosian, Karpov, Kramnik) generally outperform their Chessmetrics ratings more often than "dynamic" players do: (e.g. Larsen, Fischer, Kasparov)?

I'm thinking the dynamic group would be more likely to demolish lower-rated opposition (boosting their Chessmetrics rating) while the undynamic group would get more draws against lower-rated folks (hurting their Cm rating).
Posted by: greg koster at June 17, 2006 01:22

Jeff, Sorry--

I mean, when playing in matches: do the "undynamic" players tend to out-perform their Cm ratings more often than the "dynamic" players do?
Posted by: greg koster at June 17, 2006 01:27

"I'll leave it to better mathematical minds than my own to determine whether Kasparov underperformed his rating in the Karpov/Kramnik matches."

Okay. You had no trouble trying to do it earlier. I will let the chess fans decide whether a 30 point elo or chessmetrics gap really can really be used to predict to within one game the logical outcome of a chess match. I will also let them make their own decisions as to whether such a difference in rating is an indication of actual difference in strength or a matter of who played who and where over the last couple of months, as well as size of margin and choice of aggressiveness in the final games of the match.

"Chance plays a much, much higher role in team games, NBA and NFL results, than it does in chess. "

I am not sure we could discuss this on any sort of objective basis . . . especially since chance takes a very different form in chess from physical sports and I would agree that it plays a greater role in physical sports.

"A better analogy might be the individual, low-chance contests of weight-lifting, sprinting, pro boxing, or college wrestling, where the better player might beat all comers dozens, or even a hundred times in a row. In such sports, and in chess, beating the tar out of the #3, #5, or #6 guy doesn't mean you can beat Vasily Alexeyev, Edwin Moses, Mike Tyson, Dan Gable, or Kasparov at their best."

The first two competitions you mention are not even head to head, but rather against the standard, (weight, time) while the others have no concept of a draw, which seriously wrecks the idea of under-over performance due to style.
In any of these competitions, when a person triumphs over top contenders with ease it is generally taken as a sign of him having greater strength and being more able to win than somebody who has not had as much ease winning against them.
Posted by: Yuriy Kleyner at June 17, 2006 12:53

Yuriy--

Jeff noted my error...

In the five matches, Karpov actually out-performed ratings expectations by about 6 games, 2 games, 2 games, zip, and -2 games. Kramnik outperformed expectations by 4 games. Thus Kasparov had 12 unpredicted losses over those six matches. Looks significant to me... but maybe over so many games it's not. I'll defer to the experts.
Posted by: greg koster at June 17, 2006 20:54

Greg,

The first match is obviously a significant underperformance by Kasparov. I think we can agree that Garry froze from inexperience in that one and wasn't able to get his game going till much later in the match and till Karpov's endurance started to be a factor. I also question whether you could really have called Kasparov a two game favorite going into this match. (3.8 over 48 games=1.9 over 24) Maybe the candidates tournament tends to prop one's score signifcantly up?

It looks like one match is an exception (the first one) and the rest amount to 1 (.9+.9-.8+0=1) point (or 2 game) overperformance by Karpov over the course of the 4 matches. This is assuming that the two GMs relative strength really did waver to the specific points postulated by chessmetrics--my guess is that when you are talking about 2800 gaining 20 points this is more likely to be due to how often you play rather than any significant change in skill level.

The Kramnik-Kasparov match is a completely different boat. We were talking about whether Garry underperformed against Karpov. I have said above that I agree that Kasparov's performance in the Kramnik match was subpar and it is exceptionally different from expected by any rating measure you choose (including the one you suggested above of wins against players of similar strength and lack of losses against weaker players). What the explanation is we don't know. But it does not support the theory of "positional" over "dynamic" in matches over any other explanation one can propose for the result. Especially when the "positional" Kramnik underperforms even more severely rating-wise to the "dynamic" Shirov two years earlier.
Posted by: Yuriy Kleyner at June 17, 2006 23:46

Greg,

The first match is obviously a significant underperformance by Kasparov. I think we can agree that Garry froze from inexperience in that one and wasn't able to get his game going till much later in the match and till Karpov's endurance started to be a factor. I also question whether you could really have called Kasparov a two game favorite going into this match. (3.8 over 48 games=1.9 over 24) Maybe the candidates tournament tends to prop one's score signifcantly up?

It looks like one match is an exception (the first one) and the rest amount to 1 (.9+.9-.8+0=1) point (or 2 game) overperformance by Karpov over the course of the 4 matches. This is assuming that the two GMs relative strength really did waver to the specific points postulated by chessmetrics--my guess is that when you are talking about 2800 gaining 20 points this is more likely to be due to how often you play rather than any significant change in skill level.

