The Discipline of the Empty Data Table: When Chess Analysis Is Forced to Say 'Insufficient Information'

In November 2026, in a small apartment in the Tianhe district of Guangzhou, I opened a data packet sent to me by an analytics partner after the seventh round of a national team championship. I opened the file at 22:47, after pouring my third cup of tea. The game-ID column was empty. The PGN column was empty. The engine-evaluation column was empty. The player-list column was empty. The tournament name read "N/A". The source read "N/A". The publication date read "N/A". And in the middle of the analysis template, a line of instruction text still sat there, never answered: "Identify entities from the information points above."

There were no information points above.

I have worked as a data journalist for twenty years. I once sat in the press room in Nizhny Novgorod in the summer of 2026, when Japan led Belgium 2-0 and then lost 2-3, and I was one of only seven female data journalists accredited to that World Cup. I was once rejected by a veteran editor for daring to write that Japan won 14 of 18 aerial duels after Marouane Fellaini came on in the 65th minute — he told me, "Women who look at football data only pick the numbers that suit them." I once predicted the collapse of Guangzhou Evergrande on fitness grounds before they lost 1-2 to Jiangsu in the 2026 final, when their total distance run was 6.4 kilometres less than their opponents'.

But I had never before opened a completely empty analysis template.

And I realised: this was not a minor technical glitch. This was the moment my profession was tested harder than any game of chess, because an empty template does not announce that it is empty — it simply stays silent. A young writer, an inexperienced writer, or simply a writer crushed by deadline will fill that void with whatever they have on hand: memory, guesswork, and narrative. People call that a shock; I call it data that has not been read. But this time, there was not even data to read.

Chess is the most perfectly data-rich sport humanity has ever created. Every move is recorded. Every position is scored by an engine down to the centipawn. Every game leaves behind a PGN string that can be reconstructed to the second on the clock. No other sport enjoys that advantage: football has only twenty-two players and a ball in continuous motion, whereas chess is a finite, discrete, machine-verifiable sequence.

For that reason, I always believed chess was where data must beat narrative. But the truth is the opposite.

In twenty years of writing about chess for the Chinese market, I have witnessed a paradox: the more data there is, the less people read it. A game can be analysed by an engine to a depth of forty moves, yet the article about it usually only retells the story of "who won, who lost, who cried". The ACPL metric — average centipawn loss per move — is the most objective number for measuring form, but it rarely appears in a headline. People prefer to talk about "soaring form" rather than a 47-centipawn-per-move decline compared with last season.

To do the work decently, I set myself a three-source discipline. A rating figure only appears in my writing when I have cross-checked at least three genuinely independent sources: the official list of the International Chess Federation (FIDE), the live ratings board of 2700chess, and the game database of ChessBase or TWIC. If all three point back to a single underlying figure, they are not three sources — they are one source counted three times.

I learned that after Euro 2026. While the whole football world chased the gegenpressing fashion, I sat down and checked Italy's data. Jorginho completed 1,247 passes, the most in the tournament, with 93.4 percent accuracy, but only 2.1 percent of them were line-breaking passes over thirty metres. Italy's PPDA was 9.9, while the tournament average was 11.8. I concluded that Italy were not pressing like madmen — they were controlling through positioning and short passing. I predicted their title while the mainstream media were still worshipping the pressing game of France and Germany.

Jorginho does not need to run fast, because he reads the maze before the crowd can see it. That method does not belong to football. It belongs to anything that can be counted. And it applies intact to the chessboard, where — sadly — people find it even easier to fabricate.

To understand why an empty template is dangerous, one must understand what a full one looks like. A serious chess analysis has eight layers. I will go through each one, and at each layer I will point out exactly what is lost when the data disappears.

The first layer is game technique. Here everything begins with the opening: the name of the system under its ECO code, which variation was played, and — most importantly — whether there was a novelty never seen before in the database. A high-level game can turn on a single seventeenth move. Without the PGN, I do not know whether that move existed. Without the engine evaluation, I do not know whether it was a stroke of genius or a fatal blunder. Without clock data, I do not know whether it was played in three seconds or thirty minutes.

ACPL is what separates a clean game from a sloppy one. A player with an ACPL below twenty in classical chess is considered accurate; above forty means a game full of errors. But ACPL only means something when the phases can be separated: the opening may be spotless because it was prepared at home, and then the middlegame explodes with error as the clock drains. If all I have is a single total ACPL figure, I am reading a game without knowing what story it tells. And in particular, if the game ended in time trouble, every technical conclusion must carry a warning — because an error on move thirty-five with forty seconds on the clock is not the same in nature as an error on move twenty with two hours left.

The second layer is player data. A player does not exist as a name; they exist as a coordinate with four axes: classical rating, rapid rating, blitz rating, and recent form. The first three axes are published periodically by FIDE each month, while recent form must be drawn from tournament results and compared against the live ratings board of 2700chess. At the very top, the gap between 2,750 and 2,700 is the gap between a title contender and a qualifier. But if I do not have a player's name, I cannot place anyone on any coordinate.

