Greengard's ChessNinja.com

Corus 08 r3: Drawfest Monday

| Permalink | 55 comments

Sorry, didn't get up in time to put up a round 3 item before the games started. But if you slept through this entire round of the Corus tournament you didn't miss a whole lot, at least not on the crosstable. Unless Radjabov errs in his defense of a theoretically drawn but tricky endgame against van Wely, all seven games in the A Group will finish without a winner today. As far as I can recall that hasn't happened in the many years I've been covering Wijk aan Zee. Lots of days with six draws, but never seven. So come on, Loek! [Now drawn.] The 50-move rule will likely kick in on move 95... [Garry sez Black missed an easy win with 36..Bf5. Regarding the endgame, he points us to the well-known Salwe-Rubinstein, Prague 1908, game in which Black won thanks to White wandering his king into the center.]

Most of the games were fairly interesting and ended with obvious equality so don't get all preachy about giving 3.14 points for a win or whatever the anti-draw conspiracy of the week is. As I predicted at the start of the round, having the heavyweights with black is often a recipe for short draws. Anand and Kramnik equalized without undue effort and were happy to accept quick draw offers from Eljanov and Mamedyarov. Ivanchuk-Gelfand only lasted 22 moves but it took a long time. Ivanchuk spent 50 minutes on his fourth move! Gelfand faced 1.c4 e5 2.g3 c6 3.d4 e4!? with White last week against Karjakin (and lost) and decided to give it a try today with black. Ivanchuk pondered deeply before playing 4.d5 instead of Nc3. A promisingly unbalanced position led to rapid exchanges and a draw. Another sharp opening, the d5 pawn sac line of the QID, also fizzled in Aronian-Leko. Eljanov played an enterprising g4 sac against Anand but got nothing against accurate play.

Similar things were afoot in Polgar-Topalov, a rare, if here expected, Sicilian sighting. Larry Christiansen was enthusiastic about Polgar's attacking chances on Chess.FM, but the Qg3-f4 plan didn't pan out for her. White had a few more weaknesses in the endgame but it was never more than a draw. That game was part of the plague of opposite-colored bishop positions we had today. Adams-Carlsen was another. The teenage co-leader defended well in his Open Ruy and seemed to have the better chances before simplifying and forcing the draw. 29..Rc2, keeping the active rook, looked logical enough. Mamedyarov got a tiny something of something against Kramnik's Petroff but his passed pawn never had any gas in the engine.

So Carlsen and Aronian stay in the lead with 2.5/3, nicely setting up their encounter in tomorrow's 4th round. It's a rematch of their thrilling candidates battle of last summer in Elista. Carlsen took Aronian to the limit before finally succumbing in blitz.

The main excitement of the day, other than wondering if Chucky was going to flag on move 4, was Nigel Short going down in spectacular flames against Hou Yifan. It was sort of a Berlin, only after the Allied bombing. What on earth is that knight doing on b7? It was pretty much over by move 18, although he hung on to move 23. Quoth the Nige: "It's embarrassing to lose to someone who is three years younger than my daughter!" She turns 14 next month. At this rate in a few years she'll be in the A Group with Short in the C! Of course the chess world already had to deal with this sort of onslaught when Judit Polgar came up. In fact, she has a terrific record against Short. Macauley Peterson, on the scene in Wijk aan Zee for the ICC, reported that Short was keeping his spirits up.

That was the start of a bloodbath in the B Group, which saw only one draw today. Smeets and Bacrot are in the lead now with 2.5/3. Caruana and Braun were the only leaders to win in the C, so both stay perfect. Macauley interviewed Caruana yesterday and he said he was disappointed when Miroshnichenko dropped out due to visa issues since now he won't be facing anyone higher-rated than he is. (How about taller?)

Playing as the favorite is often difficult for young stars, but so far Caruana hasn't "suffered" the fate of Radjabov and Karjakin and other prodigies, who are often thrown in with the sharks so early it's hard for them to shift gears to playing for anything other than survival. Carlsen mostly avoided this and learned to play for a win, although Radjabov seems to have recovered well enough. I admit that despite my frequent cheerleading for stronger opposition for Hou Yifan I was a little concerned for her tender hide against 2700's in the B Group, especially after she started with two losses. But obviously she's tough, not just strong, and will keep fighting even if she does finish near the bottom as expected.

