Blank Scoresheets: How the Data Gap Is Erasing Vietnamese Women's Chess

The laptop opened on the third row of seats and lit up an empty folder. Beside me, a coach typed his student's name into the search box, then typed it again, more slowly. Six months earlier, that student had won four games in a row at a team event. There was no PGN file. No digital scoresheet. The organisers posted the final standings on a board, photographed it, shared it on social media, and three weeks later the post vanished along with the event page. Those four games once existed. They now exist in no form an analyst can reach. Thirteen years of watching this industry taught me something uncomfortable. A game only exists if someone writes it down. Everything else lives in a coach's memory, and memory exports to no format a computer can read. I learned to read chess from paper scoresheets. Before the commentary chair I was a track athlete; when injury ended my competitive career I spent four years as a chess commentator for VTC. There I understood what no classroom teaches: in a tournament hall, the most important person is the one keeping the scoresheet. The arbiter runs the game. The players decide the result. But the recorder decides what survives. A game exists in three forms. The first is the experience of two people across a board, untransferable, unverifiable, gone when they forget. The second is the paper scoresheet, which organisers are required to keep but almost nobody digitises domestically. The third is the PGN file, which only appears when a game is played on an electronic board wired to a computer. At a standard tournament, only a handful of boards carry transmission equipment — enough to cover the boards considered interesting. At men's events, the top boards. At women's events, usually fewer. Some tournaments wire two boards. Some wire none. The rest of the event runs on ordinary boards, ends, and leaves no technical trace. That is my starting point. Not playing strength. Not prize money. Infrastructure. FIDE's rating system works on a simple principle: every game played in a controlled event produces a result, the result converts into a rating differential, and the differential accumulates into a player's Elo. The list is published in cycles; each cycle is a snapshot. The snapshot gives one number per player, but tells you nothing about how many games built it, at which events, against whom, over how many months. For Vietnamese women, the sample is thin. According to public FIDE lists, only a few dozen Vietnamese female players hold official ratings and compete internationally with any regularity, while the number of classical games each plays in a year can be counted on one hand if she does not travel often. The names that shaped the national women's game — Pham Le Thao Nguyen with her multiple national titles, Vo Thi Kim Phung with her regional medals, Hoang Thanh Trang, the Hanoi-born player who moved to Hungary and won the European Individual Women's Championship in 2026 — all have far richer international records than domestic ones. The paradox is right there. A Vietnamese player is often better documented abroad than at home, because only abroad is someone recording her games. In the summer of Russia, I went looking for queens who were not on the map. In 2026, while the world watched a men's football World Cup, I and three friends started a small project called "The Forgotten Queens," interviewing former women's internationals long abandoned by the media. One of them talked for three hours about matches nobody remembers the score of. She left me a line I have carried for seven years: we do not need pity, we need to be followed. To be followed, in the language of data, means something very concrete. It means your games must sit in a file someone else can open. When I turned my attention to chess, I found the same problem. Vietnamese women's chess does not lack talent. It lacks recorders. The metric most used to compare players by machine is average centipawn loss, or ACPL. The calculation sounds simple: run an engine over every position, deduct the centipawns each move loses relative to the engine's best choice, sum it, divide by the number of moves. One number. And a single number always appeals to a writer. The technical problem is that ACPL depends heavily on three variables almost nobody reports. First, analysis depth. The same game analysed at depth twenty and at depth thirty yields different figures, sometimes differing by ten centipawns per move. Without a common standard, cross-tournament comparison wobbles. Second, position type. In quiet positions the error band is narrow, so an imperfect move costs a few centipawns. In sharp positions the band is several times wider, so the same quality of thinking produces a far larger loss. A player who forces complexity will always post a higher ACPL than one who simplifies. The metric punishes courage. Third, time control. A five-hour classical game and a three-minute blitz game are not the same unit of skill. Averaging them together is a methodological error. In Vietnamese women's chess, none of these three variables is routinely recorded. So when someone cites a metric to claim one female player blunders more than another, I usually stay quiet — not because I have no opinion, but because I know the two datasets were computed differently without anyone