The Kramnik-Kasparov match is a completely different boat. We were talking about whether Garry underperformed against Karpov. I have said above that I agree that Kasparov's performance in the Kramnik match was subpar and it is exceptionally different from expected by any rating measure you choose (including the one you suggested above of wins against players of similar strength and lack of losses against weaker players). What the explanation is we don't know. But it does not support the theory of "positional" over "dynamic" in matches over any other explanation one can propose for the result. Especially when the "positional" Kramnik underperforms even more severely rating-wise to the "dynamic" Shirov two years earlier.
Posted by: Yuriy Kleyner at June 17, 2006 23:47

The general argument, I believe, is the overall (un)reliability of TOURNAMENT results in predicting the result of a top-level MATCH between an exceptionally dynamic player and an exceptionally non-dynamic player.

Where Kasparov was an exceptionally dynamic top-level player, and Kramnik and Karpov were/are, exceptionally non-dynamic top-level players, the results of the six Kasparov-Karpov/Kramnik matches are an interesting test of this theory.

Shirov-Kramnik (9 games) would count against the theory, but the (159 games) in the six Kasparov-Karpov/Kramnik matches tend to affirm it, although, perhaps, too faintly to matter.

It's unlikely that any system would have predicted Kramnik's two-point victory over Kasparov. But a system giving less positive weight to beating lower-ranked players and less negative weight to drawing them wouldn't have made Kramnik a two-point dog. I suspect it would have rated his chances closer to 50-50.

But these ideas re tournament versus match performance are far out-running my ability to mathematically prove them. So it's probably (past) time for me to give this subject a rest.
Posted by: greg koster at June 18, 2006 02:53

Okay. I will also probably make this my final post.

With exception of Kasparov's first nine games with Karpov and his last match against Kramnik the numbers affirm Kasparov underperforming by about 1 point over the course of his matches. This is over 135 games, and without considering how much of each man's ratings was due to their match victories. So you can take that as your evidence. Or you can consider Kasparov's two worst stretches ever as an indication of his real strength as a chess player.

You can also click on Sonas's link above and take a look at matches between positional and non-positional players from the past. The problem is there are relatively few such events (Karp-Kasp, Petro-Spass, Shirov-Kram, but how dynamic is Korchnoi? do I really want to take Fischer's matches in 1971 as any sort of proof? ) and the results don't form any sort of dominant pattern.

One last thing. The strongest tournament performance of all time as well as the second most impressive tournament results in history belong to Karpov. It appears that strongest players are able to win tournaments and matches, regardless of style.
Posted by: Yuriy Kleyner at June 18, 2006 12:38

As long as we're on the topic, I ran some historical numbers for 36 all-time greats about their relative performance in match, tournament, and team play. There is a lot of calculation behind these numbers, and I don't want to go into all the details, but basically this is supposed to tell you whether players did unusually well/poorly in the three types of event. For instance, these numbers suggest that Viswanathan Anand will have a performance rating 35 points above expectation in a match, or 12 points above expectation in team play, or 5 points below expectation in a tournament. The numbers balance out because he plays so many more games in tournaments than in team play or in matches.

