Dortmund, aka Big Vlad's plaything, is a supertournament, it is generally agreed. Along with Corus Wijk aan Zee and Linares it is one of the majors, as a golf fan might have it. The M-Tel Masters in Bulgaria is a relatively new and welcome addition that has joined the ranks of annual super events along with the Aerosvit tournament and the Tal Memorial. Biel has long been considered second-tier, but that line of thought is looking increasingly obsolete. Last year it was a category 18 event, which is about where I'd put the cutoff for earning the "super" sobriquet these days. Actually, instead of rating numbers, which are rising like the price of gold, I'd rather go with something like "only one player outside of the top 40 and at least three members of the top ten". That prevents disqualification due to the presence of a local invitee (hello Yannick! hello Arkadij!) There should also be a minimum number of participants and/or rounds and an all-play-all format. Dortmund's seven round sprint is barely less unimpressive than Essent's minimalist four players and six rounds. Of course supertournament is an unofficial term anyway, much like "super-GM."
From the start of 2007 until today, we've seen around a dozen supertournaments. Plus the Melody Amber events, the world blitz ch, and a few other strong rapids. With the new FIDE GP coming online, we are enjoying a bumper crop of strong all-play-all tournaments. Taking only cat. 18 and higher, 2007-present, Anand has won three out of the five he played in (Mexico WCh, Linares 07 & 08). Kramnik is 2/5 (Dortmund and the Tal Memorial). So that takes care of almost half the events we're counting. Ivanchuk has two clear firsts as well (Aerosvit 07 & MTel 08). Carlsen just earned a big win at Aerosvit 08. He also took first at Biel 07, though that was on blitz tiebreaks after sharing first with Onischuk, so that's a split for our purposes. Topalov has MTel 07. The others are a little messier. Topalov, Aronian, and Radjabov shared first at Corus 07. (I'll bend the rules and give them all a half-point.) Aronian and Carlsen split Corus 08 honors. Gashimov, Wang Yue, and Carlsen divided the Baku GP.
To recap the scoring, that's Anand 3; Kramnik, Carlsen, and Ivanchuk 2; Topalov 1.5; Aronian 1; Radjabov, Wang Yue, Gashimov, and Onischuk 0.5. Honorable mention to Morozevich for his impressive score at the cat. 18 Sarajevo 08, but he was the only player in the top 30 in the field of six. Gata Kamsky has to get full props for winning the 07 World Cup, taking out big guns Shirov, Svidler, Carlsen, and Ponomariov in the process. There were also many impressive performances on teams and in rapids. Let me know if I missed anything.
Kramnik, Carlsen, and Ivanchuk will have a chance to add to their totals in the coming months, although we'll have to stretch the rules a bit for both. Sparkassen Chess-Meeting Dortmund begins on Saturday, June 28. The field: Kramnik, Leko, Mamedyarov, Ivanchuk, van Wely, Nepomniachtchi, Naiditsch, and Gustafsson. I'm all for new blood and new faces, so I don't want my invented elitist criteria to be taken too seriously. Dortmund taking in the winner of the Aeroflot Open each year is a brilliancy that other supertournaments should imitate. (Corus already does this with its internal group promotion policy.) The young Nepomniachtchi won't have it easy, but his talent is the real deal. It's interesting that Jan Gustafsson is playing instead of top-rated German and national champion Daniel Fridman. But it's Kramnik's tournament to lose as always, something he rarely does. Big Vlad hasn't played a classical event since coming in =7-8 at Corus in January. I don't think Kramnik has played anyone as low-rated as Naiditsch (2623) and Gustafsson (2603) in years, other than simuls and blitz. Oh, he did draw with Smeets (2538) in a Dutch league game last year.
The top event at the 41st Biel International Chess Festival is further along, starting July 19. Carlsen plays, which is all most people need to hear to tune in, with good reason. But by necessity there are other players as well: Bacrot, Alekseev, Dominguez, Onischuk, and Pelletier.
Excellent Enlightenment Mig, as usual. However, please let us know more about La Miglett's forthcoming welcome to earth day and how's The Bride? We are all very happy for you and your Family is in all of our prayers I am sure.