saying so. The biggest blind spot of ACPL: it does not measure pressure, time remaining, or what an opponent prepared for you two weeks earlier. Chess is a contest between two humans inside a clock, and this number sees half of it. The opening tree matters next, because the data gap does not merely blur a player's image — it changes how she plays. A decent opening tree needs thousands of games in the same system, filtered by opponent strength and date. International databases cover popular lines well and are nearly empty in rarely played branches, many of which carry the fingerprints of women's chess. For a Vietnamese player preparing for an international event, most reference material comes from foreign opponents' games, not from her own compatriots'. A dense domestic database would let an eighteen-year-old know exactly which structures failed against her own teammates and which worked. Without it, preparation rests on oral transmission. Oral transmission is not wrong, but it does not scale, cannot be verified, and cannot resist the bias of the person transmitting. Worse is the loop this creates. When you do not know which openings have been tried and failed, you choose the safe option. Safe options produce less volatile games. Less volatile games are harder to sell to broadcasters. Games that cannot be sold get broadcast less. Games broadcast less leave less data for the next generation. The circle closes within a few seasons — precisely when the women's game needs to open up. At national-team level the consequence is asymmetry. Our opponents have data on us, because every game Vietnam plays abroad is recorded automatically. We have little on them if they have never travelled, and even less on ourselves. That asymmetry never appears on a scoreboard, but it appears in the opening decisions a player makes in the first twenty minutes of a major game. Over years of note-taking I built three metrics of my own to read a female player without an engine. I write them by hand in a notebook, because I do not trust spreadsheets other people computed. First, time deficit. In every game I record both players' remaining time at three checkpoints: after move fifteen, after move thirty, and at the end. The gap at move thirty reveals who trusts their position, whatever the final result. Many players win games having been nearly out of time long before, and you only see it if you write it down. Second, conversion rate. I flag every position an engine evaluates as favourable by roughly one and a half pawns or more, then count whether that side actually won. At elite level the rate is high and stable. At developing level it swings wildly — far more useful information than an Elo, because it points to the exact skill that is missing. A player can outplay an opponent strategically for thirty moves and then throw it away in the endgame. Third, accuracy in the last ten moves. I count serious inaccuracies in the final ten moves and divide by the moves in that stretch. This is the zone most governed by nerves and stamina, and the zone least analysed by the media. A game is decided at move sixty, but people only retell move thirty. None of these three metrics requires an engine, an electronic board, or a budget. They require one person to sit down after each game and write. That is what makes the situation irritating: the cheapest solution is the least practised. One Saturday in Shanghai I sat in a small tournament hall writing a nineteen-year-old's moves in pencil. She played the way only the fearless young play: pushing pawns, keeping her king central, refusing simplification. On move twenty-eight she offered a material sacrifice the engine later called correct. I noted her remaining time at that moment: forty-one minutes. Her opponent: nineteen. Afterwards I asked who was recording the game. She said the organisers kept paper scoresheets. I asked who would enter it into a computer. She laughed and said she did not know. The answer was not funny. In my notebook, across three seasons of domestic and regional women's events, most games I witnessed left no technical record beyond the slip of paper in the arbiter's hand. The games I transcribed myself far outnumber those I could find in any online database. This is my own data, not official statistics, and I say so plainly because I do not want it cited as a report. Even so modest, it supports a conclusion. The data gap in women's chess is not the quirk of one tournament, one province, one season. It is structural. A player's three rating curves tell three different stories, and we usually read only one. The classical curve is built from long games with large thinking time. The rapid curve from shorter games. The blitz curve from very short games, mostly online, where a player can log thousands of rated games a year. The gap between the curves is diagnostic. A player whose online blitz rating far exceeds her classical rating is usually a fast calculator with sharp tactical reflexes who has not yet had the chance to show what long games demand. Mostly, the gap is not a flaw. It is the trace of missing opportunity. For Vietnamese women, classical games played and recorded each year are far fewer than online blitz games. So the public image of these players is built from half the data: the fast, loud, shareable half. The classical half — quieter, more accurate — is left in the