Gukesh, the Indian Wave, and the Blind Spots of Chess Data

On the night of December 12, 2026, I sat in front of a screen in Da Nang with Vietnam's clock already past midnight. Game 14 of the World Chess Championship in Singapore drifted into a rook endgame. Ding Liren had the white pieces, and the position was so balanced that the engine evaluation bar read dead level. On move 55 he chose a move that, seconds later, sent the bar swinging to the other colour. Gukesh Dommaraju, eighteen years old, set the piece down, his face unchanged. The final score was 7.5–6.5, and he became the youngest world champion in the history of the game.

What I remember does not sit in the scoreline. It sits in the silence right after the machine changed colour — a silence no algorithm can explain, about why a former champion chose that move, after fourteen tense games and months of media picking apart his every step.

Elite chess is shifting on its axis faster than at any point in two decades. Magnus Carlsen has declined to defend the classical title since 2026, pouring his energy into rapid, blitz and commercial events. The classical crown passed through Ding Liren's hands and on to Gukesh within two years. India suddenly unveiled an entire generation: Gukesh, Rameshbabu Praggnanandhaa, Arjun Erigaisi, Nihal Sarin.

In Vietnam, the gap to that group remains visible. Le Quang Liem, our number one for more than a decade, hovers around 2700 Elo and is strongest in rapid and blitz. Nguyen Ngoc Truong Son is still a pillar of the national team. SEA Games golds keep arriving, but the regional stage and the world stage are two different worlds.

FIDE publishes three separate rating lists: classical, rapid, blitz. Specialists follow live ratings through 2700chess. Most Vietnamese fans, meanwhile, meet chess through score tables and short bulletins, where the result is treated as the entire story.

Here I want to stop at something that looks harmless: the metrics.

The ACPL metric — average centipawn loss per move — has become the default yardstick for chess quality. A player with a low ACPL is considered accurate; a game with a high ACPL is called poor. But ACPL only states the deviation from the engine's choice. It does not say why a player deliberately plays a less accurate move to push an opponent into a trap that exists in no book.

Based on my experience following matches across many seasons, the same ACPL level can come from two entirely different games: one in which a player is dragged into a position the opponent prepared, and one in which that player willingly concedes ground in the middlegame in exchange for a technically winning endgame. The analysis table does not distinguish between the two.

Then there is engine match rate — the percentage of moves matching the machine's first choice. This figure lands in bulletins as a measure of class. In reality it reflects opening preparation more than overall strength. A player with a strong backroom team and a private database will post a very high match rate over the first twenty moves, the stretch the engine understands best, then fall away in the stretch where human instinct decides.

I call that the blind spot of chess data. It is not a shortage of numbers. It is an excess of numbers, enough that readers believe they have understood the whole game.

The Niemann and Carlsen affair of 2026 is the clearest example. At the Sinquefield Cup in St. Louis, Carlsen withdrew after losing to Hans Niemann, and the story exploded into the biggest controversy in modern chess. Anti-cheating statistical models were dissected in public. The problem is that those models were designed to screen hundreds of thousands of online games and flag anomalies warranting further investigation. They were never designed to deliver a verdict on one person in one specific game. But that is how they were read.

When a statistical indicator leaves the context it was built for and enters the court of public opinion, it stops being a tool. It becomes evidence, and evidence requires an entirely different standard.

Meanwhile another kind of data has quietly taken up space in chess coverage: the opening heat map. Deep red squares on the board, frequency counts for every move, colour-coded by win rate. They are attractive, easy to read, and make viewers feel the game has been decoded. But a heat map only retells the past. It does not say which move will appear in this game, because the players themselves change their choices to avoid those deep red squares.

I once sat counting every beat at a domestic junior event, where a coach opened a heat map and told his student that a certain move wins 58 percent of the time. The fifteen-year-old nodded. Nobody told him that the figure was computed on a dataset of players at different levels, in events with different time controls.

With a polymath's eye, I look at a chessboard the way I look at a hurdles lane: same distance, same finish line, but an entirely different stride pattern. Pulling back the curtain on speed in another sport, I met a whole generation running; in chess, that generation runs on a clock.

There is one more layer few mention. Women's chess events are promoted ever more loudly, but most of the sponsorship flows in as corporate social responsibility packages rather than because organisers believe in the commercial pull of the event. The paradox is that female players appear more often in the media, while prize structures and schedules are still designed by old habit. Look at the sponsorship, and you would think everything has changed. Look at the three-year contracts and the number of rounds, and it has not.

In Vietnam, the data story runs to a different beat. We do not lack good players; we lack a record-keeping system long enough and transparent enough to reveal trends. A young player who wins three events in a row is immediately called a generational talent. What nobody does is place those three events side by side, normalise for time control and opponent strength, and only then draw a conclusion. People talk about speed, but I find time — enough time for a trend to reveal itself rather than be forced into view.

The counterintuitive angle sits here: the more data there is, the less chess fans understand the game.

The reason is not that the data is wrong. It is that data is being used as a substitute for observation. Once everything has an indicator, people no longer have the patience to look at the board long enough to see a player changing rhythm, deliberately slowing down, dragging an opponent into the time zone he wants. Those things do not appear on the evaluation bar. They only appear to someone actually sitting and counting.

The same thing happened in other sports. Basketball has shot heat maps, football has movement heat maps, and both were once used to draw conclusions about a player while ignoring his real role in the tactical system. Chess is walking exactly that road, only a few years behind.

The most uncomfortable point is this: the Indian wave is not a miracle of a perfect training system. It is largely the product of digital infrastructure — millions of online games, free platforms, and a generation that grew up with them. Look at the results, and people praise the academies. Look at the underlying data, and it is a story about cheap internet bandwidth rather than curricula.

From the sandpit runway to the virtual arena, sport's homeland has no borders, but its language is always mistranslated. Vietnamese chess will not move forward because of more metrics. It will move forward when someone sits down, counts every move, and tells the truth about where the evaluation bar changed colour — not where it is convenient to say it did.