How does Aerosvit 08 + Biel 07 + Corus 08 + Baku GP = 2?
"How does Aerosvit 08 + Biel 07 + Corus 08 + Baku GP = 2?"
I would like to hear the answer to this as well. :)
Obviously Mig considers Corus and Baku shared, so 1+0,5+0,5.
still shared would be 1+1+.5+.5 = 3
Obviously I can't say that my math skills are better than Mig's.
"Actually, instead of rating numbers, which are rising like the price of gold, I'd rather go with something like "only one player outside of the top 40 and at least three members of the top ten"."
this is misleading on two levels:
1) the difference in strength between number 1 and number 40 (or 100, or x as some 2-3 digit number) is smaller today than ever before - 10 years ago it was big, 20 years ago it was very big. this follows logically from the big increase in strong, active players world-wide.
hence, a "top x players" measure is by no means any better than using rating (if the hypothesis is that the inflation is huge) - it just has the opposite effect, making it seem that being nr. 40 in (this year - y) or 2008 - 20 = 1988 was as elite as being nr. 40 in 2008. it simply wasn't.
it's the same kind of logical fallacy as claiming that being among nr. 10-50 (or top 100 if you like) was as great in 1978 as it is being among nr. 10-50 (or top 100 if you like) in 2008. because it wasn't - it takes clearly more to be top 100 today, in terms of skills and commitment, because the competition is much worse.
2) a statement like "rating numbers ... are ricing like the price of gold" (honestly i haven't checked the stock exchange for the latest prices...) seems to imply something along these lines: a) there is inflation, b) there is _heavy_ inflation, c) the rate of the inflation increase is higher than ever.
most will agree on a) even if the notion of "inflation" is somewhat intangible: ratings are supposed to be an _indirect_ measure of strength, but it's only a relative measure. the ratings aren't measuring something constant - when the price of gold goes up, we're talking about how much to pay for one ounce, typically, but where did anyone define how much playing strength you should get for 1 or 100 rating points? rating points aren't distributed due to such definitions and never were.
the exact meaning of a rating number (even if still just an estimate) is only defined within the pool of players present in that exact rating list where this rating number was published. even in the very next list, *the exact same number would mean something else*, irrespective of what "inflation" markers would say - and notably, even if all "inflation" markers would say that there had been no inflation at all. the reason? the player pool would have a different constitution with shifted relative strengths, with some players gone inactive and some players gone active or added - the ratings would define an ordering within an altogether different group.
the two other possible implications, b) and c) above, would need some agreement of how "inflation" should be defined, in order to to be measured at all, but with most sensible definitions of "inflation", everything i've seen and calculated so far, indicates that b) is wrong and c) is very wrong. i still think there's a good chance that many readers will interpret mig as saying all of a), b) and c) above, and if so, mig's just strengthening a myth, which is the opposite of spreading enlightenment, imho.
"Aerosvit 08 + Biel 07 + Corus 08 + Baku GP = 2"
based on wang yue and onischuk getting 0,5 points in mig's summary, we must assume that carlsen should be clear 2nd in the rank above, with 1 + 0,5 + 0,5 + 0,5 = 2,5 points
I'm new and need a little help understanding the rating cut-offs to create say a category 18 tournament. Anyone have a link to help clear it up?
jimbo, categories are based on average rating of all participants in a round robin.
cat. 15 = 2601-2625
cat. 16 = 2626-2650
cat. 17 = 2651-2675
cat. 18 = 2676-2700
cat. 19 = 2701-2725
cat. 20 = 2726-2750
cat. 21 = 2751-2775
cat. 22 = 2776-2800
i think you can figure out the lower categories yourself :o) due to the shift in the rating distributions, with more players having higher ratings, and ratings being what ratings are, comparing events based on category over much more than 5 years, isn't too meaningful, though.
In fact I don't see the need for categories at all- why not just simply give the average rating? Afterall a tournament with average rating 2725 is more similar to one with 2726 than 2700. The average rating needs no further decoding either.
I meant 2701 above- so that the 2 lower ones are the same category despite 24 apart, while the 2 top ones are different category despite only 1 apart.
I agree with REMIT. Giving the average rating is enough. A category has no usage.