Alexander Alekhine: +7 Elo overperformance in tournament play (523 games)
Alexander Alekhine: +3 Elo overperformance in match play (166 games)
Alexander Alekhine: -78 Elo underperformance in team play (51 games)
Viswanathan Anand: +35 Elo overperformance in match play (97 games)
Viswanathan Anand: +12 Elo overperformance in team play (41 games)
Viswanathan Anand: -5 Elo underperformance in tournament play (710 games)
Mikhail Botvinnik: +5 Elo overperformance in tournament play (421 games)
Mikhail Botvinnik: +2 Elo overperformance in match play (218 games)
Mikhail Botvinnik: -17 Elo underperformance in team play (129 games)
David Bronstein: +33 Elo overperformance in match play (51 games)
David Bronstein: -1 Elo underperformance in tournament play (1191 games)
David Bronstein: -5 Elo underperformance in team play (155 games)
José Capablanca: +13 Elo overperformance in team play (14 games)
José Capablanca: +10 Elo overperformance in tournament play (345 games)
José Capablanca: -70 Elo underperformance in match play (52 games)
Mikhail Chigorin: +51 Elo overperformance in match play (101 games)
Mikhail Chigorin: -13 Elo underperformance in tournament play (394 games)
Max Euwe: +27 Elo overperformance in match play (271 games)
Max Euwe: +20 Elo overperformance in team play (59 games)
Max Euwe: -16 Elo underperformance in tournament play (513 games)
Bobby Fischer: +15 Elo overperformance in team play (62 games)
Bobby Fischer: +1 Elo overperformance in tournament play (382 games)
Bobby Fischer: -20 Elo underperformance in match play (56 games)
Boris Gelfand: +3 Elo overperformance in tournament play (875 games)
Boris Gelfand: +2 Elo overperformance in team play (132 games)
Boris Gelfand: -58 Elo underperformance in match play (49 games)
Efim Geller: +2 Elo overperformance in tournament play (1381 games)
Efim Geller: -6 Elo underperformance in team play (235 games)
Efim Geller: -64 Elo underperformance in match play (32 games)
Vassily Ivanchuk: +2 Elo overperformance in tournament play (843 games)
Vassily Ivanchuk: -1 Elo underperformance in team play (172 games)
Vassily Ivanchuk: -46 Elo underperformance in match play (29 games)
Anatoly Karpov: +15 Elo overperformance in team play (144 games)
Anatoly Karpov: +2 Elo overperformance in match play (406 games)
Anatoly Karpov: -2 Elo underperformance in tournament play (1396 games)
Garry Kasparov: +24 Elo overperformance in team play (122 games)
Garry Kasparov: +8 Elo overperformance in tournament play (732 games)
Garry Kasparov: -33 Elo underperformance in match play (274 games)
Paul Keres: +3 Elo overperformance in tournament play (1115 games)
Paul Keres: -9 Elo underperformance in match play (64 games)
Paul Keres: -10 Elo underperformance in team play (268 games)
Viktor Korchnoi: +4 Elo overperformance in match play (360 games)
Viktor Korchnoi: =0 Elo performance in tournament play (2501 games)
Viktor Korchnoi: -4 Elo underperformance in team play (464 games)
Vladimir Kramnik: +18 Elo overperformance in team play (48 games)
Vladimir Kramnik: =0 Elo performance in tournament play (614 games)
Vladimir Kramnik: -15 Elo underperformance in match play (78 games)
Bent Larsen: +5 Elo overperformance in tournament play (1323 games)
Bent Larsen: -12 Elo underperformance in match play (115 games)
Bent Larsen: -62 Elo underperformance in team play (80 games)
Emanuel Lasker: +3 Elo overperformance in match play (178 games)
Emanuel Lasker: -3 Elo underperformance in tournament play (243 games)
Joel Lautier: +27 Elo overperformance in team play (172 games)
Joel Lautier: -3 Elo underperformance in tournament play (628 games)
Joel Lautier: -83 Elo underperformance in match play (30 games)
Frank Marshall: +78 Elo overperformance in team play (25 games)
Frank Marshall: -2 Elo underperformance in tournament play (586 games)
Frank Marshall: -9 Elo underperformance in match play (121 games)
Alexander Morozevich: +35 Elo overperformance in team play (75 games)
Alexander Morozevich: -9 Elo underperformance in tournament play (303 games)
Miguel Najdorf: +5 Elo overperformance in team play (149 games)
Miguel Najdorf: +4 Elo overperformance in tournament play (847 games)
Miguel Najdorf: -56 Elo underperformance in match play (66 games)
Aron Nimzowitsch: +4 Elo overperformance in tournament play (347 games)
Aron Nimzowitsch: -153 Elo underperformance in match play (10 games)
Tigran Petrosian: +3 Elo overperformance in team play (314 games)
Tigran Petrosian: =0 Elo performance in tournament play (1267 games)
Tigran Petrosian: -6 Elo underperformance in match play (140 games)
Harry Pillsbury: +9 Elo overperformance in tournament play (239 games)
Harry Pillsbury: -61 Elo underperformance in match play (43 games)
Judit Polgar: +23 Elo overperformance in match play (22 games)
Judit Polgar: =0 Elo performance in tournament play (518 games)
Judit Polgar: -7 Elo underperformance in team play (60 games)
Samuel Reshevsky: +4 Elo overperformance in tournament play (654 games)
Samuel Reshevsky: -7 Elo underperformance in match play (111 games)
Samuel Reshevsky: -24 Elo underperformance in team play (86 games)
Akiba Rubinstein: +28 Elo overperformance in team play (19 games)
Akiba Rubinstein: +26 Elo overperformance in match play (34 games)
Akiba Rubinstein: -3 Elo underperformance in tournament play (521 games)
Alexei Shirov: +156 Elo overperformance in match play (33 games)
Alexei Shirov: +1 Elo overperformance in team play (156 games)
Alexei Shirov: -5 Elo underperformance in tournament play (1034 games)
Nigel Short: +46 Elo overperformance in match play (122 games)
Nigel Short: -2 Elo underperformance in tournament play (1015 games)
Nigel Short: -19 Elo underperformance in team play (199 games)
Vassily Smyslov: +15 Elo overperformance in team play (237 games)
Vassily Smyslov: -1 Elo underperformance in tournament play (1512 games)
Vassily Smyslov: -15 Elo underperformance in match play (121 games)
Boris Spassky: +40 Elo overperformance in match play (243 games)
Boris Spassky: -6 Elo underperformance in team play (202 games)
Boris Spassky: -6 Elo underperformance in tournament play (1320 games)
Wilhelm Steinitz: +22 Elo overperformance in match play (204 games)
Wilhelm Steinitz: -25 Elo underperformance in tournament play (200 games)
Mikhail Tal: +4 Elo overperformance in tournament play (1680 games)
Mikhail Tal: =0 Elo performance in team play (247 games)
Mikhail Tal: -47 Elo underperformance in match play (136 games)
Siegbert Tarrasch: +4 Elo overperformance in match play (106 games)
Siegbert Tarrasch: -2 Elo underperformance in tournament play (464 games)
Johannes Zukertort: +24 Elo overperformance in match play (116 games)
Johannes Zukertort: -21 Elo underperformance in tournament play (131 games)
Posted by: Jeff Sonas at June 19, 2006 02:24