By the way, after the round Anand said it was the first he'd heard of missing the repetition claim against Radjabov in round one! He said he'd lost the thread by then and didn't think about it. Fair enough, as it wasn't consecutive by a long shot. But still, in such a position and not in time trouble it would be worth keeping it in mind and checking your scoresheet, I would think. So the gaff ruling on the field stands.


Your report on Hou-Short game was hilarious. I laughed from the bottom of my heart. thanks for making my day.
Nigel, don't give up. you can still win a place in the top of the table.

to short's credit, he won yesterday w/ the evans!

I didn't see any of the other games - but Eljanov-Anand, there was general booing on ICC for White to have gone for the bloodless draw. according to smallville et al, there was play for white.

11.g4 has been played by Irina Krush. (Although that was 10.g4 but the same position. Eljanov-Anand included 8..Ba5 9.Qc4.)

Mig wrote:
Anand said it was the first he'd heard of missing the repetition claim against Radjabov in round one! He said he'd lost the thread by then and didn't think about it. Fair enough, as it wasn't consecutive by a long shot.

Maybe Anand could invest in a MonRoi. I am told (could be wrong)the device enables the player to see 2D diagrams of any/every previous position in the game. Seems like that would be an advantage over pen & paper.

I am not so sure it was wise of the MonRoi people to provide diagrams of any position other than the latest, when their device is in formal tourney mode.

corus is pretty much boring without'the moscow artists'(morozevich and grischuk).first with his unconventional style,always going for complicated positions,the second outplaying(mostly as white)the'giants'only to get in big zeitnot and many times missing easy wins.each of them would have score around 7 decisive games,dependig on their mood they could score big+ or big-,but surely they could stir things up.

3.14 points for a draw! I LOVE it! This has to be my favorite Mig line since his comment about the hippies having to pry the remote for his A/C from his dead fingers.

GeneM: "Maybe Anand could invest in a MonRoi. I am told (could be wrong)the device enables the player to see 2D diagrams of any/every previous position in the game. Seems like that would be an advantage over pen & paper."

Yep, you would be wrong, at least acocrding to Monroi's site:

While in the recording mode, MonRoi's Personal Chess Manager only enables the recording capabilities. It disables any programs that could help a chess player during the game. It also disables access to information from the outside sources.

Looks like Carlsen is beating Aronian. They are 2.5 hours into the game.

Disadvantage of having big match at the year end is the lack of theoretical novelties from the big guys. Kramnik playing solidly yet getting the upper hand, Vishy going for queen pawn etc

Didn't Anand lose against Adams in Linares in the same line?

danyplayer, while you are entiled to your opinion, i don't think this corus is boring. I think we are witnessing some real fighting chess in all the 3 groups..dude i sure like what i am seeing.

Looks like Carlsen threw away a win. Oh well.

Kramnik's pressure with white pieces is simply phenomenal. He gets huge pressure from positions where anybody else would play the draw. He is taking 0% risks with white but is close to victory in all of them. With black he's a rock.

In my opinion, Anand will get the same kind of treatment that Kasparov had in London. He'll loose one or two games with black, and will have zero chances in all left games.

I can't remember of Kramnik being as strong in his career as he is today. In my opinion, within a few months he'll break Kasparov's rating record.

Then, even the blind will be able to see who's the all time strongest chess player.

What is funny is that I was saying exactly the same thing when Kramnik had lost 100 points, went down to 2729, was loosing everything...

At that time I was receving huge negative feedback here on this site by people insulting me when I was just saying that Kramnik was just ill, that he was the all-time best chess player, since he defeated Kasparov in a match and had positive records against Kasparov, Anand and Topalov (along with a subsequend track of victories in major tournaments).

Most of you were laughing at me... most of you were saying at that time he's a coward, and so on... so funny to think about it today.

A lof of you wouldn't have played a penny on him in Elista... most of you judged him simply on the rating difference between Topalov and him at that time.

Why do people only look at rating and don't look at the chess moves??? Those of Kramnik are sublime...

