Bryan Caplan  

Chess and the Flynn Effect

Senate Oil Manipulation Hearin... The Market and the Government...

I'm not a chess geek - I prefer games where people laugh. But I suspect that my many chess geek friends will be interested in this paper that uses chess data to argue for the real-world importance of the Flynn effect. From the abstract:

Average IQ score has been rising for several decades but researchers dispute whether population intelligence really is increasing. Clear real-world evidence of a rise may settle the issue. I first examined the domain of chess, where performance can be readily measured and tracked over decades and people of all ages compete. The young increasingly have dominated the game since the 1970's, outperforming older players at progressively earlier ages. The median age of the top 50 players dropped from 38 years old in the 1970's to 29 in 1995, and the proportion aged under 25 more than doubled. The median age of the top 10 dropped from the late-30's in the 1970's to the mid-20's in the 1990s. The median age of world championship contenders dropped from 37 in 1971 to just 26 in 1994. The Soviet team which won the 1970 Chess Olympiad had a median age of 40 and the Russian team which won the 1998 Olympiad had a median age of 22.5. The longstanding record for youngest grandmaster, set in 1958, has been broken four times since 1991.
Do any chess geeks (or IQ geeks) care to comment?

Comments and Sharing

COMMENTS (18 to date)
publius writes:

The Soviets first dominated chess because the goevrnment needed a harmless outlet for the smart people, so that they wouldn't scheme to overthrow the government. Similarly, young people have been shut out of the job market in the Marxist states of Europe and North America, so the smart and young are finding another way to exercise their brains.

Jordan writes:

I'm no chess expert either, but there seems to be a corresponding phenomenon regarding the gradual inflation of Elo ratings over time.:

The average Elo rating of top players has risen over time. For instance, the average of the top 100 has risen from 2645 in July 2001 to 2665 in July 2006.[4] Many people believe that this rise is mostly due to a system artifact known as ratings inflation, making it impractical to compare players of different eras.

So it would seem that the methods the researcher used to determine that the youth "dominated" are important, and that there may be an underlying upward bias in chess ratings that don't make them objective in comparison with IQ ratings. Then there are other elements to consider, e.g. the energy of youth and so on that are often involved in playing competitive chess. That is, there are non-IQ related elements that may give youth an advantage. Simply showing that there are near-simultaneous upward trends among youth in both chess and IQ test data doesn't mean there is any correlation.

Fabio Rojas writes:

Quick observation: I've read a number of math biographies. Among older mathematicians, they wouldn't learn a lot of basic math till they hit graduate school. Now, a prodigy can quickly zip through the most demanding material by high school age. There is simply more information out there and it is more accessible. You can even download tons of free math material over the internet. I bet there's also been improvements in teaching complex materials, like chess or math, so people can process it more quickly than in the past.

Add improvements in nutrition, child care, and health, you have a smarter population with access to more and better information. So it's not surprising that you get younger and younger experts.

Tyler Cowen writes:

Totally and obviously true, that is my comment!

Steve Sailer writes:

But we've also seen the same phenomenon of dropping peak age in a lot of women's sports, such as in ladies' golf, where a 19 year old won a major championship yesterday. Remember the trend in Olympic gymnastics that led from matronly 20-somethings winning the gold to 14-year-old Nadia Comaneci? They finally had to set a minimum age limit in Olympic gymnastics.

Dropping peak ages in women's sports came from more intensive training of little girls. I suspect that chess players get more intensive training at the highest competitive levels today at younger ages through the Internet and computer programs. That would reduce the "experience" premium and stress the young man's advantage in raw processing speed.

Greg Hamer writes:

I would attribute a lot of the improvement in younger players vs older ones to the use of computers. In the seventies and eighties a computer was no help in studying chess. Now it is indispensable. People that grew up using computers will have an advantage.

KDeRosa writes:

I agree with Greg Hamer.

Acquiring expertise in the skill of chess playing requires thousands of hours of effortful study against "players' better than you. Computers are now able to provide this willing slightly better opponent for as many hours as needed. It is now much easier to get the practice needed to develop expertise in chess which is why the level of play has increased in the past two decades or so.

Arnold Kling writes:

This is about a tail, not a mean. If you have a larger cohort of younger players, they could dominate the upper tail even if, on average, the cohort is no smarter.

I am not sure about the computer effect. I cannot imagine a fifty-year-old chess master having problems figuring out how to run a chess program.

Lord writes:

It is only a matter of time before chess ends up as tic tac toe. All paths will have been analyzed and a canonical set of play developed that will yield a draw if played correctly. It will then become simple memorization and lose interest.

stanfo writes:

In other news, a look at work done by a random sample of scientists and mathematicians leads to the conclusion that current scientists and mathematicians are obviously more intelligent and have higher I.Q.'s than 100 years ago.

Erroneous conclusion, but not much different than the chess study.

But another interesting phenomenon, in the same vein, is that younger kids are better and quicker at learning video games than older people.

NeedleFactory writes:

First, observe the research about the average age when certain people, such as scientists and mathematicians, do their best work. Assume the same is true for chess players. Next, observe that it becomes easier and easier to learn chess. When I was a kid, it was a struggle to find a few good chess books (in my native language, English). Later good chess books became more plentiful, and still later, the internet arrived to provide a treasure trove of chess resources. Perhaps the "true" average age of best work is beginning to assert itself as the difficulty of preparation melts away.

