When the Analysis Returns Nothing: The Limits of Data and the Weight of Memory in Chess

On my computer screen in Saint Petersburg, a column of grey cells runs down the page, each one carrying the same two characters: N/A. No title. No source. No name. No date. The machine had read an article about chess and returned exactly nothing. I looked at that empty column for a long while. Outside, November snow fell thin over Nevsky Prospekt. Inside, my tea had gone cold. And I realised I had seen this scene before, not on a screen, but from the stands. On 1 July 2026, at Luzhniki, before 79,000 spectators, Russia were pinned back by Spain to the point of holding just 25 percent of possession. The scoreboard read 1-1. Read only the scoreboard and you would say nothing happened. But Igor Akinfeev, then 32, played like a steel wall, saving from Koke and Iago Aspas in the shootout to send Russia through 4-3. The scoreboard recorded two numbers. The story lived somewhere else. A zero on a scoreboard and a zero in a spreadsheet are not the same thing. One is a result. The other is silence. Confusing the two is the gravest error a sports writer can make. In 2026 I began my career as a chess player and tournament organiser, then moved into media. Twenty years later I still sit in the same chair, but the world around me has changed enormously. Chess has gone from a silent hall of competition to a content industry. Every major event generates thousands of articles, hundreds of machine-made analyses, tens of thousands of posts. Nobody reads them all. Nobody verifies them all. Within that flow, most content follows a familiar template. A young player wins a few games and it is called a new wave. A champion loses once and the reign is declared over. These stories are not entirely false, but they are pre-assembled, waiting for an event to fit into them. When no event arrives, people write anyway, changing only the names. The machine I am describing works on a simple logic. It reads an article, breaks it into information points — names, events, dates, numbers — then moves to deep analysis. When the first step returns an empty list, the second step faces two choices. One is to report an error. The other is to construct a plausible story. The second choice is far more dangerous, because it does not look like invention. It looks like expertise. I have seen something similar in my own trade. In 2026, when the pandemic suspended the Russian league, my newsroom in Saint Petersburg stood empty. No matches, no scores, nothing to report. Some colleagues began writing about matches that did not exist, hypothetical scenarios presented as news. I chose differently: I reopened the archive. In May 2026 I started the series A Stadium Without a Ball. The first instalment covered the 2026 European Championship final in Paris, where the Soviet Union beat Yugoslavia 2-1 and Viktor Ponedelnik scored in the 113th minute before 105,000 spectators. I sat alone in the empty newsroom, listening to wind move through vacant stands in old footage, and understood something. Football never needed the ball to exist. Memory was enough. I carried that lesson into chess, where it became far sharper. In chess, a draw does not mean nothing happened. This is the hardest thing for outsiders to grasp, and the thing automated analysis systems get wrong most often. In November 2026, in London, Magnus Carlsen and Fabiano Caruana contested the world championship. Twelve classical games. Twelve draws. On the results sheet, it was the longest run of zeros in the history of title matches. A machine reading only the results would conclude there was no information. Anyone in the playing hall those days knew the opposite. Caruana had prepared opening lines of extraordinary depth, in one game pushing Carlsen into a passive position past move thirty. Carlsen defended with lines the experts called old-fashioned but accurate to the square. Each draw was a compressed battle of wits. In the rapid tiebreak, Carlsen won 3-0. The final results sheet records one line: Carlsen champion. But those twelve draws were where the match truly happened. That is the nature of zero in chess. It does not mean data is missing. It means the data has not been read correctly. Sports science taught me to separate two things sports journalism routinely merges: a small sample and an empty one. A small sample still permits hypothesis, provided uncertainty is stated. An empty sample permits nothing. You cannot call a system strong or weak on zero games. You cannot assess a player across zero tournaments. You cannot conclude a trend exists when it has never been observed. In sport this is not an academic detail. It is the line between analysis and fortune-telling. When the extraction step returns an empty list, every tier behind it loses its anchor. Technical assessment needs a game with a scoresheet. Rating comparison needs a name and a publication date. Tournament tiering needs an event name and a format. Industry impact needs a dated event. Every tier needs at least one concrete fact. Without them, all conclusions float, even when written in a confident voice. This is where my trade and the data engineer's trade meet. Both face pressure to produce content. Both know an empty report looks worse than a report with content, even when that content is wrong. That pressure does not come from readers. It comes from the structure of the industry itself. The next tier in any chess analysis is player data. Elo, recent form, head-to-head