FIDE: Three in Four Female Chess Players Faced Bias or Harassment — And What the Survey Has Not Said

There is a sound I have kept from a Friday evening at a small chess club on Nguyen Chi Thanh Street in Da Nang: a twelve-year-old girl pulling her chair closer to the board, maybe ten centimetres, then folding both hands in her lap. Nobody told her to do it. Her chair was already close enough. She moved it because the previous week an adult had stood behind her, leaned down, read her move out loud to the whole room, and laughed. She said nothing. She simply learned a reflex: sit tighter, cover the board with a shoulder, shrink the space that other people feel entitled to walk into.

I begin with that detail because it is the only thing in this story I can state with certainty. It needs no survey, no sample size, no confidence interval. It needs someone to sit long enough in a tournament hall to hear a chair move. My career began with a mistake — reading an opposing striker's name wrong three times in a row into a microphone at Hoa Xuan Stadium in 2026 — and from that I learned to keep two notebooks: one for verified facts, one for personal tremors. When the pandemic emptied the stands and I commentated Viettel against Ha Noi FC in a stadium with no spectators, I learned something more: it is the silence of the terraces that signs every line of commentary. A microphone does not make a storyteller.

Today a FIDE statement puts one line across both notebooks: three in four female chess players report having faced gender bias or harassment, according to a survey conducted by FIDE. That is the entire substantive content available to me. One sentence. No sample size, no response rate, no publication date, no definition of "harassment", no split between bias and harassment, no regional breakdown. And something more telling: the page carrying the item devotes most of its body text to advertising and privacy boilerplate about personalised ads. An automatically assembled, ad-monetised aggregator page republishing a finding by the world governing body of chess.

That asymmetry is where everything below begins. I am not disputing the finding. I am saying that the way it reached readers places a weight on a bare percentage that the percentage itself cannot yet bear.

First, separate two kinds of question. A sociological survey can tell you how many people experienced a defined behaviour in a defined period. It cannot tell you who did it, where, under what circumstances, or how severe it was. A prevalence estimate describes an environment; it is not a finding of misconduct against any identifiable person. Conflating the two is the single largest legal and ethical risk in this story, and it happens at precisely the step where a headline moves faster than a methodology annex.

Second, the phrase "gender bias or harassment" collapses two very different severities into one sum. A patronising remark from a parent outside the playing hall and a physical violation do not share a scale, a consequence, or a response. FIDE's underlying report almost certainly separates them, with materially lower figures for the more serious category. A headline does not need a table. It needs a compact number. The collapse therefore likely inflates the most serious category in the version that travels furthest — and it does so to the disadvantage of the very people it claims to protect, because it hands critics an easy target.

Third, an opt-in multinational survey almost always carries self-selection bias. People who have been harmed have a stronger motive to respond than people who have not. This is a standard methodological critique, not an accusation. But it has a practical consequence: if the sample sits in the low hundreds spread across more than a hundred federations, the confidence interval is wide enough that a figure of 75 percent should be read as a directional signal, not a measurement. Making such a number stand up requires a large, geographically stratified, multilingual sample and a public methodology annex. Those things live in the primary report, not in the page I read.

Fourth — and this is the analytically most valuable layer, independent of whether the number is right — most incidents plausibly occur in local clubs, amateur events and online rooms rather than in closed elite tournaments with cameras and arbiters. If so, every safeguarding reform at elite level, however correct and necessary, touches only a minority of cases. FIDE governs events and licensed officials. It does not govern a club corridor at nine in the evening, nor the inbox of an online chess platform. That power gap is the largest blind spot in the architecture, and a global survey tends to flatten it into an average beneath which federations with weak safeguarding infrastructure simply disappear.

Fifth, the mechanics. People pin a percentage to a bulletin board, but what a chess system actually runs on is a selection pipeline: a girl learns the rules at eight, plays junior events at twelve, moves into open sections at fifteen, and at eighteen either enters a national squad or leaves. In the 12–18 hinge, every female player lost is not a lost medal prospect but a person who would otherwise have become a 2400-plus opponent in an open section. Based on my experience following tournaments and junior events, the pipeline does not break at the top. It breaks in the middle, where nobody takes photographs, nobody conducts interviews, and absolutely nobody keeps statistics. Sporting culture is a graveyard of names, kept alive for one more round only when someone bothers to read the headstones.

