When the Chess World Analyses on Faith: From an Empty Report to Tashkent 2026

At 2 a.m. on 13 December 2026, I replayed game 14 of the World Chess Championship in Singapore for the fourth time that night. I was not rewinding it because I failed to understand the game. I wanted to watch the board change hands with my own eyes, not through an engine evaluation line. In that game, a single rook retreat to f2 by Ding Liren was enough to bring everything down, and Gukesh D, eighteen years old, became the youngest world champion in the history of the game.

Three weeks later, a content partner sent me an analysis package on that very match. Twelve pages. Eight major sections. Neatly ruled tables, bold headings, a tidy appendix. I read it from the first page to the last and came away with exactly one kind of information: insufficient data to assess.

No player names. No game number. No opening code. Not a single move recorded. What I held in my hands was a flawless skeleton, carefully packaged, and completely hollow.

I kept it instead of throwing it in the bin, because it is the cleanest specimen of a disease spreading through chess analysis.

A sport rich in data and poor in information

Chess has never attracted this much attention. After the wave set off by the 2026 television series, the number of online players soared, streaming platforms turned every game into an event watched by hundreds of thousands at once, and prize money at the elite level followed. In Astana in 2026, Ding Liren beat Ian Nepomniachtchi in a rapid tiebreak after a 7-7 classical draw. In Singapore in December 2026, Gukesh D defeated Ding Liren 7.5-6.5 in a match with a total prize fund of 2.5 million US dollars. Magnus Carlsen, a five-time champion, left the throne after 2026 and threw his weight behind the Freestyle Chess Grand Slam Tour, launched in 2026 under Fischer Random rules.

India is the epicentre of this period. Arjun Erigaisi crossed 2800 Elo in December 2026, becoming the second Indian after Viswanathan Anand to do so. At the 2026 Chess Olympiad in Budapest, India won gold in both the open and the women's team events, with the open section drawing close to two hundred teams. A generation of players born after 2026 grew up alongside analysis engines, opening databases, and private training rooms in Chennai, Bangalore and Hyderabad.

Vietnam runs on a different track. Le Quang Liem and Nguyen Ngoc Truong Son carried Vietnamese chess for more than a decade. Le Quang Liem once entered the world top 20 and won the World Blitz Championship in 2026, a rare milestone for Southeast Asian chess at the highest level. Behind those two names, the talent pipeline remains thin, and the number of international events strong enough to build rating each year remains small. A young Vietnamese player aiming to pass 2600 often has to fly abroad ten times a year, while an Indian of the same age can do it within a three-hundred-kilometre radius.

Between those two realities, a new profession exploded: commentary and analysis. The number of podcasts, newsletters, columns and game-summary videos has grown faster than the number of trained arbiters. That is when hollow skeletons began to be mass-produced.

Engine move-match rate measures obedience, not understanding

The two most common metrics in chess reports today are ACPL, the average centipawn loss per move, and the match rate against the engine's first choice. They are useful. They also carry a fatal blind spot.

Inside a forced sequence, a weak player still posts a very high match rate, simply because only one legal move exists. The reverse also holds: a player with deep understanding can deviate from the engine several times in a game because he is pursuing a long-term plan, and his match rate drops below that of a weaker but safer opponent.

I once compared two reports on the same game. The first concluded that player A played more accurately because his match rate was three percentage points higher. The second, written by a coach in Chennai, pointed out that player A had missed exactly the plan player B spotted, and that the match rate did not reflect this at all. The second was right. The first was beautiful. The first was read twenty times more.

What the engine cannot tell you about game 14

The engine will attach a red question mark and a negative evaluation to Ding Liren's rook retreat to f2. It tells you the move was wrong. It does not tell you why a man who had just played fourteen tense games in three weeks, who had just come through a period of mental crisis, chose that move at the exact moment when everything was still salvageable. Gukesh replied with 56.Re3, trapped his opponent's strongest piece in the corner, and Ding resigned.

Three symbols compress the entire ending of a title match. The real cause sits in the earpiece, in the clock, in the breathing, in the fact that Ding had to hold himself together for fourteen games against an opponent twelve years younger.

Based on my experience following these matches, I drew one rule: the engine answers which move is best, only a human can answer why the best move never appeared. A report that contains only the first half is a report not yet written.

In 2026, when the pandemic closed every tournament, I built my own podcast series to rewatch twenty games from the 2026 to 2026 World Cups. In that year of empty stands, I found the tactical diagram lying quietly under the seat instead of the roar of the crowd. And in that silence, I found a purely chess-based attacking pattern that had appeared long before anyone named it: young players pushing the queenside pawn to open a file, instead of waiting for the opponent to collapse. Three years later, reports began calling that phenomenon by a new term, with charts. Nobody mentioned that it had been sitting quietly inside old games for years.

An empty skeleton is the product of a structure, not the accident of an individual

When a newsroom sets a quota of eight sections per analysis, the writer fills all eight. When there is no data, he still fills all eight, except the content inside consists of sentences like insufficient data to assess. The frightening part is not laziness. It is that the skeleton remains formally valid, still passes review, still gets published, and still gets read as professional analysis.

I tested this on myself. For years I kept a bad habit: starting to write before finishing the data, because I trusted my reflexes. My reflexes were usually right. When they were wrong, they were wrong in the most expensive way, wrong exactly where nobody checks again.

