Why are men better than women at chess?
This is the question that Carole Hooven, author of the excellent book Testosterone: The Story of the Hormone that Dominates and Divides Us, takes up in a new article in Quillette. What I like about the article is that it appears to consider every available hypothesis, and uses a scientific approach to finding evidence that either supports or weakens many of them. She tentatively settles on one that may be evolutionary in its origin, but I’m getting ahead of myself. Click below to read:
I believe Hooven got interested in the question when FIDE, the international chess federation, recently decided that transwomen would not be permitted to compete in their official chess events that were limited to natal women. Since transwomen are natal men, this implies that there is some advantage in being a natal man when it comes to winning at chess. Of course this move by FIDE could be considered transphobic, as it already has been, but in fact the evidence is that, regardless of cause, men are much better than women at chess.
How do we know this? Because, although some chess tournaments are limited to people of one sex, there are also mixed-sex tournaments in which women play against men. And those show a result similar to that in tennis: rankings based on those tournaments show that the top women chess player would probably rank below the top 200 or 300 men.
But why is this? After all, in tennis and other sports, men outdo women because there is an inherent athletic advantage associated with the male body: more muscles, higher bone density, greater grip strength, and so on. These advantages become prominent at puberty because they’re associated with the higher testosterone of males. (This does not mean, of course, that no woman can ever beat a man in mixed-sex sports; it is a difference, and a big one, in averge performance.) Likewise, transwomen, biological men who assume the identity of a woman, also retain these athletic advantages over natal women, especially when they transition after puberty. That’s why several sports associations have banned transwomen from women’s athletics.
But chess? None of the athletic advantages I mentioned should obtain for chess, which involves only moving light pieces of wood or plastic around a board. So why are men so much better at playing chess?
Hooven lists a number of hypotheses, which I’ve divided into the following categories in bold (my words). Carole’s text is indented.
a.) More males take up chess in the first place. If this is the case, then regardless of average performance, the top players will be weighted with more men, simply because even if the average performance is the same, a bigger curve for men (frequency versus score) means that the upper tails of high performance will contain more men. This will be the case regardless of the variation in performance itself, and simply reflects the fact that at every performance level, there will be more men than women.
This is a reasonable hypothesis, but one Hooven thinks is weak because of Scrabble and bridge, which more women than men take up but ultimately the championships are heavily dominated by men:
In a game of Scrabble, as most readers will know, two competing players earn points for creating words using one or more of the seven lettered tiles in their inventory, which they place on a grid-spaced board. Like chess, Scrabble uses a version of the Elo rating system. But unlike in the chess world, women dominate the recreational ranks of Scrabble, accounting for about 85 percent of all recreational players. Even at the competitive level, women generally outnumber men (which isn’t that surprising given that Scrabble is all about words, and verbal ability is one area in which women tend to outperform men). So if the participation-rate hypothesis were correct in this context, then women should be dominating the elite Scrabble ranks.
But they’re not. Instead, men dominate Scrabble’s upper tiers, as they do in chess. And the same goes for Bridge, another game that’s dominated at the recreational level by women.
Scrabble tournaments usually feature separate divisions, which are classified according to Elo ratings. Players with the highest ratings compete in the first division, in which there are few women. As the skill level goes down, the proportion of women increases, until you get to the lowest level, where women vastly outnumber men. No woman has ever won a national or World Scrabble Championship. (However, just last year, Ruth Li from Toronto did win the North American Championship in the High School division, becoming the first female to ever win any such regional championship.)
Of course, Scrabble and chess are different games that require different skills, and lessons from the former may not cleanly translate to the latter. But even if one confines one’s focus to chess, the participation-rate thesis doesn’t present a convincing explanation for the observed sex differences in performance.
In addition, over the last few years women’s participation in chess has increased substantially, but the average gap between men and women hasn’t narrowed much, though it has a bit in some places. Overall, though, this hypothesis seems weak.
