Peter Zhdanov
5 May 2022

Rating Inflation in Chess: True or False?

The question whether there is rating inflation in chess or not is one of the most hotly debated topics in the chess community. Some players accept it as a universally acknowledged fact that the ratings keep growing year by year and hence assume that top masters from the past would have been rated higher today than they were back then. This belief is further enforced by interviews of former world-class players who are currently inactive. Some of them like fondly recalling their glory days when there was a limited number of 2600s in the world, and only a couple of 2700+ GMs. By mentioning this fact, they proceed to make a claim about the ratings having inflated. But is this true if we consider objective playing strength, not ranking?

Jose Raul Capablanca was a merciless chess machine who hasn’t lost a single tournament game from 1916 to 1924.

In his article titled “Would Carlsen have beaten Capablanca?” , GM Daniel Gormally tried to explain to the audience that chess standards have risen a lot over time. As a result, if we used a time machine to bring any of the masters from the past into the modern day, they wouldn’t stand much of a chance. The article attracted an incredible amount of attention from the readers and was received rather poorly. Let’s see what Daniel opined about Capablanca:

But what about Capa? How would he compare strength-wise to the players of today? I think he would come off rather badly. The difference in terms of knowledge and understanding between the players of today and the players of the 1920s and 30s is enormous. I know people who think Capablanca would struggle to break a 2500 rating if he came back today. I'm not sure if I completely agree with this, because Capablanca's raw chess talent is much higher than a 2500 player, but chess is also a game of information. It's difficult to see how Capablanca would have coped with the greater knowledge and sharp chess theory of today's chess elite. Sure, give him six months, and he might be able to bridge the gap, but it really is a huge leap. It's like trying to compare Jesse Owens to Usain Bolt, he'd be left many yards behind.

Here is what Dr. GM John Nunn had to say about one of the participants of the Karlsbad-1911 super tournament featuring such legendary names as Alekhine, Rubinstein, Nimzowitsch, Tartakower, Spielmann, Marshall, etc. In fact, they were all contemporaries and rivals of Jose Raul Capablanca to one extent or another. The quote is from “John Nunn’s Chess Puzzle Book”:

In order to be more specific about Karlsbad, take one player: Hugo Süchting (1874-1916). At Karlsbad he scored 11.5/13.5 or 'minus 2', as they say these days - a perfectly respectable score. Having played over all his games at Karlsbad I think that I can confidently state that his playing strength was not greater than Elo 2100 (BCF 187) - and that was on a good day and with a following wind.

Let me quote one more passage from the same source:

Returning then to the question as to how Süchting scored 11.5 points, the answer is simply that the other players were not much better. If we assume Süchting as 2100, then his score implies an average rating for the tournament of 2129 - it would not even be assigned a category today.

This assessment seems quite reasonable to me. I also had an impression that the winners of the tournament used to play at below IM level by modern standards. However, when you tell that to typical chess fans, they start puffing, rolling their eyes wildly and accusing you of blasphemy! The same people also love whining about the “overrated” top players nowadays and claiming that someone like Morphy would have beaten them all. They seem to fail to acknowledge the difference between the terms “talent” and “playing strength”. Paul Morphy was definitely an extremely talented individual who was miles ahead of his contemporaries in terms of chess skills. However, back then the overall level of chess in the world was so low that he had no chance to become a proficient player by modern standards. After all, our knowledge about all aspects of chess has increased tremendously since that time, especially in the last few decades of computer revolution. That being said, some ex-World Champions remain objective and admit that. For instance, when people start bugging Garry Kasparov who is the better player: Magnus Carlsen or Garry in his prime, Garry always points out that the next generation tends to surpass the achievements of the previous ones and offers analogies such as that modern students know more about science than Isaac Newton, but it doesn’t diminish his discoveries. It’s just that you can’t directly compare chess players (or scientists) who have lived in different eras.

Another reason why people tend to overestimate the prowess of the players of the past is that we typically remember only the best masterpieces of their lifetimes. And if you check out all the games, you will realize that they have been blundering left and right on their offdays. Even with game collections, it is not that simple. I recall a funny conversation with IM Vladislav Akselrod on this matter. One of his students asked him whose games to study as an example of attacking chess. Vlad’s first instinct was to give him a tip of advice that dates to the Soviet days: “Study Alekhine’s games since he was an exceptional tactician!”. However, when Vlad himself opened a book of Alekhine’s selected (!) games…at this point he made a vivid grimace as if he had tasted something sour! As a result, he decided that, with all due respect to Alekhine’s greatness, nowadays there are probably better role models for aspiring competitive chess players.

