Records in chess and press coverage

[Christopher Kreuzer]

The subject of records in chess is quite topical at the moment, seeing as it is now official that in the January 2013 FIDE rating list, Carlsen (2861) has passed Kasparov's record (2851) to set a new mark for the highest FIDE rating ever. But what does this really mean? Will the upcoming Candidates tournament have more significance? Who really are the best players in the world at the moment, and what are the sort of records in chess that really mean something, as opposed to merely being attention grabbing?

My view is that it is the gap between the top players that really matters (Carlsen is 50 points in front at the moment, ahead of Kramnik on about 2810), but even that can be misleading (matchplay versus tournament play comes to mind).

One thing I have been doing is trying to follow some of the coverage of this record in news reports. Some of the coverage was back in December (Chessbase did a round-up), some of it is now. Some of the coverage I found was in other languages, which I thought might be of interest, but from what I can tell there has really not been that much written about it so far:

http://www.deltaworld.org/sport/Magnus- ... -Kasparov/
http://jurnalul.ro/stiri/externe/record ... 32780.html
http://www.formiche.net/2013/01/02/classifica-elo/
http://www.jornada.unam.mx/ultimas/2013 ... de-ajedrez
http://www.dradio.de/dlf/sendungen/sport/1963041/
http://sport.adressa.no/sport/sjakk/article262496.ece

It makes me wonder how much press coverage a world championship match with Carlsen as one of the players might generate. There was moderate coverage of Anand-Gelfand, from what I can remember. Probably impossible to ever match the coverage generated by the Fischer-Spassky match of 1972. Chess records involving younger players tends to get a bit of coverage as well.

[Neill Cooper]

I assume Wikipedia news counts as press coverage: http://en.wikipedia.org/wiki/Main_Page

[IM Jack Rudd]

Differences between contemporary ratings mean something intrinsic to the system in a way that the actual values of ratings don't, certainly. Trying to figure out how Carlsen's rating of 2861 now compares to Kasparov's peak rating of 2851 is non-trivial.

[Christopher Kreuzer]

Differences between contemporary ratings mean something intrinsic to the system in a way that the actual values of ratings don't, certainly. Trying to figure out how Carlsen's rating of 2861 now compares to Kasparov's peak rating of 2851 is non-trivial.

Yeah. Anyone want to speculate what Carlsen's peak rating might end up being? (It was Kasparov, I think, who said over 2900 is very possible). What age do chess players tend to reach their peak rating? Kasparov was 36 in July 1999.

[Ray Sayers]

It's a personal opinion, but I think with rating inflation what really matters when you compare someone like Carlsen to previous top players in the world is how he performs against the worlds best players.

At the moment Carlsen reminds me of Karpov in the the late 70s/early 80s - top grandmasters could get into pretty even endgames with Karpov, but he would just keep nibbling away and suddenly they were lost. Carlsen is doing this pretty regularly to the other top players.

I don't know how easy it is to compare Carlsen to Kasparov, as Kasparov was phenomenally successful in just blowing people off the board and walking away with tournaments. But then Carlsen has plenty of time to chalk up those tournament victories (he's not doing badly so far!).

Has someone done a rating list for previous top players which takes into account inflation so they can be compared to players now?

[Roger de Coverly]

It's a personal opinion, but I think with rating inflation what really matters when you compare someone like Carlsen to previous top players in the world is how he performs against the worlds best players.

Is it rating inflation, or have the standards improved? Carlsen is now over two hundred points above the 100th rated player, but isn't the 100th player a bit better than ten to fifteen years ago,? All that database study and work with chess engines must count for something.

A fifty point gap to the second player is impressive, although Fischer, Karpov and Kasparov probably had hundred point gaps in their best periods.

[JustinHorton]

It's opened up very quickly. Am I making it up, or did I read that had Aronian beaten Carlsen in Bilbao they would have been practically level?

[Mick Norris]

It's a personal opinion, but I think with rating inflation what really matters when you compare someone like Carlsen to previous top players in the world is how he performs against the worlds best players.

Is it rating inflation, or have the standards improved?

Interesting question, debated here:
http://www.thechessmind.net/blog/2011/1 ... nswer.html

[Ray Sayers]

Interesting article. I personally don't buy it.