Oh by the way, two players are leading the field in Corus. Carlsen and Aronian. Two of Kramnik's favorite victims :o)

"...so funny to think about it today"
must be the dumbest line I have ever read :-)

It's like "haha, I am sooo smart, and you all are soooo silly"

You were having av "nice" monologue up there, Rusland...


With all respect, I doubt that Kramnik will ever surpass Kasparov's rating. It's true that Kramnik is solid as a rock and loses "once a year", but to improve rating that is already very high one needs to win more frequently than Vladimir does or will: his very own extra-solid style will prevent him from doing this. But in a match he is a crusher!

Speaking of Kramnik's rating. Does anyone have his overall rating performance for the time since his 'comeback' (from the Turin olympiad till now). it must be the best of everyone? What is anand's for the same period?
As a curiosity does anyone have Kramnik's rating performance for the games where he played the catalan during the same period? that must be through the roof.

Kramnik performed 2836 in 2006 and 2824 in 2007.. of course it was the highest of all players in both years.

I'm not obsessed with ratings. I'm not saying this proves that he is better than Anand, or anything, but if you perform well over 2800 for so long you can't be that bad a player ;-)

Here are Stefan Fischl's stats for 2007:


and for 2006:


Actually I think his play was even more convincing in 2007, despite the (insignificantly) lower performance rating.

those are great stats acirce,

my proposal for an antidraw tournament:

Sort those stats by draw percent and start your invitations through that order. Set a minimum rating and exclude those who play too little.

van Wely beats Topalov!

Thanks acirce.
I honestly don't put too much emphasis on the ratings either but as a measure of past performances (as in 'who was the best of x year') I think that they tend to be a quite good measure. At least both the 2006 and 2007 list corresponds quite well with what I subjectively would have thought.
Luckily we don't have to speculate whether Anand or Kramnik is the best (provided the match actually gets played). They are obviously very close though I would (subjectively) give Kramnik a small plus, but in a match anything can happen.

The figure that really counts in those stats is the losses. Kramnik's lost three times in two years: two of those were in the Topalov match when the circumstances were hardly normal, and the other one was to Moro, which can happen to anyone.

I don't keep track of ratings, but can someone who understands these things tell me how if he's performed at 2830 for two years his FIDE rating is only 2799?


The answer is that FIDE ratings have way too much 'memory' - the influence of great past performance lasts too long. It also takes too long for them to reflect recovery from a bad period (as with Kramnik when he was ill).

The 'professional ratings' that had influence for a while attempted to address this by giving greater weight to the most recent N games. However, N was 100, so 100 games over 10 years was treated the same as 100 games over 3 years. As a result, sometimes they seemed better than FIDE, sometimes worse.

Whatever one may think about Jeff Sonas chessmetrics ratings overall, he did a specific study on this question of rating lookback. He found that as a predictor of immediate future results, the winning approach (not surprisingly) was to use a fixed period of look back, rather than a fixed number of games, with recency weighting within this period. The period needs to be neither too long nor too short. When combined with recency weighting, he claimed 4 year look back was optimal.

I see Polgar beat Gelfand again. She has an amazing score against him.

Actually, Aronian was better after the opening and time trouble around move 30 make him to play some inaccuracies and then Carlsen later got the advantage. So, apparently Aronian could win this game around move 30 (and even according to Aronian's opinion, his exchange sacrifice would be winning but he made some "mistakes" later on) d Carlsen could win this game after move 34; overall a very interesting game.

35. Rxa6 was much better for Carlsen against Aronian. 35. Qg3 threw away any advantage. I am just saying at move 35 I think Carlsen has a win.

Great stats acirce.
2006/7 are very interesting to compare.
Has anyone noticed how Karjakin has jumped up from No.37 to 14 (even past Carlsen, if only by a margin)?
And Moro's draw percentage is just out of this world.

The future of chess is leading! Great job Carlsen, Aronian and Radjabov.

Obviously Ruslan and all the other Kramnik fans have forgotten about the phenomenon called Rating Inflation. Ever heard of it? Well guess what, Fischers highest rating of 2785 attained in July 1972 is equivalent to (if not slightly higher than) the 2851 rating achieved by Kasparov in 1999.

