Thursday, 8 March 2018

Variation in reverse

This is International Women's Day, which gives me a good reason to mention some recent research on women in chess that has got a little less publicity than one might have hoped: certainly less publicity than attended Nigel Short's successful efforts to make an arsehole of himself a while ago. That went worldwide: if the research received any wider public circulation than the Yorkshire Post I'm not aware of it. Hey ho.

Anyway, it was recently published in the journal Psychological Science. Here's what it looks like, and as you can see its title is Female Chess Players Outperform Expectations When Playing Men, which is something that should perhaps interest us on principle

but which is also, apparently, contrary to what some research had previously led people to believe. This is a question of stereotype threat - described in the paper as a situation
whereby an individual's awareness of a negative stereotype influences their performance
which in the particular case we are considered here, would involve women, playing against men and aware that they may be judged, as women, by the results, suffering in their peformance as a result of that awareness.

That is, if I understand it correctly. Don't rely on me - the actual paper is available here. Its author is Dr Tom Stafford of Sheffield University and his conclusion, from analysing the outcomes of a prodigious number of FIDE-rated games, is that not only does the effect appear not to exist in this particular context, but in fact a small effect in the other direction can be noted, that - as the title says - women appear to overperform when playing men.

Another way to put this is that men, considered as a group, may underperform slightly when playing women - to quote the paper again
one plausible mechanism is a degree of male under-performance rather than female overperformance. This coule be due to male underestimation of female opponents, misplaced chivalry or choking due [to] the ego-threat of being beaten by a woman.
Who knows? More research required, as ever. I should say I'm not in a position to judge the merits of the paper, having jacked in mathematics when I was sixteen. (Just on the paper's terminology, I'm not sure that "throw a game" is usually taken to mean what Dr Stafford uses it to mean, and at one point there's a confusion between "game" and "match" that has irritated me before. But these are not important points.)

Anyway, read the paper. (I'd print it out if I were you, it's almost impossible to read on a smartphone.) And after that we might try and put some thought into how we can try and address the absolutely execrable ratio between male and female chessplayers that exists in English chess.


Tom Stafford said...

Thanks for your interest! I am happy to admit my expertise is psychology and data analysis first, and chess second. Apologies for the game/match confusion. By "throw a game" I meant to imply "by accident" not deliberately, so sorry if that it confusing.

This result is new and (to me) exciting, but it is always worth remembering that no single analysis is definitive. See this paper for a contrary finding - women chess players make worse moves against men (they claim)

Anonymous said...

I read a long time that men tend to play more aggressively against women. That is to infer 'over-aggressively', that is to postulate 'unsoundly'. This study seems perhaps to confirm that?

The blueweasel -lettingHisTitsGetInTheWayInSomeKindOfBlueJokeThatOhGodWhatTheHellIsThisCrapIAmWritingAAgh

Chris Goodall said...

"And after that we might try and put some thought into how we can try and address the absolutely execrable ratio between male and female chessplayers that exists in English chess."

Withdraw all funding from men's international chess and give it to the women.

And challenge every man in the grading list to find the cleverest woman they know and teach her to beat him at chess.