The Flynn effect is thus but one explanation. I assume the internet will effect math and science "scores" similarly, with all its good free info.

Hyrum writes:

I agree with the comments before and think that more research is required from other groups like sports players to get a better picture of the society as a whole.

symme7ry writes:

What is happening is this:

Chess skill is a combination of experience/knowledge and raw mental power. Our raw mental processing power peaks in our early 20s. Up until the advent of good computer chess programs it was hard to get enough experience by your early 20s to compete with the best players, so the top players were in their 30s even though they weren't quite as mentally quick as they once were. With the training that computers now enable, players can accumulate much more experience by the time they reach their mental peak, meaning their overall chess playing peak will move closer to their raw mental power peak.

Dr. T writes:

I do not believe that access to computers is a primary reason for younger players achieving International Grand Master ratings. Bobby Fisher became the youngest champion ever without benefit of computers. However, he and others of his generation benefited from the worldwide availability of hundreds of chess books, including books analyzing the games of the top players (including the Russians). The ability to study such games, identify patterns of play, and devise counter-strategies is a big advantage.

Computers certainly help chess players, and computers can now do chess books one better (by storing the games electronically for move-by-move playback). But, there may be disadvantages to reliance on the computer. Top chess players can visualize chess games in their minds, and they can look at many probable outcomes. (One international grand master stated during an interview that he had worked out a checkmate 27 moves ahead!) I believe that excessive use of computers may 'spoil' young chess players and decrease their abilities to visualize games.

From what I have read about recent chess champions, only Gary Kasparov spent much time with computers, and that was because he was being paid to develop chess programs. Computers seem useful in the beginner to intermediate level and for working out 'mechanical' end games. Otherwise, at the 2400+ level, their biggest utility probably is as a repository of games and analyses.

KDeRosa writes:

From The Expert Mind

John Nunn, a British mathematician who is also a grandmaster, recently used a computer to help him compare the errors committed in all the games in two international tournaments, one held in 1911, the other in 1993. The modern players played far more accurately. Nunn then examined all the games of one player in 1911 who scored in the middle of the pack and concluded that his rating today would be no better than 2100, hundreds of points below the grandmaster level--"and that was on a good day and with a following wind." The very best old-time masters were considerably stronger but still well below the level of today's leaders.

Then again, Capablanca and his contemporaries had neither computers nor game databases. They had to work things out for themselves, as did Bach, Mozart and Beethoven, and if they fall below today's masters in technique, they tower above them in creative power. The same comparison can be made between Newton and the typical newly minted Ph.D. in physics.

Brad Hutchings writes:

I wouldn't dismiss the computer aspect. Great chess players need vast experience and a vast library of openings to draw from. Developing that experience may have taken years and years prior to computer chess programs becoming widely available., hence the higher median age of the best players in the 1970s. Now, any high school kid can get an advanced computerized chess tutor with hundreds of megabytes of historical chess data for $50. Where you would have to physically visit tournaments 40 years ago, you can find thousands of players online spanning most of the ability curve anytime you like. And young people generally have more time and energy to devote to such pursuits. 40 years ago, time to devote wasn't as valuable as accumulated experience, which took years and years to develop.

But all that aside, we all know the unspoken truth that chess has become a young person's game mostly due to performance enhancing drugs. <grin>

Jason Malloy writes:

Howard's theory, and paper have been criticized. The authors make the following arguments:

A) The trend in Chess Howard describes, unlike the FE, is not linear and began in the 80s.

B) Data explicitly fails to show psychometric differences between elite and non-elite chess players.

C) Two trends in coaching better coincide with the rise: Books and computer software. Literature about chess in the household is a strong correlate of playing ability.

D) The authors show that there has been a rise in practice time. (perhaps due to the fact that incentives have risen in the form of greater prize money) And there has been a rise in "chess professionals" (people making chess their sole career path).

E) There has also been a rise in participation; the International Chess Federation has 30x as many people as it had in 1970.

F) Also the starting age in chess has dropped significantly. (the Mozart/Tiger Woods Effect in the fostering of excellence should not be discounted - although this contradicts Judith Rich Harris dogmatism)

I would add the following serious problem: The Flynn Effect does not represent a rise in g factor intelligence, the predictive component of IQ tests.

Jim Glass writes:

I was in the chess world for a fair period of time and knew some of the top grandmasters.

My opinion is basically Arnold's: There's been a huge growth in the number of young players since the 1950s to 1970s.

Then you compound this by the fact that training techniques have gotten much better in chess just like virtually every other field of competition. The latter players are always better than the prior ones as a group.

Using that compound effect, a lot more of the way larger group of young players are going to move into the top 50, moving the 40-year-olds out, with no Flynn effect.

Interestingly, just about the only field of competition that I know of where performance hasn't steadily improved over time is horse racing.

NFL football, NBA basketball, MLB baseball, etc, the average team of any era would kill the top championship teams of 30 years earlier. If you look at film of Lombardi's Packers they look like high-school players today. And by size they just about are, (way smaller than even today's lower-level college teams). Same with Olympic athletes in individual sports. There are swimmers who as 40-year old seniors train as fast as they swam in the Olympics at 20. The gold medal winners of 20 years ago wouldn't make the team today.

But horse racing, zip progress. The best times haven't moved for decades. I don't know why.

Comments for this entry have been closed
Return to top