I have seen an organiser not invite a local player just because of the potential category downgrade. Inviting the local player would have only dropped the average rating by 2 points, but under the category cut-off.
I think counting Aerosvit or Biel for the same amount as Linares is pretty silly. Corus and Linares -always- have stronger fields than most of these tournaments, depending on how many locals get into Corus in a given year.
Dortmund can be on and off; Anand had his worst result in years there a few years back, and never returned, yet Naiditsch has won the tournament. Weird.
I think one purpose of categories was for ease of norm calculations - e.g., FIDE would print tables that would break down the points required for an IM/GM norm based on the category of the opposing field in a round-robin.
That could be the reason an organizer would see a big difference in an average rating of 2433 vs 2435 for example. However, those category norm calculations are out of date now, though, so even that purpose is no longer there.
I like having categories as it gives chess some more "statistics". It's fun to mess with anyway. The real big "deal" should be made of ratings inflation which is at risk of rendering past players accomplishments less meaningful. Does anyone at all believe that we have 20+ players approaching Kasparov/Fischer/Karpov's peak abilities? Nah, didn't think so!
Has there ever been a cat. 21 or 22 outside of a match?
kgd, there have been several cat. 21's (e.g., Mexico City 2007, Linares 1998 (the first category XXI), Linares 2008 (the last cat. XXI)), but never a category XXII.
I think they made once a XXII tournament in Gran Canaria, it was something like Kasparov, Karpov, Kramnik (not sure), Topalov, Anand and so on, but just six or eight players. For the rating inflation, I do believe Fischer would get smashed in Foros by everyone, and Karpov wouldn't win many games (and lose a few too). For Kasparov, well, he was already having a hard time with the new generations, and not only because he was getting less strong, but also because the others were getting much stronger. It's a natural process, they've all learned from him and first Kramnik and later the others, have found ways to withstand Kasparov's strenghts while getting some chances to win (remember Leko-Kasparov from his last Linares)
Well Alejandro,
If todays players are as good as Kasparov, why do they not have rating 2851, like Garry had at his peak? As for relative strenght, Fisher, Karpov and Kasparov are better than any of todays top players. As for absolute strenght, only a few are having the same rating as did Karpov and Fisher at their peak, and, as already stated, none of todays players are at Kasparovs level (yet).
I didn't want to create controversy, I was just giving my point of view. I said that there were two factors regarding Kasparov strength at the end of his career: he was getting weaker and the others were getting stronger. I agree that I have still to see someone as strong as he was at his peek, but in general the level of resistance is much stronger today. I wasn't saying everybody would beat Fischer and Karpov today, but the elite players i think they could. Remember what Kasparov himself said in MGP: Fischer, that at the moment Karpov became WC he was already stronger than Fischer, incorporating the qualities Fischer already had. Well; now almost everyone in the elite is as prepared as Kasparov was and they have much more tenacity as they had then (see Carlsen, Karjakin, Leko, even Anand and Kramnik are much stronger in this field) And I would add; that's something Aronian lacks for being really strong; as he is now, I think he won't be stronger than Carlsen, Karjakin etc etc
You're fooling yourself if you think that Carlsen and Aronian could resist Kasparov as well as Kramnik or Leko.
"I agree that I have still to see someone as strong as he was at his peek, but in general the level of resistance is much stronger today."
Based on what evidence? Case in point- checkout Carlsen's latest tournament victory. Do you think that Karpov, Fischer, or Kasparov would have lost these drawn endgames? No way.
"Well; now almost everyone in the elite is as prepared as Kasparov was"
Again- based on what evidence?
Alejandro, if you can pinpoint more clearly and confirm that category XXII, that'd be great.
"Well; now almost everyone in the elite is as prepared as Kasparov was"
No way.