Thanks Jeff--

Surprising(?):

Capablanca... -70 underperformance in matches
Short... +46 overperformance in matches
Shirov... +156 overperformance in matches!
Posted by: greg koster at June 19, 2006 02:45

Yeah, one problem with my ratings is that Lasker's rating expired just before his match against Capablanca, so Capablanca's games were unrated in their 1921 match.
Posted by: Jeff Sonas at June 19, 2006 04:08

There goes your theory greg... reading your posts I can't help but ask about your ELO (USCF) and how well do you know the ELO scale. It seems to me that your experience with chess ratings is somewhat limited and that you do not understand what +20-30 points in reality mean.
Posted by: Meke at June 19, 2006 07:10

Some of the ratings are somewhat easier to explain than others. Especially in these days, where match play sadly is rare. One good/bad effort will count for a lot. So Shirov's one good match against Kramnik will count for a lot. Also, if a GM performs as expected for most matches and his career and then suffers blow out such as that suffered by Larsen in 1972 or the two suffered by Geller to Spassky it would impact his rating a lot.
Posted by: Yuriy Kleyner at June 19, 2006 09:37

Meke--

Since you didn't offer an argument about my "theory", I'm unable to respond.

I had a low USCF rating a decade or three ago (I don't remember what it was), which probably overstates my current level of chess "ability."

If I understand correctly (always a big "if") a 20-30 point rating edge over an opponent translates to a performance expectation of about .53 to .54 against that opponent.

I humbly acknowledge that in this blog I'm routinely outclassed in chess-strength, statistical expertise, and argumentative ability by folks such as yourself.
Posted by: greg koster at June 19, 2006 09:52

Thanks for the info greg. As for the theory, I was referring to Jeff's post. I think the numbers there were rather clear-cut... as for the ELO ratings, .53 is indeed much much closer to 1/2 than to 1, isn't it? all the best...
Posted by: Meke at June 19, 2006 11:35

A performance expectation of any value (or rather perhaps delta between two p.e.) has no meaning whatsoever without also stating the number of games over which you expect the edge to accumulate.

Kasparov would have been expected to beat Kramnik by a full 15 in 2000 . . . had they been playing a 94 game match.

Jeff, thank you so much for your work. Chessmetrics is an endless amazing source of information about chess past.
Posted by: Yuriy Kleyner at June 19, 2006 11:40

jeff, fascinating stuff. Esp Tal, who I'm convinced was a much better player than that dismal loss to Botvinnik in the 2nd match.
Posted by: d at June 19, 2006 12:23

Yuriy--

Kasparov rating edge going into the 2000 match:
2871-2802=69 points.

69 points yields a .60 expectancy in results.

