The Empty Board: When the Data Stream Breaks in the Chess World

THE EMPTY BOARD: WHEN THE DATA STREAM BREAKS IN THE CHESS WORLD

1. A white night and an empty record

On the screen, the record appeared with exactly one kind of character: a dash. No title. No source. No player name. No date. The "information points" field was empty. The "core viewpoints" field was empty. And the "entities involved" field carried a sentence pointing into the void: identify from the information points above — while above there was nothing at all.

I sat for a long time in front of that record, in a room looking out over the Neva. Early snow fell thin as chalk dust. In forty years of following the chess world, I had lost sleep over a rook move missed on the sixtieth move of an endgame; I had rewritten an opening paragraph three times because I could not find the right verb. But never had I met a chess story that vanished before it could be told.

That emptiness was not silent. It was loud in its own way: loud with the assumptions that both readers and writers fill in automatically. With no player name, memory immediately supplies the most familiar name. With no tournament, it supplies the tournament being talked about most. With no date, it supplies today. That is the moment a sports report stops being a report and starts being a novel.

And that is why I chose to write about it. In a chess world believed to be drowning in data, the most worthwhile question lies on the opposite side: what happens when the data stream breaks?

2. How many lines of data does the chess world run on?

Outsiders often imagine chess is the most thoroughly recorded sport on the planet. That image is half right. A chess game leaves behind a perfect record: every move, every interval of thought, every tap of the clock. No other sport preserves its action in such detail. But that record speaks only of the pieces, never of the person sitting behind them.

The chess world's information infrastructure rests on a few rather thin pillars. The Elo rating list of the International Chess Federation (FIDE) is published on the first day of each month — a pulse that is slow, steady and predictable. Aggregators such as 2700chess track live ratings game by game, filling the gap between two publication dates. Game databases such as ChessBase store millions of games in standard notation. The Week in Chess, maintained weekly by Mark Crowther for nearly three decades, is the backbone for a great many chess journalists. Alongside them sit two enormous online platforms and more than two hundred national federations affiliated with FIDE.

Based on my own experience following matches, those pillars almost never collapse at the same time. But they collapse individually, quietly. A news page behind a paywall. A page that renders only after the browser finishes running its scripts. A live stream with sound but no transcript. A placeholder record entered into a system that nobody returns to replace.

What every one of those failures has in common: it does not produce a visible gap. It produces a gap filled with whatever was already at hand.

One detail is worth pausing on. In a chess report, the four basic anchors are name, date, tournament and result. A competent extraction process running over a professional chess article will almost always find at least one proper noun — because chess journalism is written about people. When the output contains not a single entity, the likeliest explanation is not that the article had no one in it, but that the article never reached the system at all.

That conclusion is procedural. Its consequence, though, is sporting: somewhere out there is a chess story nobody is tracking.

3. The anatomy of a chess report

A decent chess report needs at least four anchors: a name, a date, a tournament, a result. Lose one and the piece tilts. Lose two and it collapses. Lose all four and what remains is a sheet of paper with a headline.

Behind those four anchors sit several layers of analysis, each demanding a different raw material.

The game layer is the sturdiest. It needs only the score. From that one can compute ACPL — average centipawn loss per move, where lower is better — and the engine match rate. ACPL carries a trap: the same value means entirely different things in a locked position and in a wide-open one. It is a measure, not a verdict. A player with a low ACPL in a simple position has not played better than a player with a higher ACPL in a chaotic fight.

The human layer is far more fragile. It needs Elo, age, head-to-head history, and knowledge of who is whose bogey opponent. Elo is a single number carrying an enormous semantic load, and precisely for that reason it is the most misread. Classical Elo, rapid Elo and blitz Elo for the same person can differ by hundreds of points; using one figure to speak for all three is a common error. Tournament performance rating is another matter entirely: it measures what is happening now, not what has accumulated. And on a small sample, performance rating is wildly unstable — it can leap after three wins against strong opponents and crater after two draws.