hall. People do not measure what she plays. They measure what she gets to play. Sample size is what I wish every sports writer understood. If a player's true win probability is fifty percent, the chance she wins four of five games is still high enough to prove nothing about her level. At five games the confidence interval is so wide it is nearly useless for any long-run claim. At thirty games it is still wide. Yet most articles about a female player's "surging form" are built on three to five games. I have written that way myself, and I know how comfortable it feels — you have a story, you have a character, and you do not want to ruin it by admitting your sample is too small. Coaches do not make that mistake, because they see sixty games a season. Journalists see five, and journalists decide the public story. Building a personal archive requires nothing dramatic. A notebook, a pen, a phone. I record in standard algebraic notation so any software can read it later. I photograph results boards before organisers take them down, because that photo is the only surviving source. I request files from the arbiter running electronic boards, and I note which boards had none. I film the last thirty seconds of each game to capture both clocks, because clocks are data recorded nowhere else. I tag openings by international classification code. And when analysing, I fix engine depth identically across games so comparisons remain meaningful. None of this needs permission. It needs patience and a slightly crooked belief that a twenty-year-old's game will matter in ten years. In 2026, when the pandemic closed stadiums, a women's football team I followed came close to dissolving after its sponsor withdrew. I interviewed the players and found most were selling goods online to survive. I proposed a series on empty stands and ran a fundraising livestream; in three days we raised twenty thousand dollars. The captain told me something I still remember: thank you for not abandoning us. In 2026 we did not save the club. We saved the belief that the club deserved to exist. I tell that story in an article about chess because the same logic applies. Believing a game deserves to be recorded is not an aesthetic belief. It is the material condition for a sport to understand itself. The commercial value of a game depends on whether it was broadcast, not on how good it was. That is the long story of the sports industry, and in chess it shows more clearly than anywhere because transmission costs so little. The value of a female player is not on a price list; it is in the fate of girls who dare to dream. When we debate gender equality in chess we argue at three levels: entry quotas, prize parity, airtime. All three matter. All three are expensive, contentious, and require negotiating with competing interests. There is a fourth nobody mentions: recording infrastructure. One wired board and one data-entry procedure cost far less than a media campaign. They are not beautiful, do not go on a poster, do not cut into short clips. But they are the first condition for every later argument to have a basis. The blind spot is specific. Reward presence and you attract people good at appearing. Reward depth and you attract people good at thinking. Global sport chose the first branch long ago, and chess is copying that choice without much deliberation. The second blind spot concerns how talent stories are manufactured. A female player usually becomes famous through a viral clip, a brilliant online game, a moment. That is a rational path in a short-attention environment. But it creates bias: players with solid classical foundations, strong across many seasons, rarely have a moment to go viral. They are not weaker. They simply lack the recording infrastructure for anyone to see their curve. People call that a surprise. I call it fate. There is an objection I often hear: if a game is truly good, it will find its way to the public. I do not believe it. A good game sitting in a notebook will find its way nowhere. It will simply sit there, correct and invisible. That objection also ignores something dry: international databases do not spontaneously learn about a small provincial event. They learn when somebody uploads a file. Uploading has never been a default procedure at most domestic events. The fix is not expensive. An arbiter who can enter algebraic notation, a connected computer, a rule requiring the file before prizes are paid. Together, cheaper than one closing ceremony. Its political cost is higher, because it forces organisers to be accountable for something others will later read. Accountability is what sports bodies tend to avoid, particularly in sports without broadcast rights revenue. I have seen a generation willing to press, and I believe they will never retreat. The female players growing up now play with a confidence I never saw when I sat in the commentary box. They analyse their own games with software, post them, exchange across borders, and wait for nobody's permission. What they need is not another compliment. They need a scoresheet entered properly, a file stored in the right place, a platform that lets someone ten years from now reopen the game they won on move forty-one. Her press does not live on a tactics board; it lives in a heart nobody has yet dared to believe in. And if the recorder is not in the room, nobody will ever know.