It doesn't require one to be a quantum astro physicist to decipher this. For example, if only ratings where to be considered, then most of the current grandmaters and IMs of today would be stronger than the likes of Capablanca, Alekhine, Lasker, etc, during their heyday.

So please, let's not have these misinformed comparions of ratings that are so many years apart. It's just doesn't work, nor is it an intelligent thing to do.

Amos wrote:

"Obviously Ruslan and all the other Kramnik fans have forgotten about the phenomenon called Rating Inflation... For example, if only ratings where to be considered, then most of the current grandmaters and IMs of today would be stronger than the likes of Capablanca, Alekhine, Lasker, etc, during their heyday."

I agree that there is rating inflation, but it is also true that today's players are vastly superior to their predecessors (not necessarily more talented, of course).

It is pretty clear that Capablanca - lacking the experience, training, tools and theoretical knowledge knowldege available today - would not be, say, a top-10 player in spite of possibly having much more talent than most players.

Capablanca vs. Kramnik et al.

There is a good reason why chess players are usually compared to their contemporaries. We could argue that without the 20th century advances in theory and training methods Capablanca could play (and win) a simul against most of the Corus 2008 field. Being first and making an unexpected discovery in any field is much more difficult than 'training well and knowing theory'.

Amos, you understand quite little about chess ratings, maths and statistics.

Just consider a random group of 1000 chess players. Give them all a 1600 rating, and let them play for a while. Once you come back, you'll have a wide range of ratings, going from quite high to quite low.

Now split again this group of players in two groups, the group with the 500 strongest players, and the group with the 500 weakest. Start again from 1600. Strongest players of both groups will have the same ratings, as the weaker players of both groups.

That is, first conclusion, a rating has strictly no value in itself.

Let's talk about what you call inflation now. You, as many chess fans, can see that the ratings of the top 10 today are higher than the ratings of the top 10 five years ago, higher than ten years ago, than 20 and son on.
So you decided to call that an "inflation". But... why?
Have you been thinking about the evolution of human population?
Have you been thinking about the evolution of the number of chess players?
Have you been thinking about the evolution of the number of professionnal chess players?
Have you been thinking about the evolutions in training techniques, in the evolutions in chess theory?

Here you can compare chess with, let's say, athletism. The difference between chess and athletism is that 2800 today has no specific reason to be worth the same thing that 2800 yesterday, whereas in athletism 10 seconds on 100 meters is 10 seconds and will remain 10 seconds forever.

For instance, would athletism have been based on such a rating system like the elo one, Zatopek would have been high above 2800 in his time.
Today, any pro marathon runner will run in less than 2h10. Zatopek's world records were around 2h25, far better than anybody of his time.

Here is my point. Even if Zatopek was the best man of his time, the most gifted in his field, he was just able of 2h24. You have ZERO reason to believe that if he'd train like today marathons specialists, he'd have been able to achieve better results. Maybe that 2h24 was quite close to his physiological bareer.

The reality is in fact quite simple. People become stronger at running and at chess as well because training techniques improve. A raw comparison between Alekhine and any 2500+ GM (that is, a chess game) would end up in today's GM victory as certainly as any 2h20 marathon runner (that is a very good amateur or a semi professionnal) would leave Zatopek 10 minutes behind him.

Besides this elementary reality, if you talk about probabilities, when you've got 100 times more players playing a game (there are certainly 100 times more chess pros today than there were in 1930), it's more than likely that the level of the world champion of that era will be somewhere in the values of today's top 100 (here I'm not talking about improving training techniques, just about raw talent).

Well think just about it. 10 years ago, the top 6 was Kasparov, Kramnik, Anand, Guelfand, Ivanchuk and Topalov. Today, apart from Kasparov who left, the 5 other players are just part of the top 10, because now you have Aronian, Radjabov, Mamedyarov, Morozevich, and Carlsen and Karjakin are coming. Why? Do Kramnik has less talent than he had ten years ago? I don't think so. Today, you just have MORE talented chess players because you have MORE chess players.

Therefore, I don't think that fischer's rating of 72 was worth Kasparov's all time peak. And I'm also quite certain that in terms of raw talent, Kasparov and Kramnik are the two all-time strongest players.

