The 9...a5 antidote in the King's Indian: inside an opening DVD sold under a Capablanca 2026 headline

At a small chess club on Tverskaya Street in Moscow last winter, a friend set a clear plastic case down on the table. The cover showed a man in a pale suit, seated sideways, chin on his hand, eyes on the board. Beneath the photograph, in capitals: CAPABLANCA — BUENOS AIRES 2026. He bought it for that line. He expected to watch a classic game, perhaps one of the last played by an ageing champion before Europe collapsed.

Then he put the disc in.

For the first forty minutes there was no Capablanca. No Buenos Aires. No 2026. There was an electronic board, an even-toned narrator, and a position any King's Indian player recognises within three seconds: 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 O-O 6.Be2 e5 7.O-O Nc6 8.d5 Ne7 9.b4. The Bayonet Attack. And the answer printed in bold on the back of the box: 9...a5.

I watched all forty-five minutes. Not because I needed to learn the ninth move — I have known it for years and know where it sits on the theory map. I watched because of the distance between the headline and the disc.

From the touchline, I see the whole game.

Why the Mar del Plata is a memory problem

The King's Indian is the strangest defensive system in the modern canon. Black hands White an enormous centre — d4, e4, c4 — builds a wall around it with Bg7, d6, e5, and waits for that centre to swell until it suffocates itself. Then Black pushes ...f7-f5. The entire doctrine compressed into one pawn move.

The Mar del Plata Variation emerged from the tournament of that name in 2026, where 7...Nc6 appeared densely at the top boards. There is nothing mysterious about it: it aims at d4, prepares ...e5xd4 to open the e-file and, above all, keeps the ...f7-f5 plan intact. After 8.d5 Ne7 the centre locks. White has space on both wings. Black has two wings to counterattack with and a knight looking for a rim square.

The position becomes a two-wing race. White pushes c4-c5 and b4-b5, opens the b-file, swings the queen queenside. Black pushes ...f7-f5, opens the f-file, swings the queen kingside, sometimes throwing in ...g6-g5-g4. Each side attacks a different part of the board and nobody is allowed to be slow. In such positions the loser is usually the one a tempo behind, not the one who misread the plan.

Which is exactly why 9.b4 appeared.

The Bayonet Attack has a precise target: take queenside space before Black can deploy counterplay there. The b4 pawn controls c5, threatens b4-b5 to drive the c6 knight away and, more importantly, paralyses ...b7-b5 and ...c7-c6. When the Bayonet became fashionable it dragged a whole chain behind it: thousands of club games, hundreds of lectures, and a memory arms race that amateurs can never win.

That is the ground this DVD stakes a claim on.

9...Nh5 and 9...a5: two different worlds on one board

The old route against the Bayonet is 9...Nh5. The knight goes to the rim, ...f7-f5 follows, and the position opens up. Black accepts a calculating battle: every move has a reply, every branch has a sub-branch, and whoever misplaces a single square inside a twenty-five-move forced sequence loses. That is the world of technique and memory.

9...a5 does something else entirely. It strikes directly at the b4 pawn and forces White to decide on the spot: either push b4-b5 and freeze his own queenside, or allow the exchange and accept a long-term weakness. No immediate win. No tactical trick hidden in the corner. Just a structural question.

The DVD describes this trade-off in four sentences, and those four sentences are the entire commercial argument: 9...Nh5 opens the position, creates forced lines and therefore hands the advantage back to whoever has computer support. 9...a5, by contrast, produces sound positions with very few concrete lines, where strategy outweighs tactics.

The real value of 9...a5 is not that it is stronger than 9...Nh5, but that it changes the kind of game being played.

That is a repertoire-direction argument, not a novelty argument. The product promises no refutation, no forced advantage, no unknown move. It promises a different entrance. For a seller this is far safer positioning than claiming to have broken the Bayonet — every claim to have broken something can be dismantled by a single game.

And here is a detail I respect, buried in the marketing language: the product states that the objective assessment of the position is equal. Not better for Black. Not better for White. Equal.

In more than fifty years of reading opening material, I have rarely seen a product diminish itself this way. Most opening DVDs sell by claiming their line is better than someone else's. This one sells by claiming its line is equal but unpleasant. Those are very different pitches, and the second is far more credible.

The closing line of the advert — that 9...a5 is dangerous for White if Black knows what he is doing — is another way of saying the same thing. Practical danger, not objective advantage. The buyer is not promised a win. The buyer is promised a position the opponent must think through himself, and most club opponents are not trained to think through closed positions.

Who actually benefits from this structure

Based on my experience of watching games, I would split players into three groups.

The first is the club player, roughly 1800 to 2200, who plays the King's Indian as Black and meets the Bayonet a few times a season. For this group the value is real. They have no time, no seconds, no private database. A low-memorisation line with a clear plan gives them more than a theoretically correct line requiring twenty-five exact moves. They do not need to be stronger. They need less to remember.

The second is the computer-prepared player above 2400. For them 9...a5 loses most of its appeal. They have engines, databases and time. A closed, low-forcing position is precisely the position they can prepare best, because plans can be learned and tactical reflexes cannot.

The third is the coach. For a coach, 9...a5 is a far better teaching tool than a forced line, because it forces the student to answer the right structural question: is my queenside healthy or weak, and where should I play first. A position that can be taught outlives a position that can be memorised.

These three groups produce three different verdicts, and that is what no advertisement ever says: an opening product is not good or bad in the abstract, it fits or it does not.

What the advert leaves out: the clock

One variable is missing from the whole sales argument, and it is the most important variable in practical chess — the clock.

