Greatest Chess Master in History

Oh yeah, we're a passionate bunch.  

https://www.chess.com/forum/view/chess-players/dang-it-long-time-for-fischer-kasparov-but-now-admit-magnus-might-be-best

My vote is pending.  Read and comment on my post, if you like.

There's a difference between who's the best (human) player of all time (so far) and who's the greatest player of all time. 

The best human is clearly either Kasparov (highest rating of 2852, and the player who stayed actively above 2800 for the longest time) or Carlsen (highest rating of 2882, and the only player other than Kasparov to maintain that rating for at least five years). 

When we compare careers, there's no contest though. Kasparov was World Champion for 15 years, defeated players from three different generations, and won more elite tournaments than anyone other than Karpov. In almost every critical game, Kasparov outperformed Karpov. When he had to win, he did win. 

Fischer's record as world champion should disqualify him from any discussion of "greatest".

I believe you already voted for Kasparov and if you did, I recorded it. You say that Fischer's record as world champion should disqualify him (presumably because he essentially retired after winning the title) - but that wouldn't even disqualify him if this was about the greatest world champion in history. Fischer was clearly the best player in the world. Any player who was or is the greatest master of his time deserves some consideration as the greatest of all time.

You're right. Ratings aren't useful for comparing players of different eras on an absolute basis. Because of inflation, today's highest rated player will almost certainly have a higher rating than the highest rated player of two generations ago.

However, ratings can be useful for comparing players of different eras on a relative basis. Since we're evaluating masters on the basis of their results against their strongest competition, we can use ratings among other things (world championship, major tournaments, etc.) in that evaluation. Since SmyslovFan argues that Fischer doesn't even belong in the discussion, he shouldn't have mentioned ratings. At the time of the 1972 world championship match, Fischer was rated 120 points higher than the second highest rated player. I'm quite sure there has never been a gap that large between the top two highest rated players before or since. Fischer's rating was more impressive than Kasparov's or Carlsen's - because what matters is his rating relative to the strongest grandmasters of his time. 

And he has not beaten them.

In 1975 Fischer had not played with him. First match with Kasparov was interrupted.

So Karpov is underestimated, despite he made a great chess career.

Karpov was one of the best of all time but I would bet most people would rank Fischer and Kasparov higher.

One of the great chess myths is that there's been rating inflation over time. 

Kenneth Regan, an IM and a professional statistician, has proven that not only has there been no rating inflation over time, Elo is actually an accurate measure of objective strength and can be used to determine how well people have played historically. 

Check out his work at CSE Buffalo. 

https://www.cse.buffalo.edu/~regan/publications.html#chess

I wouldn't consider the conclusion of one statistician proof. Scientific studies of the same phenomena often result in different conclusions.

But the argument about rating inflation is neither here nor there. Suppose there is no rating inflation. All that means is that the highest rated player today is stronger than the highest rated player of a couple of generations ago. We've already conceded that top tier players today are stronger than their counterparts of generations ago. No one in his right mind would suggest that Paul Morphy was as strong as Magnus Carlsen on an absolute basis. Once again, we're evaluating masters of different eras on a relative basis. If ratings are germane, they're only germane to the extent that they measure a master's strength against his contemporaries. 

Who has challenged Regan's conclusions with science? 

SmyslovFan is correct. Add my vote for Kasparov.

By the way, will someone nominate Judith Polgar? She obviously deserves it.

To me the idea of Intrinsic Performance Ratings is one way to look at things and very much not the best way.  While I have no doubt that the statistical analysis is spot on, it's also based on the premise that "skill [should be] based on the quality of decisions made rather than the outcomes of contests"  which would be all well and good if the players all had the same available knowledge, understanding, experiences and training to make decisions of equal quality.  To me (again) trying to apply an absolute is a faulty method because there are far too many subjective variables.  No one would deny that Carslen is objectively stronger than, say, Lasker, but that doesn't make Carlsen de facto a stronger player than Lasker since Lasker could only be as strong as his situation allowed.  The relative or subjective part (to me yet again) is far more revealing than number crunching.  
IPR is a skewed and less impressive measure of strength. 

Agreed.

We can estimate absolute strength of players. We have all needed data. We need a programmer and a mathematician.

We should take stockfish (or another chess engine) and analize all Fischer's games in 1970-72 or Karpov's games in 1975-1985 (or someone else in his best time) and calculate inaccuracies, mistakes and average centipawn loss by move.

1.  "Attempts have been made to measure the strength of players by running their moves through a strong computer engine and seeing how well they match, but of course that can say more about how "computer-like" a player was, rather than objectively how strong their moves were."  By Jeff Sonas from the article about rating inflation. 

2.  I think Sonas article is a better argument than Regan's.

Thanks for all the interesting stuff guys. Batgirl recently sent me a computer analysis, which included a CAPS score. All of Whites 'mistakes' came in a totally won position, so which move wins the quickest is pretty irrelevant. I trust the sonas numbers a lot more than computer generated game evaluations!! Many great players have tried  to outplay their opponents, and win games, rather than play the 'best move'. The article by Spectrowski that i linked earlier is very revealing in that respect, and has Lasker and Kasparov at the top of the list. My view is very simple - and reinforces the point that batgirl makes - anyone who is a great player within their own era is a great player. There are many examples, Steinitz for example, who rarely get mentioned in debates like this!

 @Philidor_Legacy thank you.

They are very close one to another. Kramnik has surprized us!

But it is the best we have! Percentage of moves that matches with Alpha Zero moves will be lower, but still close one to another.