Ruslan , something that could prove you wrong is if we compared the average rating, and showed that it is higher now than what it used to be (of course there is the technical problem that it used to be that no rating below 2000 existed some time ago so it is not clear what averages to compare).

Also, when you compare the best player, you are far from the average. Therefore the statistic factor of having more players may not be so significant. The correct question, is how many standard deviations far from the average is the top player. Then you can make a comparison. I am pretty sure that with this measure, Kasparov, Fisher did best and next was Karpov probably followed by Kramnik. Proper analysis like this (I believe) have been made by Jeff Sonas e.g. http://db.chessmetrics.com/CM2/PeakList.asp?Params=199510SSSSS3S000000000000111000000000000010100

Obviously, all this question depends on how big time period you consider (Karpov for example was 1 or 2 in the world for maybe 25 years playing continouusly).

ruslan has posted some crazy things before, but I agree with most of what he has said here. I very much doubt there has been any inflation in chess ratings at all. If anything, there is a deflation - that is, due to improvement of training techniques, it is even harder to keep the same rating as before, if you are not improving.

I like to compare this to physics. Today we have hundreds if not thousands of physicists that have greater understanding of physics than Einstein had 100 years ago. It doesn't mean they are as great as him because one's greatness is measured against one's contemporaries and contemporary knowledge. But they sure know physics better than him - because they have better knowledge and training techniques and because they are standing on the shoulders of people like Einstein. In the same way, an average GM of today is not as a great a player as Lasker, as far as dominance of the chess world is concerned, but it wouldn't surprise me in the slightest if this average GM was objectively a stronger chess player than Lasker. Hell, I wouldn't be surprised if some of the present IMs were objectively stronger than Steinitz.

No rating inflation? Hm.

"In November 1986, FIDE decided to grant 100 bonus ELO rating points to all active female players except Polgar, which knocked her from the top spot in the January, 1987 FIDE ratings list."

Source: http://en.wikipedia.org/wiki/Susan_Polgar

"Here is my point. Even if Zatopek was the best man of his time, the most gifted in his field, he was just able of 2h24. You have ZERO reason to believe that if he'd train like today marathons specialists, he'd have been able to achieve better results. Maybe that 2h24 was quite close to his physiological bareer." - That is true. And it is doubtful modern training techniques would help someone like Capablanca whose inability to train hard was as innate as other sides of his personality like good chess intuition. But the more important thing is - how Zatopek (or Morphy or Capablanca) would do if they use the modern training techniques is irrelevant. The arguments like "well, if Capablanca was transported by a time machine to present, he would destroy most of the Corus A field" never cease to amuse me. Well, time travel is impossible, folks. Capablanca, by definition, has NO access to todays knowledge and technology and training methods. And if he had such access, he would cease to be Capablanca, in any meaningful sense. So we can't go by fairy tales and "arguments" that involve time travel. We have objective facts - that is the games of people like Lasker and Capablanca, as well as of modern players, and if we examine them, would it be really that surprising that people like Lasker and Capa, who lived 100 years ago are inferior to people with 21st century understanding, technology and training methods? I mean, that is the case with pretty much any other field - why shouldn't it be the case with chess?

There is no point (or way) in comparing apples with oranges...

The only thing you can compare when speaking of how good a player is, is how much above the average contemporary was he. By how much, I mean how many standard deviations far. I.e. how extreme, how unlikely was that a player so stronger than the rest, existed. This takes into account the total number of players as well as the possible increase in level.

Kramnik is not in a totally different era than Kasparov, and I think that their achievments can be compared more directly. And yes, Kasparov was for much longer in the top and even when Kramnik is in the top he hardly shows the performance that Kasparov (or Karpov actually) did.

As for the quality of chess (not the results), it is quite subjective and I am not entering this discussion. Only the reminder that his unbeatable match tradition is not that impressive. He bearly won Topalov (+1) drawn with Leko and ok beat Kasparov with +2 (the only achievment that gives him a position in history.. And the reason that this is so great is only because kasparov was great)

An off topic final note, being a theoretical physicst, I strongly doubt that an average nowdays physicist is better than Einstein. Take his computer away, his calculator also and all his books, and ask him a general physics question (not about some non-existing things called strings :-))...