Second, over the last 50 years or so, female participation in chess has increased measurably around the world—a fact that should, according to the participation-rate hypothesis, lead to a narrowing of the sex gap at the highest levels of play. And in a few cases, that has happened. In France, for example, the female participation rate increased from 6 percent to 15 percent from 1985 to 2015, and the sex gap in ratings also significantly narrowed. But overall, the evidence is mixed. In the mid-1940s, the Elo difference between the world’s highest-rated male and female chess players hovered around 150 points. Eighty years later, that figure hasn’t really changed. (Note that such comparisons are based in part on retrospectively calculated Elo ratings, as FIDE didn’t start using them until the late 1960s.)
b.) Sexism: women are driven out of chess or don’t take it up because of misogyny in the game. Sexism can manifest itself in many ways: simple harassment of women (which is reported), not taking women seriously, which can lead to a lack of self-confidence, lowered expectations, and a higher dropout rate. This should be mitigated to some degree by the existence of all-women’s leagues and tournaments. But Hooven doesn’t think that this is an important hypothesis because reduced sexism over time hasn’t narrowed the performance gap:
Such reports [of sexism] should, of course, be taken seriously. But I’m far from convinced that sexism and harassment are the main reasons why men outperform women at chess. We’ve already come a long way in battling sexism during my lifetime. And yet, even as women have made great strides in such areas as medicine, law, engineering, and academia, the sex gap in chess has barely budged since second-wave feminism took off in the 1960s. This all suggests there’s something else going on.
c.) Men and women have the same average performance, but men have greater variance, manifested as relatively more players in both the highest and lowest tails of the performance distribution. Hooven calls this the “greater male variability”, or GMV, hypothesis. The variability can involve many traits possibly involved in chess success: spatial ability, drive to win, willingness to practice, and so on. The key here is that there need be no average difference between men and women, but still the greater variation of men ensures that in the upper tails, where the champions reside, will be mostly populated by men. (The hypothesis can still hold even if there are some differences in means.) GMV may be the case for intelligence, as the average performance of men and women on IQ-related tests are about the same, but men are more variable. But again, this doesn’t seem to be telling for chess, though it could be important in STEM fields:
The GMV hypothesis is the explanation often given for sex differences in STEM fields, particularly the “hard” sciences such as physics. The idea is that even if there’s no male-female difference in average math or physics ability, there would still be more men at the very high (and low) end of the ability spectrum. These are the extreme outliers who are most likely to earn prestigious faculty positions, file many patent applications, and win career achievement awards. And there is, in fact, strong evidence supporting the hypothesis; many traits do tend to be more variable in men than in women.
But if the greater male variability hypothesis explained the male advantage in chess, then we should observe that Elo ratings [these are measures of chess proficiency involving games won as well as the quality of the opponent] for males would be more variable than those for females. That is, we would expect more male grandmasters not because males are better at chess, but simply because there would be fewer females at both the high and low end of performance.
But in most populations of chess players, that statistical pattern isn’t reflected in the distributions of Elo ratings. Those for males are not more variable than for females. In many cases, in fact, the variability among female ratings is actually higher.
d.) Males are innately better in traits that lead to success in chess. These involve average differences in traits and not just variances, and could include spatial ability, degree of aggression, drive to win, other aspects of cognitive ability, dedication to the sport so that one practices a lot more, and so on. Note that “innately” implies the differences don’t result from socialization or sexism, but are the same kind of differences that gives men advantages in “regular” sports. Of course these innate differences could interact with other factors, as the phenotype here (chess performance) always involves an interaction between genes and one’s environment.
Ultimately, Hooven considers this the best explanation because there is independent evidence that men excel in the kind of motivation, competitiveness, and “obsessive passion” that leads to monomaniacal focus not just on winning, but on practicing:
A more promising explanation for male dominance in elite chess involves motivation. A large body of research strongly suggests that the sexes differ in their preferences for competition. As both Kasparov and Repková have intuited, men are simply more competitive—that is, they have a stronger motivation not just to compete, but to win, in formal physical and non-physical competitions of all kinds.