I cant find the exact category in the net, maybe someone with chessbase and a very big database can find it: it was gran canaria 96, and the field was Kasparov, Karpov, Anand, Kramnik, Ivanchuk and Topalov, on a double RR. Carlsen had actually a better position in Reikjavik (rapid, ok) against Kasparov but ended drawing, and then he was really a kid. I think at the last two or three years of Kasparov's career there were about three or four players as strong as him. See the two last Linares tournaments Kasparov played. The others were maybe not as agressive as him in his opening preparation, but also very deep. Garry wasn't getting his usually extremely "good positions, crushing attack, 1-0 (or 0-1)"
Alejandro,
I found the Gran Canaria tournament you are talking. In the Chessbase database it is listed as "Las Palmas". It was an extremely strong event, but the ratings of the top players weren't as high as those today:
Kasparov 2785
Karpov 2775
Kramnik 2765
Topalov 2750
Anand 2735
Ivanchuk 2730
A great, great category XXI event (avg. 2757)--well short of cat. XXII, but who could ask for a better lineup? It would be great to see a cat. XXII event now (Bilbao in the fall looks like it will be just shy of XXII), but even then it may not be as strong or interesting as that XXI in '96.
Off topic: its odd that Kasparov would do his master class with the kids in New york and not Moscow. apart from helping Russian chess it would also be good PR.
Frankfurt Chess Giants in 1998 was Cat 22 and had Kramnik (2790), Anand (2770), Kasparov (2825) and Ivanchuk (2740), but it was only rapid.
The Category system has diminishing utility after one gets beyind Category 18. By then, if an event has an average rating of about 2700, one can asume that all of the participants are already GMs, so categories for Norm calculations are moot, anyway. I doubt that too many organizers care enough about the Category listing to change the lineup of a tournament, so that the Average rating of the event is, say, 2726, rather than 2725....
Does anybody know whether it is possible to earn a (GM) Title Norm if one achieves a Negative score in that event. A GM norm corresponds to a Performance rating that is at least 2600. If the player competed against a field of 2800 players, he could obtain a big minus score, and still maintain a 2600 performance--probably good enough for dead last. Against a 2800 field, a player could reach a 2600 performance by garnering 1 point out of every 4 games (or 1 draw from every 2 games). That means from a 10 round event, a 2600 performance might only require a score of -5, or 2.5/10....
With so many players converging on 2800 now--Anand, Carlsen, Ivanchuk, Kramnik, Morozevich, Topalov--the chances are improving that we will see some sort of Category 22, or even Category 23, event.
I say: if you can get 5 draws from 10 games against 2800 players, you're well worth your 2600 rating.
Theorist is correct; the ELO system actually understates the strength of the very top players (which is why Moro, Mamedyarov, etc. score so heavily in weak tournaments, run up their ratings, etc., but don't always perform so well against the regulars at Linares).
I too see Carlsen as 2.5
I basically agree with Mig that the current rank of players in a tournament is more relevant than the category or average rating.
Getting the ranks is not worse, as frogbert would suggest. Yes it is harder to be no. 20 now compared with 30 years ago, and it takes more to be no. 20 now compared with 30 years ago, but so what? One has to be measured against his/her peers.
Kasparov knows much more than Steinitz, but they are still on the same list of world champions, while Svidler or Gelfand, among others, who almost certainly also know more than Steinitz, are not on that list.
It is not an issue that a current rank 20 player knows more than a rank 20 guy in 1970. It could be argued that the earlier (1970) rank 20 guy, if he has current challenges and facilities, might also reach the knowledge of current 20 ranked. So comparison of strength between different periods is not relevant. But the rank, i.e. comparison against peers, is more relevant. Hence it is a better measure of tournament strength.
Alternatively, we could compare ratings based on Sonas, or similar systems, which suffer less from the rating inflation.
I guess a super tournament is one where a top ten player has to really compete, rather than just finish two points ahead of everyone else with a 2900 performance.
The first point is very valid. There is no doubt Carlsen is the real deal .. but it is a stretch to say Linares and Biel are the same. Actually I have reservations about Dortmund too. As Mig writes, something with just 7 rounds shouldnt really count as a super-tournament.
Talking of Dortmund, Anand did return after his 2001 fiasco. He was sole winner in Dortmund 2004 over Kramnik, Leko & Svidler. He also played in 2007 just before Mexico and finished T-2nd.
Kramnik will be playing Tal memorial and Anand the Bilbao grand slam. Those will be their last events before the match. Interesting to see how they dodge playing theoretically crucial lines; it will be a great opportunity for their opponents to bring them down a point or two.