Why Anand will play d4 against Adams only, last time it was Linares 2005 that he played d4 in classic game ( Ihope I am correct). He never plays d4. Anyone wants to enlighten on this.


Pyada, I think it is because he has a match against Kramnik, and he wants to keep any opening novelties he has for there. Adams is a Petroff/Spanish player as black against 1.e4, and, without a novelty, Anand would get no advantage at all (with white!)
So he tried 1.d4
1. e4 e5 = that's where we are with current theory :(
That's where we have arrived with so much theory.

Ruslan's thread does not seem correct: when more people play chess and when there's a natural distribution of chess talent, one would expect the rating of the top player to be more than the one Garry left upon us at his peak. This is not the case. In addition, the referred to top 6 players of that time have not significantly improved their elos which in case of a broadening chess base would suggest a deflation rather than an inflation of the ELO rating system.

In his ChessBase videos, Kramnik described his preparation for WC matches, and I think he didn't spend more than half a year preparing for any of them. It's possible that Anand would spend more time on this, and certainly he has older novelties worth hiding, but it still looks unlikely that the world champion would sit through an entire tournament playing random openings, just to confuse Kramnik.

evanhaut, so we have 2 facts:
a) Kasparov had a higher rating than anyone now
b) top 6, top 10, top 100 average rating has gone up and is going up.

Fact (a) has no statistical significance, while (b) has. Which means-Ruslan's theory has a strongest support. Also, i think his example of marathon is a very, very good one.

jussu, I think anyone who would have a good Petroff novelty and a match against Kramnik somewhere in the horizon, would keep it.

There's a big difference between saving your most important ideas and playing completely different openings.

Would you think that maybe Anand is just warming up in using 1.d4 (besides his usual 1.e4) for his upcoming match with Kramnik? Some of them might think that Anand wouldn't improvise that way, but actually, I don't think the reason he does not play frequently 1.d4 is not for lack of opening knowledge (Anand has played plenty of time with the black pieces against 1.d4), it is more of feeling comfortable. Fischer was able to play a QGD against Spassky with success in 1972, Leko did the same against Kramnik in 2004, it is a "safe bet" to think Anand would do the same and I think Kramnik is perfectly aware of that.


On the discussion about past and present, bla, bla... it is interesting to think that even if the opening knowledge has increased through the years, does endgame knowledge and technique has increased all the time? Or there is a stationary point in history if we evaluate the endgame knowledge of top players in that topic? I think is pointless to compare Kasparov to Fischer, or Capablanca, etc. If we talk about openings there are plenty of new books, with more and more novelties because of continuous play, but if players can learn about endgames, the advanced books (like Averbakh's collection) are the same used more than thirty years ago ... What do you think?

Well, even Averbakh was (obviously) not available in Capablanca and Lasker's time.

But even parts of Averbakh are outdated now, as they have been found to be erroneus thanks to computers, tablebases and/or better analysis.

Dvoretsky's Endgame Manual (which is checked against computers/databases, I believe) and books like that were NOT available 30 years ago. Even for less advanced amateurs there are more and better books. Silman's book has just came out and if it is anything like his other work, it is a must have and it is not something one would dream about having 30 years ago. There are Endgame CDs and DVDs now. While it may seem like only our opening knowledge has increased, make no mistake - all aspects of chess training have improved and continue to improve.

To use the marathon analogy ...

Suppose I run a marathon today at 2h10m. Suppose further that twenty years from now I run a marathon in 2h05m. However, while I have shaved five minutes off my running time, the best in the world have shaved ten off of theirs.

I think that my rating (if we were to measure such things with ratings) should go down in this case, because my performance vs my contemporaries has worsened. I get the sense that some people would suggest that my rating should go up as my performance vs myself over time has gotten better.

Well, the whole argument is about whether ratings have inflated or deflated. But the question is - inflated or deflated in relation to what? It is obvious that being a 2600 doesn't get you quite as high a ranking as it used to - noone is arguing that at all. But as for the 2600 of today being actually weaker than the 2600 of 20 years ago - it is debatable.