Men are more likely to choose games that involve direct, one-on-one competition, in which the result is a clear winner and loser—such as chess. Women are less competitive even when interacting anonymously—for example, in online arenas such as massive multiplayer role-playing games. This applies even when players interact using avatars of the sex opposite to their own; situations in which social expectations and stereotypes should have a reduced influence on in-game behavior. Women’s performance and enjoyment tends to suffer when the competition intensifies; that is, when the stakes are highest or time pressure is applied. For example, the average male-female sex difference in “blitz” chess games, which allocate ten minutes or less for each player to make all of their moves, is greater than that observed in standard chess, in which each player has at least an hour and a half. Moreover, relative to men, in experimental and real-life conditions, women tend to opt out of tournament conditions.
So it’s not surprising that females, being less focused (on average, as usual) on crushing an opponent in some future tournament, might be less motivated to go in for the kind of hardcore practice that’s necessary to develop elite skills (“deliberate practice,” as it’s called, as distinct from simply practising by playing).
. . . . If your instinct tells you that males will be disproportionately drawn toward this kind of intense practice style than females, you’re correct. Studies show that boys and men are more likely to exhibit a “rigid persistence in an activity,” by which “the passion controls the individual” (“obsessive passion” in the literature). In anecdotal terms, we are talking here about the man who drops everything to become, say, a 16-hour-per-day videogamer, or a day-trader, or chess addict. Yes, some women take on these kinds of fixations. But men do it more often, and with greater intensity.
It’s long been known that measures of risk-taking, competitiveness, persistence, and aggression are higher in men than women, so this may be a key factor in the explanation. But are these differences due to evolution or socialization? After all, men are expected to be aggressive and behaviorally conform to a “male stereotype”. On the other hand, that stereotype itself could reflect behavior instilled by natural selection more in one sex than another, so it’s seen as the norm.
Hooven comes down on the evolution side, and I pretty much agree with her given these arguments as well as others (e.g., socialization should differ among human societies but the average behaviors don’t; our closest primate relatives, who aren’t socialized, show similar difference in aggression and competition, there are biological reasons to expect higher competition in males, and these traits begin to manifest themselves at a young age, presumably before much socialization can take place). Luana Maroja and I discuss similar sex differences in behavior (and their possible evolutionary roots) in our Skeptical Inquirer paper on ideology and biology.
Hooven:
That said, I don’t see evidence for the idea that socialization alone explains the stronger male tendency to focus obsessively on doing whatever is necessary to win, even at board games. And there are good reasons to think that this tendency has an evolutionary basis: In the animal kingdom, males tend to devote more time, energy, and risk to status competition, since this tends to pay more reproductive benefits for males than females. So it’s not unreasonable to suspect that boys and men have some kind of biological advantage—possibly underpinned by higher lifetime exposure to testosterone—that helps explain their over-representation in tournament-level competition in general. (While this particular brand of competitiveness may have a strong evolutionary explanation, it is unlikely to be the wisest reproductive strategy in today’s world.)
If this is the case, what about FIDE’s decision to ban transwomen from their women’s chess tournaments? (Some countries, including England, Germany, France, and the United States, don’t uphold this ban in their national tournaments.) In the end, since Hooven concludes that biological factors play a key role in men’s dominance in chess, for the time being FIDE’s ban makes sense:
Ultimately, sex differences in complex behaviors and skills are always a product of interactions between biology on the one hand (that is, our genes and their relatively fixed effects, such as hormone levels and body size) and our environment on the other (that is, factors such as our family circumstances, social dynamics, and cultural norms). Interactions between the two shape not only our skills and abilities, but also any emerging group differences. But none such complicating factors change the fact that the sex gap in chess is real and persistent. Given the circumstances that led to the creation of the female category, and the fact that many girls and women appreciate what this category offers, FIDE is correct to take the steps necessary to protect its integrity.
Of course the data we really need are the chess performance of transwomen playing against biological women, and as far as I know we don’t have that kind of data.
****************
A coda: Perhaps the thinnest book I own is called “Jewish Sports Heroes”, given to me by a Jewish relative. It’s thin because Jews are not usually among the best baseball, soccer, football, or basketball champions we can think of (Sandy Koufax is a notable exception). I’m not going to hypothesize about this religious lacuna, but what amuses me is that the last chapter in the book, and the longest one, is on chess, as Jews have always excelled in chess. If the writers wanted to produce a book of reasonable length then, they simply had to add chess as a “sport” coequal with sports like football and basketball.