What about the statistical evidence? In the paper titled “Intrinsic Chess Ratings” by Kenneth W. Regan and Guy McC. Haworth (May 18, 2011), the authors posed two important chess questions:

1.     Has there been ‘inflation’—or deflation—in the chess Elo rating system over the past forty years?

2.     Were the top players of earlier times as strong as the top players of today?

They have arrived at the following conclusion:

…there has been little or no ‘inflation’ in ratings over time—if anything there has been deflation. This runs counter to conventional wisdom, but is predicted by population models on which rating systems have been based [Gli99].

The results also support a ‘no’ answer to question 2. In the 1970’s there were only two players with ratings over 2700, namely Bobby Fischer and Anatoly Karpov, and there were periods as late as 1981 when no one had a rating over 2700 (see [Wee00]). In the past decade, however, there have usually been thirty or more players with such ratings. Thus the lack of inflation implies that those players are better than all but Fischer and Karpov. Extrapolated backwards, this would be consistent with the findings of [DHMG07], which (like some recent competitions to improve on the Elo system) are based only on the results of games, not on intrinsic decision-making.

Since Elo ratings were introduced in 1971, only three players were able to set new all-time records.

If you talk to players whose careers have spanned decades and who are still active, they will probably acknowledge that the standards have risen dramatically. In the past, Mikhail Tal used to say that he would have crushed his former self from the days when he was the World Champion since his understanding of chess has improved a lot. And that was before the computer age, which has revolutionized chess! To give you a more recent statement, last month I have stumbled upon an interesting quote by Sergei Tiviakov in the “Chess for Life” book by Matthew Sadler and Natasha Regan, published in 2016:

What areas of your game have improved and worsened?

In general, my level of chess playing has become better. I am playing a better level of chess than before. But the problem is that the general level of playing chess in the world has increased much more than my level of chess. In comparison, I am falling behind. Sometimes I look at my older games and I’m really surprised how poorly I played before. In general, I think I’m playing better than before.

Sergei Tiviakov, three-time Dutch Chess Champion. Image ©

Sergei is one of the most experienced players on the circuit, who has played a few thousand tournament games, winning dozens of events and visiting a plethora of countries. At one point, he had a streak of 110 consecutive games undefeated. He knows what he is talking about, and I can totally relate to what he is saying.

When I and my friends were starting on “serious” chess and received ratings of about FIDE 2000, we had no opening files and could play pretty much anything we wanted from scratch. Our “preparation” was finding a scoresheet from the previous round and checking out what first move your opponent made. We were rather clueless about endgames, even very basic ones, such as the Lucena position in rook endgames. Our defensive skills were pretty much negligible. Nowadays, thanks to chess engines, databases and other modern technology, kids learn to play like engines from the very start. They know sharp lines up to 15-20 moves and have much more middlegame experience due to being able to spam online training games against a diverse field of opponents. They prepare carefully with databases, engines and opening trees, relying on statistics to pick the optimal weapons against the opponents. They are much more resistant since engines teach you how to sacrifice material for the sake of getting counterplay, or to defend stubbornly in general even when things look bleak. I could go on about this for a long time, but my informal assessment is that someone rated 2000, let’s say, 20 years ago, would probably be 1700 or so by modern standards. That is, if we are talking about a time machine type of scenario.

COVID-19 has had a tremendous impact on chess ratings as well. Namely, the improving players didn’t get much chance to compete during the last couple of years. As a result, nowadays in open tournament there are many kids rated like 1500-1600 who tear apart experienced 2000+ players. This, in its own turn, affects players with higher ratings as well, since if you were farming those 2000 guys before, now their ratings will probably plummet to 1800-1900, and you would be expected to score much more points against them than previously. I have seen GM discussions on FB on this topic, and we are talking about active professionals who are close to their peak in terms of strength, not some washed-up has-beens who need to come up with excuses to justify their poor performances.

Anyway, no matter what the research findings, logical arguments and empirical experiences of the leading GMs and computer scientists are, I still realize that I am fighting an uphill battle here when trying to prove my case. The eyes of most chess fans will remain dazzled by the very best masterpieces of the good old times and by the scale of the personalities of the maestros.

Wondering how to improve faster than the chess world in general? Try out Magnus Trainer, a one-stop app created by World Champion Magnus Carlsen and his star-studded team of experts!


Don’t Miss Out!

Get monthly product updates, exclusive content, and special offers!