What I learned was not to abandon the reflex. What I learned was to install a gate: if an analysis does not contain at least three concrete pieces of information, it does not get published. A proper data pipeline must stop itself when the input is empty, instead of emitting a valid-looking container. It sounds simple. In practice almost nobody does it, because an empty report still pulls half the traffic of a real one, at less than a tenth of the effort.

Seven minimum requirements for an analysis to mean anything

Player name. Event name. Round. Opening name or code. The move number of the turning point. And at least one of the remaining three: average centipawn loss, engine match rate, or time remaining on the clock at the decisive moment.

Take those seven criteria and hold them against ten trending chess analyses online this week. I bet no more than two pass. And of those two, the odds are one was written by someone who sat in the playing hall, or by someone who rewound the footage until four in the morning.

The data gap in women's chess

At the 2026 Olympiad in Budapest, India's women's team won gold, a feat covered widely for about forty-eight hours. Four days after the event ended, the number of detailed analyses of the women's team's games was about ten times smaller than the number for the men's team. I counted. I count because this is the kind of counting I have done for thirty years, and because I am tired of having to explain that counting is not an attitude, counting is data.

Koneru Humpy, Dronavalli Harika, and on the Vietnamese side Pham Le Thao Nguyen, all played games that deserve to be taken apart move by move. The analysis devoted to them is far thinner. This gap does not come from audiences not caring. It comes from newsrooms not allocating people, writers not having time, and so the empty skeleton wins again.

The economics of a serious analysis

A serious game analysis takes roughly six to eight hours: rewinding footage, checking with an engine, cross-referencing the opening database, writing, editing. An eight-section skeleton with mostly empty cells takes about forty minutes. Both are paid per piece at most newsrooms. The effort gap is tenfold, the fee gap is close to zero.

I am not writing this to complain. I am writing it to say that the quality of chess analysis is not a moral question about writers. It is a question of cost structure. If you want good analysis, you have to pay for rewinding time. There is no other way, and every call for professional responsibility without a budget attached is empty talk.

Chess's transfer market runs like football's, minus the ticket sellers

Chess has no transfer window in the football sense, but it has an equivalent: federation transfers. Alireza Firouzja left Iran for France in 2026. Every time this happens, a chain of noise is produced: who negotiated, how long it took, what the fee was, how long the waiting period is. Agents appear, make statements, apply pressure. A story about a federation transfer is usually three times longer than a story about a good game of chess.

I read that as a sign of an immature market. When an industry lacks data standards, noise fills the vacuum. And noise is far cheaper than analysis: a post about a possible federation transfer takes thirty seconds, a decent game analysis takes three days.

This has direct consequences for smaller chess nations, Vietnam among them. When resources are sucked into stories of movement, what is left for training and for the domestic tournament system grows thinner. A fifteen-year-old talent in Hanoi needs three things: opponents of the same level, events of sufficient standard, and a coach who can read the position. She does not need another story about someone negotiating a federation switch.

A story about the writer, not about the data

In 2026, at a press conference in Delhi, I was the only woman reporter in the room. I asked a question about the opening choice of a key player, and a male colleague sneered: what would a woman know about defence. I put that sentence verbatim into my piece, along with three tables of data on that player's average move count in the period. The piece drew the highest readership of the week.

The question I asked did not look like a woman's question, but the answers they dodged did not look like a man's either. Ten years standing in the middle of arguments, and the fire lit by that question in 2026 is still burning. I tell this story not to boast. I tell it to say that chess analysis has a second layer of noise, outside the board: noise about who has the right to speak.

Where I might be wrong

I might be wrong in dismissing what audiences actually want. Most viewers do not need to know why a player missed a plan. They need to know who won, in how many moves, and who is currently the best. A short, tight analysis with three numbers serves the majority better than my long one. I have to hear that, even if it is not easy to hear.

I might also be wrong in underestimating the engine-raised generation. Gukesh, Erigaisi, Praggnanandhaa, players born after 2026, learned chess by conversing with machines from childhood. If extreme engine synchronisation produced the youngest world champion in history, then the move-match rate may not be a junk metric. It may be the right metric, merely being read the wrong way. I note the possibility and leave it open.

And I have to retell my own error. In 2026, working as a commentator at the World Cup, I mispronounced Denis Cheryshev's name three times in the first half. I got the player's name wrong, and then got the game right when the slow-motion replay appeared in front of my eyes. After that day I sat through footage of all thirty-two teams over two weeks, mostly to understand that getting a name wrong is a small error, while confidence without data is a large one. Mistakes are not frightening; bending down to pick the footage back up at 2 a.m. is what makes a forecaster.

There is one more possibility I do not rule out: that empty report was more honest than anything I have ever written. It said plainly that it had nothing. Many of my word-filled reports only talked around something I had not yet verified.

Two verifiable predictions

Before the 46th Chess Olympiad opens in Tashkent, I predict at least one major chess outlet will publish a game report containing a factual error, a wrong move name, a wrong move number, or a wrong player name, caused by an automated content process missing a verification step. My second prediction: at least one decisive game in the 2026 cycle will be settled in an endgame, with under five minutes on the clock, not in the opening.

If both predictions hold, the industry will have to choose one of two directions. Either standardise minimum data and accept that some pieces will not be written. Or keep producing beautiful skeletons, complete with eight sections, and hollow. Silent applause can kill emotion, but it cannot kill tactics. And an empty board never lies.