"Getting the ranks is not worse, as frogbert would suggest. Yes it is harder to be no. 20 now compared with 30 years ago, and it takes more to be no. 20 now compared with 30 years ago, but so what? One has to be measured against his/her peers."
henry, i think i understand what you are saying, but i need to correct you in your interpretation above - i didn't say that using ranks was worse, i simply said it had very similar "problems" as using categories - its meaning changes over time.
sonas made some method to rank "greatest tournaments in the history of chess", using world ranks of top 10 players in the event as criterion iirc (taken from memory). however, imo this measure is quite meaningless (despite possibly sounding like a smart thing to do), since the number and density of strong players are vastly greater than 20, 50 or 80 years ago (for intance).
to give an extreme example, an event with 10 players randomly taken from the top 50 players in 2008 could be seen as more "elite" ALSO relatively, than 10 players taken from top 15 in 1908, yielding a SMALLER difference between the strongest and the weakest player in 2008! the logical conclusion obviously is that the concept of "greatest ever" can't be quantified and solved by such an approach. in fact, personally i find most of the "greatest ever" debates pretty pointless - and efforts at using some such quantification to cut through the subjective opinions only adds deception to the mix as well... ;o)
of course, if the goal simply is to have a sustainable definition of "super-gm tournament" without the need to change the definition every now and then, something like mig's rough proposal could be used and have some justification. but assuming a development in size and activity of the group of (semi-)professional chess players as seen over the past couple of decades, the effect would be that future "super-gm events" would become increasingly "super" and would hence not be any more comparable to previous "super-gm events" than a current day category 18 event is comparable to a 1990 cat. 18 event.
and despite the success and usefulness of our ranking system in chess (i.e. ratings), every debate where elo-arguments are applied, seems to confirm an image of elo-ratings as a seemingly popular child that still feels alone and often misunderstood...
Dominating has its own flaw in being used as a criteria for comparing players of different times since in todays computer age chess doesn't differentiate much between the players at the top level chess. Anand taking half hour for a whole game and some GM taking the whole 2 hours, there has to be some way to differentiate that. Increase the complexity and then we can compare Fischer and Anand in terms of dominance!
"to give an extreme example, an event with 10 players randomly taken from the top 50 players in 2008 could be seen as more "elite" ALSO relatively, than 10 players taken from top 15 in 1908, yielding a SMALLER difference between the strongest and the weakest player in 2008!"
frogbert do you really believe that? perhaps you could gives a practical example of that. i don't know much who was in top 15 in 1908 but it sounds a difficult to believe.
I too don't agree with the rank thing for defining category. My feeling is, say if there were top 10 players with 10 rating points strength apart in the past, today there are like 10 players more or less the same strength as that of one of the top player of the past. And making it like 100 players in the same strength group today which had only 10 players (or 20 players if 10 sounds too much of an exaggeration) in the past.
Mig,I don't understand why Dortmund and Biel are "Supertournaments". They are cat. 18,which is very low now days. Corus,Linares and MTel Masters was cat.20,even Foros was cat.19.The simple reason that your favorite player "Big Vlad" is playing in Dortmund cannot do this tournament "Super".Nobody can't live from the past and the history,the question is who are you today? Obviously both tournaments have money problems. There is nothing wrong with that,but don't call them "Super" With all my respect Biel and Dortmund was Super in the past,but not nowdays.
Yeah Mig. Your relentless Kramnik-promotion is ruining this otherwise outstandling blog.
Yah
Poor Mig, target of ire from all camps. To Mr & Mrs Koster (aka Clubfoot), Acire & brothers inc., he is the Kasparov praising archangel.
How about using the rating gap between the tournament average and No. 1 ranked player to gauge the strength of the tournament?
Mig being relentlessly pro Kramnik makes me laugh. Mig is an associate of Garry the great who cannot forgive Vlad for taking the title away. Mig also has a preference for the more all out tactical styles of a Topalov or Moro etc. I think any "pro KRamnik" sentiments are an effort to compensate for his preference for other players. To me Kramnik is the classiest player in the world for sometime. His style is really impressive and while others around 2800 obviously have similar ratings there is a feeling that KRamnik understands certain positions beyond his rivals ability....