So in your example, Tom O'Donnell, if rating will be a function of your world ranking, then no doubt your rating has gotten lower. But if it is a function of how well you actually run, it has gotten higher - since you are running faster, and in this case you can use your time itself as the rating. But like I explained above, in chess noone argues whether a rating like 2600 gets you a better place now on the rating list than used to - because it obviously doesn't, and one just has to see the rating lists. And your ranking in the running world is irrelevant if we are only interested in examining your objective performance. So I would say that people run faster on all levels, including the highest ones, so improving from 2h10m to 2h05m doesn't make your ranking higher, just like improving from 2600 to 2605 will not make your chess ranking higher, even if you yourself are playing objectively better chess.

Whether you like it or not, there is an obvious unambiguous way to tell whether there exist inflation in ratings. It has nothing to do with top players however. Is the AVERAGE player that determines whether there is rating inflation or not...

Even in economics, inflation is defined not with respect the most expensive items in an economy. Is the increase/decrease in the value of an AVERAGE item. For example, it tottaly irrelevant if someone makes today a car that is 10 times more expensive than all the previous cars. What is relevant is the price of the average car...

Even in the marathon example this may apply: I am pretty sure, that the average runner will have similar times that he used to have. This ofcourse include not only professionals. The fact that professionals do better now, is due to better training, the fact that there are more runners and therefore bigger probability that someone is much better than the average and so on...

Does anyone know of any intelligent and (or?) authoritative stuff on ratings and inflation?

I am no statistician, but it would seem to me that since rating is a measure only of how good one is relative to another, analogies from events with an objective performance measure are not helpful.

Is it not statistically true, for example, that if you have more people playing the range of ratings, and hence those of the best players, will tend to expand? I should have thought an important factor was where your base was. If one were to assume (improbably I dare say but never mind) that the players at the top have been of constant strength, then the weaker the players at the bottom are. the higher the ratings of the top players. So if society changed such that weaker players began to play in tournaments - if for example the entire population of the world began to be ELO rated -, one would expect in time that the average rating would go up.

That's good, derida. Now we only have to invent a way to unabiguously detect the average player.

rating inflation.
Computers + opening knowledge + opening databases have made GM's better but how much do they affect to average player who spents only few hours per week to chess. I would like to say that because of these new training methods todays pros have benefitted much more relative to hobby player and this is good reason that ratings of pros have to be higher than before relative to average player.
And that is exactly what is happening.

jussu, not that difficult. An attempt would be the following. We rate all the players that play in tournaments (i.e. pretend there is no minimum ELO). We do this for now and for ,say, 20 years ago.

Then the average rating is just the sum of all ratings divided by the number of players. we then see if this number is higher now or in the past. If it is a fair and/or closed system, then the average should be the same. Otherwise there is inflation.

We can also plot the graph : # of players per rating. From this we can see how many standard deviations far from the mean is the top player of each era and can make some kind of objective claims for who was better.

derida, it is easy in theory, but hard to do in practice, IMO. It may be relatively easy to keep track of all the players now in a country like the US where even lower rated people are likely to be registered with the federation. But in a country like Russia it would be hard if not impossible, as I doubt there is a centralized database of players below a certain level even now- and there definitely wasn't one 20 years ago. The way I understand the system works is that local TDs keep track of local players that are split up into categories (one is a "Class C" player or "Class B" player, but one doesn't have an actual rating, like a class player in the US does), and only relatively strong people ever get a rating or get entered into a national database - at least that's my understanding.

You are probably right Russianbear. But we could do this study to countries like US or even easier like Greece (I am from there), and make the assumption that this is a representative sample. Of course that could (and is) not entierly true, since Russia for example is very special country for chess and ignoring Russia and say China and India in the study could bring very biased results. Still we could deduce that in EU for example the average rating has gone up or has stayed constant which could be indicative if not conclusive for our discussion.

Twitter Updates

    Follow me on Twitter



    About this Entry

    This page contains a single entry by Mig published on January 14, 2008 12:39 PM.

    Corus 08 r2: Favorites with 1st Move was the previous entry in this blog.

    Corus 08 r4: Hits and Misses is the next entry in this blog.

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