record, age against the development curve. This data is published openly and updated monthly, so in theory it is the easiest to verify. Precisely because it is easy, people get lazy. I have seen many articles call a player rising without citing a single figure. Or cite an Elo from a list months out of date. In chess, a few months can mean three tournaments, dozens of games, and a significant shift in position among the leaders. A figure without a date is nearly meaningless. But in the empty analysis I am describing, even an undated figure does not exist. No player. No rating. No head-to-head. Nothing at all. And that does not merely spoil one section of a report. It collapses the whole structure behind it, which has seven tiers, each one capable of being filled with plausible guesswork. The tournament system tier is the same. Where does an event sit in the hierarchy? Qualifier or final? Round-robin or knockout? Chess has a more complex hierarchy than most sports. An open can have three hundred players and an unknown champion. A Candidates event has eight players and every game affects the right to contest the title. Same score, same number of games, entirely different meaning. Without identifying the event, a writer cannot establish the weight of any result. Then comes the competitive landscape. Who holds the throne, who is chasing, who is rising. This is the tier journalism loves most, because it permits storytelling, and the tier most easily inflated. For over a decade, elite chess revolved around one name. When that name withdrew from the title arena, an entire industry had to rewrite its templates. People began speaking of generational turnover, of the title belonging not to an individual but to a group. There is some truth in that framing, but it is also a way of dodging the hard question: who is actually strongest, and by how much. Even with complete data, judging the landscape is difficult. Without data, it is impossible. The rules and governance tier is the most sensitive, and the one where mistakes carry the heaviest consequences. A false cheating accusation can destroy a person's career. A wrong assessment of a format's fairness can shake faith in an entire event. Precisely because the consequences are heavy, this tier demands the highest evidentiary standard. Without a named incident, a named party, an official statement, there is nothing to analyse. And nothing to rule out either. This is the fine point I want to stress. Empty does not mean clean. If a tier cannot be checked, we cannot call it sound. We can only call it unchecked. That distinction matters in every industry, and especially in chess, where major scandals usually begin with a small detail nobody noticed. The remaining two tiers are risk and public narrative. I group them because they operate the same way. Both depend on timing. A risk without a date cannot be assessed. A narrative without a timeframe cannot be placed — nascent, accelerating, peaked, or receding. And both depend on something the empty analysis lacks: a subject. At the industry tier, chess today is a long chain from youth training to digital content. A decision at the top, a format change or a new federation policy, may take years to reach the bottom, but once it does it is very hard to reverse. This is why I dislike writing that looks only at match results. A single win says nothing about the flow of the sport. A change to qualification points, an adjustment to entry fees, a new time-control regulation — those shape the next ten years. During a transfer window, people talk about deals. In chess, transfer means something else: changing federation, changing coach, changing sponsor, changing competition system. These moves are quieter than a football contract, but their impact lasts far longer. The transfer market is where people sell the future to buy hope. I wrote that line about football, but it holds for chess too. The chess world has governance fractures everyone knows. Anti-cheating is one. The fairness of tiebreak and Armageddon formats is two. Federation transfer and neutral status is three. The power boundary between the world federation and major online platforms is four. Each fracture has real precedent, real controversy, real names. Which is exactly why they are perfect traps. If someone handed me an empty analysis and asked about cheating in chess, I could write three thousand persuasive words. About a famous 2026 case. About platform ban waves. About accusations never proven in court. All real. But none connected to that empty input. I would be answering a question nobody asked, with irrelevant facts, and presenting it as a finding. Analysis does not work that way. Filling a gap with memory is a different habit, and a more dangerous one. There is a psychological mechanism that makes this hard to avoid. When people meet an information gap, they tend to fill it with the most familiar story they already hold. In chess today that story is usually one of two: an old dynasty fading, or a new wave rising. Both templates are flexible enough to fit almost any event, and familiar enough that nobody rechecks them. The trouble is that we only remember the times those templates were right. In 2026, an eighteen-year-old Indian player became the youngest world champion in history. The Indian wave story was immediately retold everywhere, and this time it had grounds. More striking is that years earlier, the same template had been applied to other names who went nowhere. Those