Something definitely changed in the late 1980's, the Super GM level had been stable at 2600 for probably 15 years and only Karpov and Kasparov broke 2700 then there has been an explosion so that now 2650 is outside the top 100. Maybe it was because of many more FIDE rated events? One of the reasons that England has relatively few GMS, and only 6 rated 2600+ is because there are practically no FIDE rated events held there.
Is Carlsen going to surpass Kasparov?
Sort:
When people are closely matched, draws are reasonable results. Just an impression, but it seems when the amateur the cries "play for a win you bums" he betrays his ignorance. I know draws are more common at higher levels, but if reasonable play ends in a draw...
IMDeviate wrote:
Moses2792796 wrote:
The studies which claim that ratings inflation isn't occuring are severely flawed, for a start they claim to prove this via engine analysis. Considering that modern players do all of their prep with engines isn't it obvious that they would play more moves which are recommended by engines, this has very little to do with the practical strength of the player.
+1
Intelligent people know this, but if you take this assertion to the cheating forum you'll get banned from the site. I know of several members who have met this fate.
Fact is you can't compare ratings from different eras or forms of chess...obviously, because the ratings pool was different and time controls/manner of play are different. Long time controls more or less create better quality games. I don't know any serious chess players who play their best at 1 minute games and their worst at correspondence.
Another fact is engines have had an impact on openings theory, especially in the last 5-10 years.
So naturally if you're looking for engine matchup while still in openings book you're going to get false positives.
Sadly, a lot of chess.com members and even some staff don't understand this.
Computer evaluations have been built on a foundation of human evaluations (e.g. Kauffman) so it can't be that people who train with computers are adopting a less human style... it's quite the other way around.
Training with a computer doesn't involve memorizing moves. Practical considerations are also taken into account. I've seen GM comments along the lines of "g5 is Rybka's suggestion which leads to crazy complications, better is to play ______ with a lesser but clear advantage to play for a win"
Obviously opening book isn't counted. I wonder what your explanation is that Capablanca's moves agree more with computers than most modern players and post-computer WCC.
So if it isn't style, memorized middlegame moves, or openings, I wonder in what way you think training with a computer will result in noticeably higher matchup rates?
And finally, when everyone seems to disagree with you, either you're the only true enlightened one or... lol. I mean how often do you think this happens?
You only have to look at the first graph in this article to see that ratings inflation has occurred, and that it started suddenly in the mid 80s. The best explanation I've seen (as discussed in the update to that article) is that the inflation is due to lowered rating floors for FIDE ratings.
People who think rating inflation doesn't occur need to come up with an explanation for why ratings were basically totally flat from early 70s through 1985, and then in 1986, practically overnight, they suddenly take off at a rate that has stayed relatively stable since then.
Its worth noting that since 1986 we have been in the steroid era. When
more testing is imposed on caffeine and the Petroff these ratings will fall back in place.