Capablanca's 2026 Gold, the Mar del Plata Bayonet, and a Mislabelled Archive

Last week, at eleven at night, I opened a product page for an opening DVD — from my room in Pune, under the same thirty-year-old habit that made me a reporter of sport's paperwork rather than of sport itself. The page's headline read: "A farewell triumph: Capablanca wins gold in Buenos Aires 2026." The text beneath was something else entirely. It was sales copy for a Black repertoire against the King's Indian Defence, Mar del Plata Variation, Bayonet Attack. I read the whole blurb twice. No Capablanca. No Buenos Aires. No 2026. Not one player's name. Two entirely separate worlds had been glued onto a single page — on one side the 2026 Olympiad where European grandmasters were stranded in Buenos Aires, on the other a single move, 9.b4.

This is not a harmless typo. To me it is the most interesting part of the whole affair. Whatever chess world is now being built on YouTube and DVD has a first crisis that is not technical but one of labelling.

The second screen was not a distraction; it was where the story learned to speak. When I pitched a five-part series on 'The Anand Effect' in 2026, my quarrel was already with this same thing — Indian chess did not get a broadcast revolution, it got a documentation revolution. The camera arrived, the clock became visible, and the game became watchable. What looked like a chess broadcast was quietly becoming a sports documentary. But the documentation revolution has a naked truth nobody writes about: what gets recorded becomes history, and whatever gets a wrong label becomes wrong history. This DVD page is a specimen of that disease.

Consider the context. The classical King's Indian — 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.0-0 Nc6 8.d5 Ne7 — and here White plays 9.b4, the Bayonet Attack. This is the most fashionable blade of the Mar del Plata Variation, which has travelled from club level to the Candidates over three decades. Black has two roads. The first is 9...Nh5, the main line: a piece comes forward, the position opens, and the play turns into forcing moves quickly. The second is 9...a5, which this DVD is selling — a structural wedge on the queenside, a slow move, thinner concrete theory.

This DVD is a product of the old market where chess knowledge was sold in a sealed file. Today that same knowledge is open, sitting in annotation columns, searchable in databases. And precisely for that reason the question of the label has become bigger than before.

The real technical debate here is interesting, and honest too. The difference between 9...Nh5 and 9...a5 is not merely one of style; it is a trade of memory for judgment. In Nh5 the moves force — White must find precise answers, Black must reproduce prepared lines, and the whole payoff of preparation arrives in engine-verified files. In a5 Black has no forcing move to lean on; he must trust the dynamic compensation that comes after f5, and that trust comes from experience, not from a file.

Based on my years of watching matches, at club level this trade often inverts. The player who wants to memorise less is often the same player who has less positional experience to handle that low-forcing position. The DVD tells him there are fewer lines to remember; it does not tell him there are three or four losing routes inside those fewer lines.

The weakest thing in this blurb is its biggest promise — the claim that computers cannot handle these complex strategic positions. That was true, roughly before 2026. Engines of that era genuinely stumbled in closed King's Indian structures and long-term positional sacrifices. Neural-network engines — Stockfish's NNUE build, Leela-style networks — now evaluate these structures far more reliably. The trope that "engines get lost here" is now the language of marketing, not of technical fact. That small detail dates the entire product: this DVD was written from an era when closed positions were humanity's last fortress.

But one part of the blurb survives, and it should be allowed to survive. The real strength of 9...a5 lies in preparation economics, not in the equality of the position. A well-prepared White will have deep files on 9...Nh5; on 9...a5 his file will be thin. That argument is sound, even if the claim of "equality" cannot be verified. The problem is that this strength is rented — the day the line is entered in an opponent's file, the advantage ends.

And now back to that headline, because the history in it is real and that makes the error more dangerous. At the 2026 Olympiad in Buenos Aires, Capablanca, playing first board for Cuba, gave his last great international performance; war had already begun in Europe and several players were stranded there. The headline is not lying — it names a genuine historical moment correctly, it simply attaches it to the wrong object. This is the same error that happens in newspapers between a photograph and its caption, and which readers never forgive.

There is another technical impossibility buried in that headline. In 2026 no rating number existed; the Elo system came into use in 2026. Which means that gold medal cannot be converted into any rating coordinate today. When a page borrows history for advertising, it borrows history's measurability too — and that measurability does not exist. The real reference point for the ceiling of Capablanca's late career is the 2026 World Championship match against Alekhine, which went 6–3 in Alekhine's favour with 25 draws. That is a number that proves something; the 2026 gold is only a sentimental photograph.

Examine the technical claims and the gaps widen. Nowhere in the blurb is there an engine match rate, nowhere an ACPL, nowhere a database win-draw rate, nowhere a citation of a single game. A product that calls itself an "antidote" has no sample of its own claim. To me this is the same old disease that hides in the fine clauses of a contract — on paper everything is fine, until you look at the date and the place for a signature.

The buyer is equally clear. This is a product built for the club-level and improving player who wants to understand rather than memorise. The content is calibrated to that segment, not to elite preparation standards. That should not be called a fault — it is an honest picture of the market. The fault is that the same buyer plays mostly at fast time controls, where positional understanding must operate under the pressure of the clock. The DVD's entire premise — think for yourself — is weakest exactly where the amateur loses the most points.

The camera did not make chess a sport; it made the clock visible. In the same way the DVD did not change chess, it sealed preparation into a file. And whatever is sealed in a file lives on the trust of its label.

Here I have to turn on my own argument, because the easiest conclusion would be that the headline is clickbait and the matter is closed. My guess is duller than that: this is probably not clickbait but a sloppy bundling error on a newsletter page, where several headlines were stacked together and the wrong header stuck to the product text. Both readings share one thing — both are a wound to content integrity, and in both the real victim is the reader who leaves the page confused. We often lighten a large organisation's large mistake by calling it a conspiracy; often it is mere carelessness, which is not less frightening.

The genuinely contrarian point lies beyond the headline. The word "antidote" itself is the weakest part of the product. If 9...a5 is not the engine's first choice — and it almost certainly is not — then its value depends on White being under-prepared, not on the move being objectively superior. A buyer who wants an antidote in the literal sense, a refutation of White's edge, will be disappointed. Sharper still: the feature of having fewer concrete lines cuts both ways. It reduces the payoff of White's preparation, and it also reduces Black's winning chances against a strategically competent White. A White who keeps the queenside closed and plays on space can neutralise this "antidote" cheaply. In the King's Indian, dynamic compensation is demanded; it is not given, it is earned.

Follow the fee far enough and you find the culture in the paperwork. Here the fee is the headline, and the paperwork is that page nobody edits.

What comes next? My estimate is that over the next two or three years the opening-products market will go where cricket's data market went — the buyer will search first and buy second. Any improving player can now enter 9...a5 into a database and see how many games were played, who won, and in what proportion. The day buyers start checking labels, products will have to be honest. Chess's archive was built by streamers; now its clean-up will be done by consumers. Every arena is a track, even when the competitors are sitting still — and on this track the next race will be between headlines, not moves.