articles are still online. Nobody deletes them. Nobody corrects them. A content industry with no self-correction mechanism accumulates error over time, like a game in which the player never admits the twelfth move was wrong. In chess there is a tool that teaches this lesson clearly: the endgame tablebase. For positions with few pieces, it gives the exact result — win, draw or loss — and the optimal number of moves. When it returns 0.00, that is a conclusion: the position draws with accurate play. When the tablebase has no data, that is a different state entirely: unknown. The two look identical on screen. They do not mean the same thing. A decent chess writer must tell them apart, and must endure the discomfort of saying: unknown. But there is a deeper tier here, and it is the one that worries me. If an article about chess did exist and the machine could not read it — because of a paywall, a JavaScript page, content locked inside a video without subtitles — then the problem is not that we analysed wrongly. The problem is that we missed a story. A story that may be unfolding, may be shifting the shape of a tournament, may need reporting. And it sits silently in the gap between systems. That is the greatest risk of the information-automation era. The machine does not crash. It returns nothing. And people believe it. In chess, a similar situation has occurred many times. A fine move missed because the commentator was watching another board. An incident known only when someone reopened a recording. A talent discovered years late because nobody attended a provincial event. The chess world's memory has holes. Automated systems do not patch those holes. They only make them harder to see. I wrote a line in an article some years ago, and I still believe it: when the stadium no longer holds the ball, memory begins to score. But one clause must be added. Memory only scores when someone still remembers enough to tell it. If nobody remembers, an empty stadium is just an empty stadium. In my trade I have a rule for handling empty analyses. First, establish whether the input is truly empty or empty only at the extraction step. The two cases require different responses. Next, if it is an extraction failure, return to the raw source text, not the machine-processed version, and find at least one anchor: a name, a date, an event. One anchor is enough to reopen most of the tiers behind it. If after all that nothing remains, the correct report is an empty report, with a clear note explaining why. The last step matters most: do not delete that note when passing the report on. So I want to say something counter to the industry's instinct. People worry that artificial intelligence will give wrong answers about chess. That worry is misplaced. Wrongness is easy to detect. A wrong move, a wrong evaluation, a wrong figure — all verifiable. Far harder to detect is an answer that is formally correct but empty in substance: an analysis presented neatly, with terminology, with structure, appearing to discuss a specific game, while in truth it merely recycles a template. This kind of content is never caught, because it is not wrong anywhere. It simply has nothing in it. For chess this is especially dangerous, because chess is a sport where most important information lives in very small details. A different move order. A castling choice. An extra minute of thought on move twenty. Those details only come from a writer who actually watched the game, or at least reread the scoresheet carefully. A content aggregation system cannot do that, because it aggregates at the level of sentences, not moves. So I hold that the most valuable thing in chess media today is not the capacity to produce more content, but the capacity to refuse to produce it. Saying I do not yet have enough information is an expensive sentence in the attention economy. But it is the only sentence that preserves credibility over the long run. One more thing, and perhaps the part I believe most. In chess, people routinely undervalue information from silent sources: original scoresheets, official organiser notices, entry lists, rating tables. These sources do not go viral. They do not generate headlines. But they are the foundation. When a big story breaks, people rush to find those sources and often discover the information had been sitting there for weeks — nobody had simply bothered to read it. The day a heart stopped, I learned to beat more slowly but more deeply. I wrote that line after a June evening in 2026, when a player collapsed on the pitch at the European Championship and the world held its breath for 109 minutes. During that time no score was updated, no statistical table refreshed. If an automated system read that match through data, it would log: interruption. Nothing to analyse. But there was a great deal to analyse. Just not in numbers. The column of N/A stays on my screen, undeleted. It is a quiet reminder that my trade has a hard limit, and that the limit is not in the tools. A heart stopped, a heart still beats. I wrote that about a footballer, but it holds for a whole industry. When the data stream stops, the rest of the trade must keep beating — more slowly, more deeply, and more honestly. What I want to leave with younger readers is this. Next time you receive a seemingly flawless analysis of a match you never watched, ask one thing. What is the first concrete name in it. If there is no name, you are reading the memory of a machine — and that machine has never sat in the stands.