A strategic, low-forcing position requires the player to build a plan at the board. Building a plan costs time. In a classical game with ninety minutes plus increment, that is a sound investment. In a three-minute blitz game, it is a loss.

The advertised strength — few forced lines, less memorisation — becomes a weakness in fast formats. A blitz player does not win by understanding structure. He wins on drilled reflex. Someone who has spent six months drilling 9...Nh5 has reflexes there; the buyer of this DVD has none, only a plan that needs time to unfold.

In other words, the product sells a good item for slow chess and a risky item for fast chess, and the advertisement does not distinguish between the two. The buyer must.

Why I still want a dataset

Seven months without football, seven months of asking why. In 2026, when every competition stopped, I used those seven months to compile a dataset of 214 goalless draws across Europe's five major leagues from 2026 to 2026, classified by nine pressing models. I did not do it because I like spreadsheets. I did it because without data there is nothing to argue against.

The same principle applies intact to an opening DVD. To know whether 9...a5 is worth playing I need four things: the practical game count for the line, the win-draw-loss split by rating, the distribution by time control, and the share of games in which White found a clear antidote.

None of the four appears in the advertisement. No sample size. No reference games. No engine match rate, no average centipawn loss, no publication date to tell me when this theory was packaged.

The system does not lie, but it can only be heard when the data is thick enough.

Here the data is zero. That does not make the technical argument wrong. It makes it unverifiable, and in my line of work those are very different things.

The headline, and what people actually buy

Back to the box on the table in Tverskaya Street.

José Raúl Capablanca was in Buenos Aires in 2026. He played first board for Cuba at the Olympiad held there, late in his career. He died in March 2026 in New York. Those are verifiable facts. The gold medal attached to the headline I could not confirm anywhere, and the disc itself never mentions it once.

The thirteen information points in the advertisement — I counted three times — contain not a single reference to Capablanca, to Buenos Aires, or to 2026. The entire technical content turns on 9...a5 in the Mar del Plata.

There are two explanations. The first: a packaging error, a headline spliced in from another product by the same publisher. The second: a deliberate editorial hook — using a legendary name to sell dry technical content.

Both lead to the same practical conclusion: the headline cannot be used as information. It can only be used as a market signal.

And that signal says a good deal. A DVD about the ninth move in the King's Indian does not sell with a picture of an open board. It sells with a picture of a person. The publisher knows this, and put the legend on the cover while leaving the technique inside the disc.

People watch the pieces. I watch the whole formation shift.

The execution blind spot: the sentence about computers

One sentence in those thirteen points deserves more attention than the rest: computers often do not know how to handle the complex strategic positions that arise.

That sentence was true around 2026. In 2026 it is false.

Modern engines handle Mar del Plata structures better than any human, and they handle closed, low-forcing positions more accurately than open ones. Where engines used to be weak was where many plans share the same value — a problem of brute-force evaluation, not of the neural era.

Read literally, the sentence is a technical error. Read as metaphor — that humans struggle inside those strategic mazes without a guide — it is entirely correct, and that is what the product actually means.

This is a common fault in the opening trade: an argument sound in pedagogy presented in technical language that has expired. The buyer loses nothing to the error, but should know he is buying advice about how humans think, not a claim about the limits of machines.

The second blind spot sits inside the promise of few forced lines. Few forced lines means fewer paths you must follow, but also fewer paths the community has verified. The less a line is played, the less data exists, and the easier it is for a well-prepared White to neutralise it with a plan nobody in the DVD mentions. The safety of a repertoire does not vary inversely with the number of forced lines. It varies directly with the amount of verified data on that line.

The third blind spot is shelf life. The product lives off a wave — the Bayonet is fashionable. When the wave turns, its value turns with it, and nothing in the advertisement says when it was packaged.

The first piece I ever wrote was stoned. Data never takes offence. In 2026, at sixty, I published 3,400 words on Leipzig's gegenpressing and was called a reactionary. I spent six weeks re-watching footage, logging 412 failed pressing situations, then published a data-backed correction. I would do exactly the same with 9...a5: six weeks, a dataset, and a conclusion that can be argued against.

Patience is not stillness. Patience is waiting for the opponent's ...f7-f5 to arrive on the right beat.

What to watch over the next six months

Four signals, written down so I can check myself later.

First, the frequency of 9.b4 in elite databases. If it keeps rising, the product has a reason to exist. If it falls, both the commercial and the technical argument lose their footing.

Second, the practical score of 9...a5. If the line drifts away from equality in Black's favour, the pitch is confirmed. If it drifts the other way, it is refuted — and refuted by data, not by feeling.

Third, the emergence of a clear White counter-plan. One strong new path is enough to turn the advantage of few forced lines into a disadvantage of few options.

Fourth, the identity of the author. An opening DVD with no named author is an opening DVD that cannot be assessed. The preparer's credentials are the decisive quality variable, and they are missing.

A forward thought

As my friend closed the case, he said something I expect to stay with me: so I bought a cover.

Not quite. He bought a structurally sound argument wrapped in an unrelated cover. In the opening trade those two things usually travel together, and the strong player is the one who can separate them.

What I want to know next season is not whether 9...a5 wins. I want to know whether a product willing to call its own position equal — willing to say outright that it gives you no advantage, only a different kind of game — can sell. If it can, the opening market is maturing. If it cannot, publishers will keep printing an old champion on the cover and hiding the real lesson at minute forty.

I am sixty-nine. I still learn from the young ones. Chess does not retire.