That mid-pipeline break explains a rating-list fact I have to mention in almost every broadcast. The women's world number one in recent years has generally sat around 2550–2620 Elo (figures requiring verification against each official publication date), while the open number one sits in the 2750–2850 band. That 150-to-250-point gap is misread nearly every time it is quoted. It is not evidence of a capability ceiling. Judit Polgar peaked at 2735 — a number that stands as an existence proof that the ceiling is not biological in any naive sense. The rating gap is a product of pool size and retention rate, not of skill limits. It is like comparing two swimming pools, one crowded and one nearly empty, and concluding that the water in the second cannot swim.

Across the strength map, depth is distributed very unevenly. China has both thickness and a long-reigning world champion; India is rising as a wave; Georgia carries a tradition; Kazakhstan and Uzbekistan have made notable strides in faster time controls. But most national federations field one or two women capable of international events, and almost no second row. A global sample therefore blends very different environments into a single division. Where safeguarding infrastructure is rudimentary, reported rates are plausibly higher; where complaint intake exists, rates may be lower — or higher, because people know they will be heard. Both directions are possible, which is exactly why a global average cannot serve as a cross-country comparison.

On governance, one point short reports skip: when an organisation publishes data unfavourable to itself, credibility is high on access and low on independence. FIDE can reach federations, players and events in ways no newsroom can. But a survey commissioned by the governing body is also an internal instrument in a budget and mandate negotiation between a commission and a council. There is nothing wrong with that. Readers simply need to know what genre of document they are holding.

And there is a data layer FIDE does not own: online chess platforms hold chat logs, report volumes and ban records. That is the densest behavioural data in the sport, and it appears in no federation report. A serious response would require platform–federation data cooperation — a structure with no institutional home at present. Today's story exposes that gap without naming it.

Here I have to turn into the wind, because this is the part most easily skipped. The popular framing is that "women's chess has a problem". That framing misaddresses the issue. The subject of this finding is not women's chess; it is the competitive and training environment — something built by adults, organisers and federations. Placing the problem on the affected group quietly converts the people who bear the impact into the cause, and turns a question of responsibility into a question of endurance. The distinction is not a matter of taste in wording. It determines where money flows: into technical coaching for women's squads, or into complaint procedures at tournaments.

There is a further paradox few state plainly: the 75 percent figure may be simultaneously a ceiling and a floor, depending on which force is stronger. Self-selection pushes it up. The absence of confidential intake channels and anti-retaliation rules pushes it down, because most people who have been harassed never tell a questionnaire. The two forces run in opposite directions and nobody knows which wins. Which means the argument over whether the number is right is consuming energy on a match that cannot be decided, while the operational question — keeping people in the game — could be acted on today without waiting for a methodology annex.

I am also not optimistic about the reaction cycle. When a percentage appears without sample, date or definitions, it will be rebutted on exactly that absence faster than any corrective data can arrive. Sceptics get an easy target. Supporters lose an argument. And the real loss, once again, is the person who moved her chair. Sport does not promise victory; it promises only a pulse willing to continue — but that pulse needs a room where nobody has to shrink their shoulders to survive.

At industry level, the most durable consequences will likely never make headlines: safeguarding clauses migrating into hosting contracts, into coach licensing conditions, into on-site incident procedures. They create a new cost line, small but real, and simultaneously a competitive advantage for organisers with the resources to certify themselves as safe. For Vietnamese chess, where the club scene and junior circuit are expanding faster than the management infrastructure, this is work that can be done before it is demanded — and doing it early is always cheaper.

Back to that Friday room. The girl moved her chair ten centimetres, and no survey recorded it. If FIDE eventually publishes a full report with sample size, response rate, a bias-versus-harassment split and continental analysis, that will be a good day for commentary as a craft — because then we can argue about causes instead of arguing about a number. And if that report never becomes public, the only thing left standing is what I heard in the hall: a chair moving, very softly, very steadily, every Friday evening. A mature chess nation is not one with more female grandmasters. It is one where a twelve-year-old can sit exactly where her chair is, put both hands on the table, and not have to think about who is standing behind her.