The tournament layer demands structure. This is where the qualification path decides everything: a place from the World Cup, a place from the Grand Swiss, a place by rating, a place from the Grand Chess Tour, and a wild card. The same player with the same form tells a completely different story depending on which door he walks through. One player enters the World Cup because a knockout format offers a short path; another chooses the Grand Swiss because a nine-round Swiss rewards consistency; a third simply waits, harvesting points at open events to take a rating place. Format shapes psychology too. A round-robin demands endurance; a knockout demands nerve; and a single Armageddon game can erase weeks of accumulation through one drawing decision and one clock offset.

The first three layers can be checked against data. The remaining layers cannot, and that is precisely where a chess report is actually written.

The landscape layer demands context measured in years: who holds the throne, which generation is rising, which national wave is arriving, and whether a golden generation has enough depth to sustain pressure for a decade or is only a short burst. The rules and governance layer demands primary documents — anti-cheating regulations, tiebreak procedures, eligibility and federation-transfer conditions. The risk layer demands data on congested calendars, on mental health, on the burnout risk facing teenage players pushed through too many events in one season. The collective-memory layer demands time — it only forms after the result has settled. The industry transmission layer demands money: prize funds, sponsors, viewership, and the migration from board to screen.

None of those layers lives inside the score. So a record complete in every move can still be empty in every meaningful respect. One can look up precisely the thirty-first move of a game and know very little about why that game mattered.

This is the central paradox of writing about chess. Our instruments measure thousandths of a second of a player's thinking, yet cannot measure what is happening inside that player's head.

4. The blind spot of collective memory

There is an almost default belief in the chess world: because every game is recorded, every story is recorded. That belief is wrong in one very concrete place.

Compare football. A football club leaves traces of minutes played, distance covered, injury history, transfer fees, wage bills, release clauses. Chess leaves traces of moves. Nothing else. No injury report, no wage bill, no release clause. The closest thing chess has to a transfer market is national federation transfers, and those are usually handled quietly in administrative procedures that sports journalism rarely touches.

The transfer market is where people sell the future to buy hope. In chess that market exists but has no listed price. And when there is no listed price, people assume there are no deals worth reporting.

The result is a structural blind spot. The three biggest fault lines in this sport — anti-cheating, fairness of tiebreak, and eligibility — are procedural rather than narrative. They produce no goals, no moments, and therefore no headlines. The 2026 affair between Magnus Carlsen and Hans Niemann was the rare exception that pushed the anti-cheating story into the mainstream press, and it managed that only because two big names stood at either end of it. Without those two names, the same affair would have stayed inside technical bulletins.

There is a second blind spot, less often mentioned. People assume a data-dense sport cannot lose a story. But disappearance in the information age is rarely about being silenced. It is usually about infrastructure. Three or four databases carry almost the entire memory of a sport; let just one of them stop updating properly and a region of memory vanishes before anyone notices. No bell rings when a story goes untold.

I once sat in empty stands during the pandemic and understood that the sound of an empty stadium is not silence — it is the sound of thousands of conversations being unplugged at once. The online chess world never has a moment like that. It is always loud. But loud is not the same as remembered.

When a stadium has no ball, memory starts scoring. In chess, memory scores through the games that are preserved, while the people fade.

5. The final move

An empty record is not a disaster. It is a signal, and that signal is far cheaper than a wrong analysis.

What I took from that night is not a call for more data. The chess world already has enough moves. What is missing is context: a name, a date, a reason. Any system handling chess news should begin with the three cheapest questions — is there a title, is there at least one information point, is there a date — before spending any resource on interpretation. If the answers are no, the right action is to stop and register the gap, not to fill it with the most familiar story available.

The greatest risk in this profession has never been misreading a game. The greatest risk is letting a story pass by without anyone knowing it ever existed.

The board can be empty. The person sitting in front of it cannot.