Three in Four Female Chess Players Report Bias or Harassment, FIDE Survey Says — But the Sample Is Still Missing

A rented hall in Balewadi, March 2026. One camera, one 4G dongle, a borrowed mixer, and a three-person crew — of which I was one. Nine rounds of six hours each, 90 minutes for 40 moves plus a 30-second increment, ceiling fans groaning, and the small honest sound of a captured piece being lifted off the board. On that webcast I said less about the standings and more about the room.

Around the same time, calling a women's event in Pune, I saw something I have not been able to file away. The organizer took the mic and opened with “She's like our daughter, she plays very well.” The player in question was rated higher than half the men sitting in that hall. Nobody said anything. I did not either. That is the category the survey calls “bias” — not a crime, not a slur, just a sentence that turns a competitor into somebody's daughter.

Now there is a number attached to that category. A FIDE survey, according to secondary reporting, found that three in four female chess players have experienced gender bias or harassment — roughly 75 percent. The report arrived through a second-hand outlet. The sample size, the fielding dates, the geographic scope, the questionnaire design and a link to the primary document are all missing. What the material does contain is one substantive chess claim and, pasted alongside it, a website's privacy notice — the fingerprint of an automated, unedited scrape. That does not make the headline false. It does fix what we are dealing with: a figure whose direction is clear and whose precision is unknown.

To understand what that figure collides with, you have to look outside the playing hall. A girl's chess career in India rests on three supports: state cash awards, residential pipelines of the Velammal kind, and family. In the age-group events I have called, a parent's shadow is close to mandatory between twelve and sixteen — the outstation trip, the hotel, the tiffin box, the night-time supervision. For boys that shadow is optional. For girls it becomes a condition of entry. Around eighteen, when a player is expected to travel alone, the support breaks — and that is exactly where the drop-out begins. It reads like a hundred separate stories. It is one funnel.

Hold one structural picture beside it, pending verification against the primary source but stable for years: the top women's ratings sit around 2550–2600, while the world number one sits near 2830+ — a gap of roughly 230 to 280 Elo. Fewer than 50 women have ever held the full Grandmaster title, under 2 percent of all GMs. Neither figure argues anything. Both show where the pipe leaks.

And here the 75 percent number fuses two different problems into one. The headline says “bias or harassment.” The “or” is doing enormous work. Bias is a culture problem: the organizer's sentence, the coach's “you're a girl, play solid,” the male player's free advice, the arbiter's suspicion. Harassment is a safety problem, whose endpoint can be a criminal complaint. A single blended percentage answers both questions and therefore tells no organizer whether it needs coach training or a designated complaints officer. The treatments are completely different. A statistic that is rhetorically strong can be administratively useless.

Then there is the oldest problem with any number: where is the denominator? If the survey had about 400 respondents, the 95 percent confidence interval on a 75 percent estimate is roughly ±4 percentage points; at 100 respondents, roughly ±8.5. The headline is not fragile, but its precision is unknown. Larger than the sample-size question is the absence of strata. Seventy-five percent cannot tell us whether the experience was elite or district level, over the board or online, Tehran or Chennai. For a policymaker, missing strata are worse than a missing sample size, because without them you cannot aim.

Third, the thing written into every voluntary survey and absent here: selection bias. Surveys distributed through federation and platform channels over-sample the engaged — the tournament-active, the online-present — precisely the group most exposed to both bias and harassment. In more than twenty years around these halls, I have learned that the players who never arrive are the ones no survey records, and their stories are the heaviest. There is also a technical gap: we cannot tell whether male respondents were asked the same question. Without a comparison baseline, nothing establishes that the problem is gender-specific rather than general.

And the hardest thing to write coldly: the most severe conduct will also be under-reported here. The professional chess community is small and dense. Coaches, arbiters, organizers and sponsors all live on the same street. A player who complains must expect to meet the same organizer's invitation next season. That is why 75 percent may not be a ceiling but a floor.

One more thing belongs in the ledger, and it never appears on a scoreboard — event selection. A player who does not want to travel to a particular city, a particular hotel, a particular organizer's event simply is not in the draw. She does not appear in any rating list that month. The most direct performance cost of bias is not distraction in preparation; it is the tournaments that never get played. Every proxy that would make this measurable — withdrawal rates by gender, participation continuity across a cycle, drop-out by age band, error rates under time pressure — is absent from the material. The 75 percent is a prevalence metric, not a performance metric.

Here is the part least likely to be said. This story will be read as a condemnation narrative — some bad men, some bad incidents, and a number underneath. That is not where the number points. Its real target is the participation structure: who stops playing, at what age, and who remains in the hall after they leave. Second, an institution publishing an aggregate while withholding strata is not sloppiness; it is a grammar. It usually precedes a policy package rather than answering a scandal. A report with no name, no incident and no date tends to announce its next step later. Third, the figure means something different in India, because risk here maps onto logistics — the shared room, the accompanying parent, the overnight train, the unfamiliar city. Every year a middle-class family runs that calculation about whether to send a daughter alone. It never enters a survey, and it is the most accurate survey this country is running.

What to watch. Whether FIDE publishes the primary document, the questionnaire and the sample size, or leaves the number suspended. Whether age-group national championships start handing exiting players a short form — the real data will come from there, not from the podium. And whether the next cycle puts a named complaints officer at every major event, or just a notice board.

At one in the morning in Balewadi I wrote that day's last line about the three pieces left on the board. The girl whose rating was higher than half that hall's had no name anywhere in the report. She will not be in any survey. The question is whether she will be at the board next year — and that answer will not come from 75 percent. It will come from our hall.