9...a5 in the King's Indian: Deconstructing an Opening DVD and the Real Price of Surprise

Black's clock jumped from 4:11 to 3:47 when his finger touched the a7 pawn. Board seven of a weekend open in a city of a few hundred thousand people, no spectators, no television, just fourteen boards and an arbiter scrolling through his phone. White had pushed b4, then b5, then placed a knight on d2 and a bishop on a3. Black — the man playing the black pieces — told me afterwards that he had paid for an opening DVD, and that the advertising had promised the line would force opponents to "think for themselves instead of reciting". He did think. He thought for four minutes. Then he picked the wrong wing, dropped a pawn on move 22 and resigned on move 31.

I tell that small story because it is almost the entire content of a DVD being sold at the price of a professional course. The product names no author, no release date, contains no sample game, and cites no engine evaluation of any kind. In the information packet I have in hand, every data field is empty. That is why I have to analyse it exactly the way I analyse a football match with no video available: on the technical claims alone, with every unverifiable assertion labelled as such rather than smoothed over.

Context: a rare line sold as an antidote

The King's Indian is the defence in which Black concedes the centre entirely for ten moves, then demolishes that centre with the pushes e5 and f5. The name Mar del Plata comes from the 2026 tournament in Argentina where the variation was first dissected in public; more than seventy years later it still carries ECO codes E97-E99 in opening manuals and remains one of the heaviest bodies of theory a club player can choose to carry. Bobby Fischer and Garry Kasparov both kept the King's Indian in their competitive arsenals; Teimour Radjabov is the modern top-level player most closely identified with the system.

The Bayonet Attack, marked by 9.b4, is White's most uncomfortable answer. White does not wait for Black to demolish the centre; White opens a second front on the queenside first. The b4 pawn advances, the bishop goes to a3, the knight to d2, the rook to b1. Black has three or four responses, and the main road is 9...Nh5, putting the knight on the rim to keep f4 free for the pawn and prepare f5. That is where every book, every video, every coach converges.

Any Black player who chooses 9...Nh5 has to memorise. Not because the line is beautiful, but because it is narrow: once White pushes b5, the black knight must choose between two squares, the black bishop between two directions, and every wrong choice is punished by a concrete tactical stroke analysed out to move 25. The memory load is heavy enough that many players spend a lifetime in the King's Indian and never dare leave a single branch.

The DVD in question offers an exit: 9...a5. Black pushes the rook's pawn, attacks the advanced b4 pawn directly, and claims that from here the game shifts into the kind of position "computers do not understand". The blurb adds that "the chances are objectively equal" and that the buyer will "think rather than memorise". Three claims. Not one of them comes with a number.

What 9...a5 actually does on the board

Structurally, the move does exactly what the blurb says. The a5 pawn forces White to decide immediately on the queenside: either push b5, or trade on b4 and hand Black the open a-file. Both are strategic decisions, not tactical ones. And that is precisely the point the blurb does not state: 9...a5 does not make Black objectively stronger, it reduces the variance of preparation cost — a perfectly legitimate practical benefit mislabelled as an improvement in the quality of the position.

More concretely: if White pushes b5, Black gets an open a-file for the rook and a clear target in the b5 pawn, while White loses the ability to open the b-file for his own rook. If White trades on b4, Black gets an open a-file for free, and the pawn structure shifts into a shape where Black's majority sits in the centre rather than on the wing. Both branches lead to the same outcome: the game leaves the motorway of pre-analysed lines and enters territory where both sides must draw their own map. That is a skill, and the skill is not evenly distributed between the two people sitting opposite each other.

That is the whole difference. A line with heavy theory gives the player a narrow corridor with walls on both sides. A line with thin theory gives the player an open field with no walls, and the player must know where to walk. For a 2300 player, the open field is an advantage. For a 2026 player still trying to remember the first fifteen moves of some other variation, the open field is a minefield.