The most interesting thing at this layer is the age curve. In classical chess, peak performance usually arrives between thirty and thirty-five — much later than in physically demanding sports. That is why Viswanathan Anand was still in the top group past fifty, and also why an eighteen-year-old phenomenon deserves watching but not yet a coronation. Age is the only variable that never lies. But to read that variable, I need progression data across many seasons — something an empty template cannot provide.

The third layer is the tournament system. Chess has a complex tiered structure that outsiders often lump together: the World Championship Match, the Candidates Tournament, the Grand Swiss, the World Cup, the Grand Chess Tour, open events, and online tours. Each has a different qualification path. The World Cup is a knockout with more than two hundred players, where two Candidates places hide inside a bracket in which one slip in round three can wipe out a season. The Candidates is a double round-robin with eight players, where every draw carries strategic value.

To assess a tournament's quality, I need the strength of the field, the prize-fund scale, the draw rate, and the reasonableness of the schedule. The draw rate is the most sensitive indicator: when the classical draw rate exceeds sixty percent, it signals eroding commercial value, because spectators pay to see decisions, not caution. Without an event name, I cannot place it within the championship cycle. Without a prize-fund figure, I cannot assess an organiser's sustainability.

The fourth layer is the competitive landscape. This is where the modern chess story is most fascinating, and also where it is easiest to fabricate. After Magnus Carlsen announced in 2026 that he would not defend his title, chess entered what I call the separation of the crown from the rating. Ding Liren took the world title in 2026, and then Dommaraju Gukesh — at eighteen — took it from him in 2026 in Singapore, becoming the youngest world champion in history. Yet on the rating list, Carlsen remains number one, with a historic peak of 2,882 set in May 2026.

That is a paradox unique to chess: the strongest player is not necessarily the reigning one. To analyse this landscape, I need a four-tier map: the throne tier, the 2,700-plus challenger tier, the rising-star tier, and the reserve tier. I need to track the Indian wave with Gukesh, Rameshbabu Praggnanandhaa, Arjun Erigaisi and Nihal Sarin; the squeezed generation born in the 1990s with Ian Nepomniachtchi, Fabiano Caruana and Ding Liren; and the longevity of veterans such as Hikaru Nakamura. But if the entity list is empty, every generational comparison is an imposition from memory, not analysis.

The fifth layer is rules and governance. This is the layer that has made chess hottest over the past decade. The affair between Carlsen and Hans Niemann in 2026 — when Carlsen implied an accusation of cheating after their game at the Sinquefield Cup, leading to a lawsuit settled in August 2026 — turned anti-cheating into the sport's central topic. The tiebreak format is also a maze: rapid, then blitz, then Armageddon, where White gets five minutes and Black four but only needs a draw to win. The Sofia Rules forbid draw offers before move thirty. Federation-transfer rules allow a player to change representing country after a waiting period.

Each of those provisions has a direct effect on results at the board. A player prepares for Armageddon in a completely different way from how they prepare for classical chess. A player who has just transferred federation may lose the right to enter certain events during the waiting period. Without a named event, rule, or governing body, I cannot assess anything at this layer.

The sixth layer is risk. Here I must be blunt: the greatest risk of an empty analysis is not a chess risk. It is an information-integrity risk. When an empty template still demands that one "identify entities from the information points above", it creates a structural pressure forcing the writer to import their own existing knowledge and present it as though it were derived from the source article. That is fabricated attribution, and it is dangerous because it turns a data gap into an apparently confident claim.

At this layer, I must distinguish two states that every data system can confuse: "no risk found" and "no data with which to find risk". These two states look identical if one looks only at an empty cell. But they are worlds apart. The first is a conclusion. The second is a failure of process.

The seventh layer is public narrative. Chess has a very seductive set of story labels: the prodigy's debut, the new king, the end of a dynasty, the revenge arc, the cheating scandal, the breakthrough of women. Each of those labels can send search volume soaring for a few days. But a narrative is only sustainable when it is underpinned by data. A "new king" story built on three wins is fragile; the same story built on twenty games with a steady ACPL below twenty-five is worth analysing. The heat cycle of a narrative passes through four stages: germination, acceleration, climax, and backlash. Without subjects and claims, I cannot determine which stage I am in.

The eighth and final layer is industry transmission. Chess runs along a current: upstream is youth training and the talent supply, midstream is tournaments and platforms, downstream is content, commerce and derivative markets. An upstream phenomenon — the Indian wave, for instance — can reshape the entire downstream within five years. The rise of online platforms such as Chess.com and Lichess has restructured how audiences reach chess, creating a streamer class and changing the flow of sponsorship. Downstream, there is an entire market in equipment, books, courses and intellectual-property licensing.

Not a single one of those actors appeared in my empty template. What is striking is that the simultaneous absence of both competitive and industry entities reinforces a single hypothesis: this is a failure at the extraction stage, not eight independent content gaps.

And this is where I must confront the greatest temptation of my profession.

I have enough knowledge to fill that template. I know Carlsen has held the 2,882 peak since May 2026. I know Gukesh became world champion at eighteen. I know the Niemann affair ended in a settlement in August 2026. I could write an analysis that looks deeply professional, full of numbers and names, and no one — not even an editor — could detect that not a single line of it came from the data of the game I was asked to analyse.