The authors accept that ratings have inflated but because modern players are more skilful. They then do a lot of statistics to 'prove' the case. But of course you can say anything with statistics. They admit Karpov in 1981 was only 2695. I looked at the live 2700 list - Karpov would be below Vallejo who is 2701 (indeed Karpov wouldn't get into the list).

I'm sorry, but if Vallejo (or indeed most of these players) had played Karpov in 1981 there would have been only one winner. To say Karpov had less skill than all those on the 2700 list I just find ridiculous, no matter how many numbers you crunch.

[JustinHorton]

I'm sorry, but if Vallejo (or indeed most of these players) had played Karpov in 1981 there would have been only one winner.

But if the Karpov of 1981 played the Vallejo of today* then that wouldn't be the case. This is the point, that in chess, just as in other sports, standards rise.

I was interested to see this research: I spent some time just before and during Xmas looking at a lot of old games, and when you do that you're always liable to wonder how good, objectively, they really were.


[* Mind you Paco has apparently packed it in, for the while at least, so Karpov could no more play him than he could play Fischer...]

[Ray Sayers]

I did say it was my personal opinion.

If you think that Vallejo is more skillful than Karpov at his peak then then you are entitled to think that.

[JustinHorton]

Well, the computer doesn't seem to think he was any weaker - and that's the present state of science in this field.

[Stewart Reuben]

50 points difference is really not much except in a longish match. Style counts for quite a lot. Thus Korchnoi had a tremendous score against Tal.

Fischer had the objective of winning every game.
Karpov had the objective of winning every tournament and didn't mind tieing for first place. He boasted to me he had just played in a tournament, had
Kasparov had the objective of winning every tournament by himself, particularly winning with white. It is understandable that they had different ratings!
Karpov never had to meet Fischer. Had that happened Tolya might hve become stronger player than he did. H improved to meet the challenge of Kasparov after his peak. Yet nearly fought him to a stand-still.

There is little doubt in my mind that there was inflation in the FIDE Ratings last century. One reason was the ridiculous rule where players got a rating of 2200, provided they got 50% in the Olympiad.

I have worked out a systematic inflationary error in the current system due to the English. That will be corrected, but these experts never noticed.
If your rating at the end of the month ends in +0.5, then it is rounded up. If it ends in -0.5 then it is also rounded up. But it should be rounded down to -1, not 0.

It would be very strange if players hadn't improved. Computer use. Information retrieval. Coaching. Opportunities to play.

By the way, games played 3/1/0 should not be rated alongside 2/1/0. While there are so few we haven't bothered.

But returning to the original thread. There are various ways to try to generate publicity for chess. Best is for a young girl to beat a GM. The highest rating ever is easy for the non-chess media to have an idea what that means.

[Roger de Coverly]

By the way, games played 3/1/0 should not be rated alongside 2/1/0.

Why not? At its weakest level, all a rating system is trying to do is to rank players using their results. So it's white win, draw, black win which are the relevant data inputs and how an event chooses to score these to determine a competition winner shouldn't matter.

[Stewart Reuben]

Roger >Why not? At its weakest level, all a rating system is trying to do is to rank players using their results. So it's white win, draw, black win which are the relevant data inputs and how an event chooses to score these to determine a competition winner shouldn't matter.<

That is simply wrong. The scoring system causes people to play with different objectives. It is thus a different set of data for 3/1/0 from 2/1/0.
Also you yourself say the rating system has the objective of ranking players using their results. It is also used to predict future results.
Take two players who play against the same field and both make 50%. With 2/1/0 they have the same TPR. If one player drew all his games and the other won half and lost half, it would make no difference. With 3/1/0, the latter player would score 2/3.
For their next series of results, the prediction for the two players would be quite different. With 3/1/0, there should be at least another figure showing the win/loss rate of the player.
You cannot seriously believe that if all games became 3/1/0, the current rating system could simply continue? If we changed the Laws of Chess substantially, e.g. banning repetition, the validity of the rating system would be diminished, though probably not by much. 3/1/0 is a bigger change than has happened in the Laws for hundreds of years apart from rapidplay and blitz. The USCF includes rapidplay in the same population as standardplay. I doubt that would find favour with many internationally.