The term "rating inflation" was first used by Nigel Short when referring to short term fluctuations due to incorrect "K-factor" in the statistics, not related to any sort of overall number (only how it may vary due to a good/bad tournament). http://en.wikipedia.org/wiki/Elo_ra...

FIDE uses a different statistical metric resulting in different numbers between the two organizations but this is also unrelated to any sort of historical comparison.

In any event, Arpad Elo its creator considered it futile to compare people from different eras, he compared it to "the measurement of the position of a cork bobbing up and down on the surface of agitated water with a yard stick tied to a rope and which is swaying in the wind." http://en.wikipedia.org/wiki/Method...

i think that notion is silly, too. and at no point has chess ratings (whatever system we're talking of, whether chessmetrics, fide or anything else i know) tried to say anything about "best ever" in terms of chess <skills>.

any rating system based on results only is relative per definition, and comparisons across eras are totally meaningless using result-based ranking systems like that of fide.

1) the <fide> rating system didn't exist in 1966, the first lists were published in 1970/71.

2) like nimh points out, doing "random" adjustments to ratings, based on some arbitrary choice of defining "rating inflation" (one of the more misunderstood terms among chess fans, i think), does <nothing> to make players from different eras comparable in terms of ratings. i.e. as a measure of how strong chess players they were.

<Just have in mind it was LMAJ's hilarious statement: <If you really believe this, you know almost nothing about chess. I would say most of the change in the numbers is pure inflation.>>

actually, and technically, that comparison doesn't prove zilch about rating inflation in itself, whether you think the comparison yields "unreasonable" results or not.

that most people (including myself) assume an amount of inflation is true, though - but there's absolutely no concensus on neither definitions nor estimates. and people typically go about the existence of inflation in utterly wrong ways. when fide and sonas seemingly think a small amount of (what i call systemic) inflation represents some kind of problem, it basically tells me that they want to use the rating system for something it's not meant for ...

<There are some attempts to fix the calculation (see: chessmetrics.com) and I assume it will be fixed in the future.>

fixed? in which way? by fooling people into thinking that a rating system (based on results only) without systemic inflation somehow would make ratings in different pools several years apart comparable, as a measure of chess strength?

oh, i beg to differ - they represent some of the best sport ranking systems in the world, in my opinion. it's only when chess players (and fans) start to obsess about

1) best ever players, and/or
2) best player in the world

ah, this was where it started?

oh well, the statement is nonsensical, but it takes too long to go into all details today/tonight.

in the fide system, fischer's peak rating was 2785, and kasparov's was 2851. those two numbers are not comparable neither before nor after doing random mathematics on them.

without any "advanced math", though, we know that both were number one by a big margin at the time they had those ratings. given their participation in a 14-man tournament in 1972 and 2000 (respectively) it would be possible to arrive at a reasonable estimate of how they might score, given the ratings of their 13 opponents and <assuming> a typical performance, in line with previous results.

but that kind of predictions is neither the primary purpose of ratings; the purpose is to provide a good and simple metric of <previous> results, so that results of contemporaries who have played mostly <different> events and opponents still can be compared. and this is something rating systems are fairly good at.

why is that useful? well, for instance i can participate in a tournament open for players in the 1900-2100 fide rating range and be <pretty sure> that the players i'll face there will be more or less of the same strength as myself, even if i haven't played a single one of them before. i think that's a <great accomplishment> of the rating system, and one that imo makes chess a much more enjoyable hobby than what it otherwise would've been.

lmaj, chess has changed in numerous ways. elite chess, for instance, has in some ways moved quite a lot from "science" to "sport" (compare time controls), while in other areas it has moved a lot from "art" to "science" (consider how people do chess analysis now compared to earlier).

also, the much increased importance of <concise, line-oriented preparation> and preparing lines for specific opponents (on every level!) to avoid preparation or pet lines have taken much more focus than it used to have. and for two reasons:

1) games are faster, and relatively more of the time spent on a game is spent on the first 40 moves

2) doing away with adjournments has implicitly lead to less <time and focus> being spent on learning and practicing end games

3) computers/engines have "killed" numerous dogma - modern top players accept "bad pawn structures", "risky" materialistic play and so on (read watson's book on modern strategy) to a much greater extent than the chess pioneers did, as long as <concrete lines> can justify their escapades

discussing whether chess objectively has improved or not is possibly the wrong discussion; the more relevant point imho is that chess has <changed> a lot. in the ways mentioned above and more. and when the competition changes (even when the basic rules are kept), the talent needed for success might also change. at least i find that a very reasonable assumption.

hence, some old time greats might not have had what it takes to become top 5 or top 10 today, while some current top players might not have what it took to be among the greatest some 50 or 100 years ago.

one thing's beyond debate, though: the number of people spending thousands of hours on top level chess in 2011 (think im and gm level strength) is many, many times higher than the similar number in 1911, one hundred years ago.

where do you guys think the level is generally higher? in a sport with 20 full-time practitioners world-wide or one with 20,000 full-time practitioners world-wide? (just a random example, not talking about chess here)

There are three dimensions.

1) Absolute playing strength. Most recent players have the advantage. In the course of time, chess knowledge accumulates, training methods evolve, and the increase of the number of people who pick up chess leads to the bigger level of talent among top level players. Obviously Carlsen is the best ever in this category.

2) Influence on the game and subsequent generations of players. The less developed a field, the easier to create new ideas, something unpredecented and longstanding. Here past players had the advantage. I don't think it's ever possible to surpass Steinitz in this behalf.

3) The dominance over the contemporaries. Naturally, the smaller the pool of active players, the easier it is to dominate. Hence again chess players of the past had more opportunities, although, not so distinctly as in the case of the previous two points above. Names to be included in this category: Morphy, Steinitz, Fischer (vertically); Lasker, Kasparov (horizontally).

To me, the incredibly outstanding players (not "merely" WCs no matter for how long) in their years were clearly:

1. Morphy
2. Capablanca
3. Tal
4. Fischer
5. Kasparov

Other greats that were WCs deservedly have their own fans, whatever. It's a debate that will never reach any agreement, and that's fine by me too, we all are entitled to our opinions.

After your line with 43...f5 44.exf5 gxf5 then 45.Qd1! threatens Qd5+ for the loose bishop on e5 and after 45...Qb5 to protect d5 (probably best), then 46.b4 and white is up a pawn with better position (K on f3 stands very well).

In the game, Shirov's plan with 41.Kf2 seemed best, and Carlsen seemed to make the best responses. Q on d5 is of course, very powerful. So much so, that 46.e5 was played prematurely for a dubious attack (big mistake at this level).

[Event "Biel Chess Festival"]
[Site "Biel SUI"]
[Date "2011.07.26"]
[EventDate "2011.07.17"]
[Round "7"]
[Result "0-1"]
[White "Alexey Shirov"]
[Black "Magnus Carlsen"]
[ECO "C60"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "119"]