Chess Needs a Data Standard, Not More Empty Analysis Sheets

On a December night in 2026, in Singapore, Ding Liren sat over a rook endgame that the engine assessed as balanced. Ten minutes later he left the board as a former world champion, and Gukesh D — eighteen years old — became the youngest world champion in chess history, winning the match 7.5–6.5. I replayed that game four times in one night in Chengdu, with two screens open side by side: one playing the recording, one showing my own data sheet. On the second screen, almost every cell was empty. I could not record a single figure solid enough to cite. What kept me up until dawn was that emptiness, more than any blunder.

Chess is the sport with the oldest measurement system in the adversarial category. The Elo rating devised by Arpad Elo was adopted by FIDE in 2026, and it remains the common currency of the chess world. Behind it sits a second data layer: average centipawn loss (ACPL), the rate of agreement with the engine's top move, the number of novelties never previously recorded in a database, the time remaining on the clock when entering the endgame, and the draw rate by format.

In the Vietnamese-language market, that second layer barely exists. Most chess content readers encounter is translated from foreign headlines, garnished with a few emotional adjectives: inspired, resilient, declining. Nobody checks what those adjectives rest on. When a data standard is missing, a writer has only two options: invent a plausible-sounding story, or state plainly that the data is insufficient. Only one of those options survives over time.

A top-level game can be compressed into a few figures: lower ACPL is better, a higher engine match rate is better, more novelties suggest deeper opening preparation. That convenience creates the first trap. The ACPL of a whole game says nothing about the opening, the middlegame, or the endgame. A player can perform flawlessly for twenty moves and then collapse over the final ten because the clock shows forty seconds — and the aggregate figure still looks fine. Seeing that requires splitting the data by phase and by remaining time. Based on my experience following matches since 2026, I began doing this systematically in 2026, after realizing that the summary tables major outlets publish usually hide precisely the decisive detail.

Engines do not manufacture class; they strip the mask off those pretending to calculate. Everything on the board is data waiting for a reader, if the reader is willing to sit down.

Elo is the most quoted and most misunderstood figure in the sport. Magnus Carlsen reached 2882 in May 2026, the highest officially recorded rating ever. But Elo is a relative system, recalculated after every event, and it answers how often a player beats another player, not how well that player performed today. Elo does not lie; the people reading Elo do. So I always place three columns side by side: classical Elo, live rating during the event, and actual performance in that specific event. Arjun Erigaisi crossed 2800 in late 2026, a rare milestone, but it only means something beside his own rating a year earlier and beside players of his age cohort.

Then comes the most neglected part: the qualification path. A place in the Candidates Tournament can come from the World Cup, from the Grand Swiss, from Grand Chess Tour points, from an average-rating spot, or from a wild card. Each path produces a different kind of player. Those who qualify through the World Cup are used to knockout series and rapid tiebreaks running past midnight. Those who qualify through the Grand Swiss are used to the pressure of needing a win in the final round against an opponent on the same score. At the 2026 Candidates in Toronto, Gukesh D earned the right to challenge for the title at seventeen, winning by playing steadily across fourteen rounds of a round-robin where the draw rate is always high and every half point is expensive.

The Armageddon format has to be read separately. White gets more time but must win; Black only needs a draw. That inverts ordinary psychology entirely. A player holding Black in an Armageddon game is not playing to win — they are playing not to lose. Feeding that result into an ordinary win-loss table corrupts the data at the input stage.

The national landscape also has to be read by model, not by ranking. India is in a phase any chess nation would envy. At the 2026 Chess Olympiad in Budapest, India won gold in both the open and women's sections, something that had never happened in the country's history. Their roster runs from Gukesh, Rameshbabu Praggnanandhaa and Arjun Erigaisi to Rameshbabu Vaishali. China took a different route: one player holding the world title for two years, plus an exceptionally strong women's system, with Ju Wenjun successfully defending the women's world championship in 2026. Uzbekistan shows a third model: structured training under limited resources, with Nodirbek Abdusattorov winning the World Rapid Championship in 2026 at seventeen. These three models do not replace one another, which is why a national ranking table becomes meaningless if it only looks at the number-one player.

Even the rulebook generates data, and that data is being wasted. In September 2026, after losing to Hans Niemann at the Sinquefield Cup, Magnus Carlsen withdrew from the event and publicly voiced suspicion. In October 2026, the world's largest online chess platform published a report of more than seventy pages concluding that Niemann likely cheated in over one hundred online games. FIDE, after its own investigation, concluded he did not cheat in over-the-board play. Niemann filed a lawsuit seeking one hundred million USD; most of his claims were dismissed by a US court in 2026.

What stands out in that sequence is not any individual. It is that a private platform and an international federation reached two different conclusions about the same person, based on two different evidentiary standards, and both were valid within their own scope. The chess community still has no mechanism to reconcile those two rulings.

In December 2026, at the World Rapid Championship in New York, Carlsen was penalized for wearing jeans, refused to change, and withdrew from the rapid event. FIDE subsequently relaxed the dress code, and he returned to win the blitz title. A debate about dress regulations overshadowed the entire competitive content of the tournament for days. For someone working with data, that is a cost that cannot be entered into any table.

There is one further layer of analysis that chess journalism barely touches: risk. The career risk facing an eighteen-year-old who has just won the world title lies in a packed calendar and expectations that spike after every trophy. Financial risk lies in a prize structure distributed very unevenly between the elite and everyone else. Psychological risk lies in players who must compete continuously in front of cameras and on online platforms, where every game is recorded and every mistake is counted. Systemic risk is newer: when every elite player trains against the same generation of engines, stylistic differences tend to narrow, and the draw rate at elite events rises cycle by cycle. That is a measurable trend, and it is being ignored because it generates no attractive headline.

On the public side, the expectation narrative always outruns the data. A young player who wins three games in a row is called a title contender; three months later, if form dips, that same narrative turns against them. The gap between expectation and reality can only be measured with a large enough sample — usually several hundred games, meaning several years of competition.

And here I have to argue against myself. The data-verification school has a trap: when there is no data, people easily turn silence into a stance. Every month I receive a few analysis sheets from younger colleagues, and roughly half of them contain nothing but carefully annotated empty cells: insufficient sample, unverified, no source. That caution deserves respect, but it only has value when paired with a plan to fill the gap. Writing "insufficient data" and stopping there is a polite form of delay, not a professional standard.

The second trap lies on the opposite side, and it is more dangerous. When data is empty, the content market fills the space with emotion, and emotion is always available. That is why stories about resilience, class and decline outlive any statistical table. They require no proof, and precisely for that reason they are useless to readers who want to understand what actually happened on the board. I no longer believe in miraculous moves; I believe only in conversion rates for advantages.

A chess data standard will not come from a federation or a technology platform. It will come from writers willing to state clearly what they are measuring, with which tool, and where the gaps lie. If a chess article does not tell you which figures were verified and which were not, then you are reading an empty spreadsheet, merely better presented.