That is precisely what I refuse to do.

The number is asceticism: one must give up ease before one can see the truth. The ease here is the ease of telling a good story instead of admitting that there is nothing to tell. And I am grateful that I have enough scars to recognise it.

But wait. Here I must be careful with myself, because data purity has its own trap. I once called everything "data that has not been read" — until a number fooled me, and I swore never to let the market surprise me again. Since then, I have been prone to the opposite trap: treating every data gap as evidence of an impending failure.

The truth is that correlation is not causation. An empty template does not prove the source article never existed. Nor does it prove the tournament had a problem. It proves only one thing: either the extraction process failed, or the source article genuinely contained no technical data, or some link between the source and me fell silent. Those three possibilities cannot be distinguished by looking only at the empty output.

This is the blind spot of the entire sports-data analytics system. We design highly sophisticated analysis templates — eight layers, dozens of metrics — but we forget the most basic question of all: did the data actually arrive? A full but wrong template is more dangerous than an empty one, because the empty one incriminates itself, while the full and wrong one wears the mask of professionalism and walks straight into the reader's decisions.

I have seen this at a much larger scale. When the pandemic halted Asian leagues in March 2026, I collected Guangzhou Evergrande's sprint data across three seasons, 2026-2026, and compared it against their post-lockdown recovery tests. Average sprint speed fell 38.2 percent; players over thirty recovered 27.3 percent more slowly. I predicted they would collapse in the run-in. No one believed me. The pandemic did not kill Evergrande; it only signed the sentence that age had already written. But what matters is this: if that recovery test had returned empty data, I would have had nothing to say. And the greatest temptation then would have been to tell a story about the decline of a dynasty — a story that might have been true, but was not the story my data told.

That is the boundary I have to hold every day, and it applies to chess exactly as it does to football. I am not permitted to call a tournament "chaotic" when I have no standings. I am not permitted to call a player "on the rise" when I have no rating curve for them. I am not permitted to write that "chess is changing" when I have no data on sponsorship flows or viewership. Each such sentence might be true, but true in an irresponsible way is just another kind of false.

And there is a subtler trap, one I call the false three-source trap. I take pride in always cross-checking three sources. But if all three point back to one underlying figure — the same press release, the same database, the same translation — then I do not have three sources, I have one source multiplied by three. This is a mistake I nearly made many times when checking ratings. FIDE, an aggregator site, and a game database may all display the same figure, but if all three took it from the same publication, their agreement is fake. The real discipline is to trace the source deeper: where was this number born, who measured it first, and has anyone independently verified it.

At a still deeper level, I must remember that my long-term vision is not allowed to forget short-term blockages. A player can have a beautiful age curve across seven seasons, but a single injury, a coaching change, or a personal shock within three months can reverse everything. So I always keep a separate short-term checklist alongside the long-term analysis: fitness, a crowded calendar, and the small psychological signals that no rating list records.

But all of that only means something when I can begin. And this time, I could not begin.

The empty template taught me something twenty years of experience had not taught me in full: the most important discipline of a data journalist is not the discipline of writing, but the discipline of refusing to write. Refusing to publish an analysis I know to be imposed. Refusing to fill an empty cell with a plausible-sounding name. Refusing to turn "no data" into "no risk".

And I asked myself: if the person who opened that file had not been me, but a young writer on deadline, what would have happened? The answer chills me. They would have written a beautiful analysis. It would have had enough big names, enough impressive numbers, enough gripping story. And it would have been entirely unrelated to the game they were asked to analyse. No one would have noticed. No one would have checked. And it would have entered the reader's memory as fact.

That is why I am writing this piece. Not to tell the story of a broken file — broken files can be fixed. But to point out that the information market is the only place where people pay for numbers that have not been verified, and in chess, where every move can be machine-verified, letting unverified information circulate is an unforgivable waste.

So what is the signal for the next cycle?

I do not need a revolution. I need one small data field: extraction status. Three values only — success, partial, empty. When the status is "empty", the system must block the analysis layer itself and return an empty report, rather than an eight-layer template waiting to be filled. When the status is "partial", it must state clearly which parts are missing. And any figure cited afterwards must carry its provenance — FIDE's official rating list, the 2700chess live board, the ChessBase or TWIC game database, the platform statistics of Chess.com and Lichess — while figures that cannot be verified must be clearly marked "pending verification".

That is a small technical change. But it is a large ethical one, because it turns silence into a signal and honesty into a format.

I closed the file at 23:12. The third cup of tea had gone cold. And I did the thing I had never done in twenty years: I wrote a report saying that I could not write the report.

People will say I am slow. They have said it throughout my career — that I am the slowest journalist on the desk. Let them keep saying it. But I know one thing that an empty template taught me more clearly than any full dataset ever could: the line between analysis and fiction does not lie in how much one can write, but in whether one dares to stop when the numbers have not arrived.

And in a sport where every move is recorded, where the truth can be reconstructed to the second, daring to say "I do not know" before a data gap may be the most accurate move I have ever made.

I will wait for the data. And when it comes, I will read it before I write about it.