> I like how the article looks at the absolute skill levels among professional competitors in chess and professional performers in orchestral music...
The chess example is controversial as it is not correct to consider the Elo ratings "absolute". They work poorly when players are separated into two groups, e.g. by time or geography, with little interaction between the groups. Many chess players believe there has been "rating inflation" causing 2700 today to be weaker than 2700 forty years ago.
Personally I don't believe the claims of rating inflation and broadly agree with the article that there are 30+ players active today making moves as good as the top one or two from the 70s.
Kenneth Regan et al have done comprehensive and persuasive work on rating the intrinsic performance of chess players by measuring move strength itself. From one of their papers: "We have shown multiple, separate, and novel pieces of evidence that the Elo system employed by FIDE has remained stable in relation to intrinsic skill level."
Kenneth has done fascinating research but somehow it just goes against my intuition as a chess player.
If I understand the paper correctly especially tables 2 and 3, then it shows that a 2600 in 1979 plays at the same strength of a 2600 today, ie no rating inflation.
This just seems so counterintuitive personally, because I know a number of 2600 players today. I managed a draw against a 2580 in a tournament game(I am 2350), while the legends of 70s and 80s at 2600 seem far beyond my grasp.
I just can't accept that Montreal 1979 was substantially weaker than your run of a mill 2700 tournament today.
Anecdata is one thing, but the proper way to critique would be to look at formulas that generate IPR in Kenneth's paper.
Do you think that at your skill level today (estimated approximately top 500 in the USCF [1]--the only recent rating percentile list I can find) you would have had the regular opportunity to play the 2600 player of yesteryear to measure yourself? (E.g., Karpov in 1972?) Or do you consider your own 2350 so inflated to the 2350 of the 1970s that you'd be in the same relative percentile and never get that chance?
From this [2] it looks like a 250 point ELO difference gives you roughly a 20% chance of a draw. And the general belief is that the higher you get into the ratings, the more likely that a draw will be an outcome (compared to a 250 difference between a 1750 and a 1500, for instance).
I can't readily find 1970s era chess ratings, but it might be interesting for you to look at the games of 2350 players circa 1975 and see whether you think you could keep up. (To test my memory, I picked some random guy I thought might be around 2600, Tigran Petrosian, and found he was 300 points better than you at his peak in 1972, which would only give you an 11% chance of a draw, by comparison).
The chess example is controversial as it is not correct to consider the Elo ratings "absolute". They work poorly when players are separated into two groups, e.g. by time or geography, with little interaction between the groups. Many chess players believe there has been "rating inflation" causing 2700 today to be weaker than 2700 forty years ago.
Personally I don't believe the claims of rating inflation and broadly agree with the article that there are 30+ players active today making moves as good as the top one or two from the 70s.