.60 of a 94-game match =

56.5 points for Kasparov
43.5 points for Kramnik
=13 point margin

Thus, if London2000 had been a planned 94-game match: Kasparov would have been EXPECTED to beat Kramnik 13-0 with the rest drawn, or 14-1 the rest drawn, or 15-2 the rest drawn, etc.

If the actual London2000 results had been extrapolated out to 94 games it's Kramnik who would have had enjoyed a 12-game or 13-game margin.
Posted by: greg koster at June 19, 2006 12:47

Greg,

According to chessmetrics (http://db.chessmetrics.com/CM2/SingleEvent.asp?Params=199510SSSSS3S068986000000111102199000000010100) Kasparov was expected to get 58 (not 60, but it doesn't make a major difference) percent of the points in the match, Kramnik 42.

.58*94=54.5 points
.42*94=39.5 points

That is a 15 point margin in the match.

Your math contains a simple error:

56.5+43.5=100 points at stake. In a 94-game match, only 94 points are at stake. That is, the 56.5 you got is the number of points Kasparov would be expected to get in a 94 game match (if he was EXPECTED to get 60 percent going in). The 43.5 you got from thinking Kasparov would be expected to get 56.5 in a HUNDRED game match.
Posted by: Yuriy Kleyner at June 19, 2006 13:32

[I don't know why we're talking about a 94-game match instead of a 100-game match, but you are right.]

Projected results of a Kasparov-Kramnik 94-game London2000 match:
Kasparov--54.5 pts
Kramnik--39.5 pts
(15-0 or 16-1 or 17-2, etc., the rest drawn.)

Actual London2000 results projected out to a 94-game match:
Kramnik wins 12.5-0 or 13.5-0.5 or 14.5-1.5, etc. the rest drawn.
Posted by: greg koster at June 19, 2006 14:26

Oh geez.

Last line should be:
"Kramnik wins by a margin of about twelve to thirteen."
Posted by: greg koster at June 19, 2006 14:38

Hey, Greg

I chose the 94 game match because that's when the margin runs to 15 games ... the length of the actual match. This was to demonstrate how performance margin has meaning only as it relates to the number of games.

For a 94-game match, the expected performance margin is 15 games.
For a 15-game match, the expected performance margin is 2.4 games.

But in reality, they are the same performance expectation!
Posted by: Yuriy Kleyner at June 19, 2006 14:55

My previous answer about Anand didn't pass the Mig-filter and was discarded, hence here is another attempt, (in the view of Jeff's finding that Anand overperformed in match play), about the myth "Anand refused to play WC matches":

- Anand didn't play in Kasparov qualifier in 1998, because he had played in FIDE WC (and had signed a contract with FIDE).
- Anand accepted to play Kasparov in 1999, and hence couldn't play in FIDE WC. The match with Kasparov being cancelled in 1999, Anand lost not only all the money for the preparation, but also all money he could have made in FIDE WC 1999 (which was won by Khalifman), and in some tournaments.
- In 2000, Anand was offered a contract by a company which didn't exist yet (MSO announced the match, then Braingames), with unknown sponsor and unknown location. Hence Anand demanded that an advance payment was made, not a unreasonable request in view of the past failures (twice which Shirov, once with Anand). Instead of negociating, the deadline for the offer was advanced by one month.

Now who is claiming that Anand genuinely refused a match in 2000? What happen is that Anand was given a dubious offer, and was discarded for questionning it.

Quiz:
- Which of the companies: MSO and Braingames, successively announced to organize the WC in 2000, ended up bankrupt? (hint: more that one).
- How many of the following matches of Kasparov's cycles actually occured with no issues: Shirov-Kramnik 1998, Kasparov-Shirov 1998, Kasparov-Anand 1999, Kasparov-Kramnik 2000, Kasparov-Ponomariov 2003, Kasparov-Kasimdzhanov 2005? (hint: less than two).
- Which major title could Anand participate in, by virtue of not having signed the risky MSO/Braingames offer? (hint: it's a World Champion title by an international chess organisation, with a top prize of $300,000)
- Ha also, how many times, the #2 player of the world (yes, Anand), was consulted for the organisation of a World Championship ? -- as you'd think the opinion or agreement of the *#2* would be of some value in any World Championship - not only in chess. (hint: not in 2000, and not for Dortmund 2002).
Posted by: zarghev at June 19, 2006 15:19

The only filters are for spam, profanity, and too many links (for anti-spam purposes). They are entirely automated. If you had tried to post something and it was blocked or moderated, it would still appear in the system and I don't see anything else from you since your post in this thread on the 14th. What happened when you tried to post?
Posted by: Mig at June 19, 2006 15:34