I spent six months of 2026 digitising 2,400 old matches from my notebooks, and the biggest lesson from that work was not about football. The empty stadiums of 2026 were the most perfect laboratory sport ever accidentally created, because they forced me to drop the habit of looking at results and start looking at structure. Applied to an opening, the question is not whether the line is strong or weak, but which type of skill the line loads. 9...a5 loads strategic judgement and unloads memory. For the seller, that is a story that sells. For the buyer, it is a bet on exactly the type of skill he does not yet have.

A documentation gap: two moves left out

There is a small detail in the blurb I cannot let pass. The move sequence presented stops at move seven with 0-0, then jumps straight to 9.b4. The two moves 7...Nc6 and 8.d5 Ne7 vanish. Without those two moves the Bayonet does not exist — White cannot push b4 into a position where the central structure has not yet closed.

This is not a chess error. It is a documentation error, and it is diagnostic: the text was written for someone who needs persuading, not for someone who will sit at the board and execute the line. A decent opening course begins by drawing the actual move tree the student will walk. An advertisement begins by drawing the move tree the reader wants to believe.

The same kind of gap appears elsewhere. The blurb never says how White replies. White has three natural directions — b5, Ba3, Nd2 — and all three are the refutation front. None is discussed. An opening DVD that does not devote a chapter to the opponent's responses is not a repertoire, it is an invitation.

The "computers don't understand" argument and its age

The claim that computers do not understand these positions was once true, and true for a long time. Before the neural evaluation era, classical engines searched depth and counted material; a closed position with no tactical stroke and no clear pawn structure made them misjudge in favour of the side with more space. That was the foundation of an entire generation of opening books.

That foundation began to crack in 2026 with AlphaZero and collapsed in 2026 when Stockfish 12 integrated NNUE, a neural network evaluation replacing the hand-written evaluation function. Around the same time, Leela Chess Zero learned by self-play and produced evaluations that grandmasters needed a week to understand. Closed, low-space, structure-rich positions — exactly the kind the blurb says computers do not understand — are where the new engines show their advantage most clearly, because that is where strategic understanding matters more than raw calculation.

Put differently: the blurb's claim is a legacy claim. It was correct around 2026. It is no longer correct in 2026. And that does not strip 9...a5 of value — it only makes the sales argument obsolete.

A memory gap, not an understanding gap

The argument that actually survives is narrower and different. In the positions after 9...a5, there are fewer forcing lines. Black does not need to remember eighteen moves; Black needs to understand three ideas. But Black's opponent still has an engine. He can still open the machine and check the position for fifteen minutes before sitting down. What 9...a5 takes away from White is not the ability to check, but the ability to check from memory.

This is the distinction the blurb blurs: "the engine dislikes this position" and "the engine has not checked this position" are two completely different statements, and the second is false. Every position on the board can be checked. Everything on the board is data waiting for a reader — if you are willing to sit down.

What 9...a5 buys is time. Specifically, the opponent's clock time in the first fifteen minutes of the game, when he realises he is no longer standing on the road he studied. A 2200 player who has never faced 9...a5 will spend several precious minutes reconstructing the structure from scratch. Those minutes, in a game of 90 minutes plus 30 seconds a move, are the entire value of the product.

And here I have to question a number. The blurb says "the chances are objectively equal". There is no engine evaluation, no game reference, no database statistic. In practice the Mar del Plata structures give White durable space and a queenside majority plan while Black plays for the initiative. That is a high-variance balance, not a symmetric equality. The claim may be approximately right at engine depth, but it is stated without evidence, so I mark it unverified.

The half-life of surprise

This is the central risk of the product, and the blurb never touches it: the benefit of a rare line is inversely proportional to the number of people who know it. A move that appears in 0.4 percent of top-level games carries enormous surprise value. A move that appears in 4 percent carries almost none, because the opponent already has it in his files.

The blurb does not state the frequency of 9...a5. It does not state the number of games in the database. It does not state the score. For a product whose entire value rests on surprise, the silence around those three numbers is deliberate. And I understand why: if they were stated, the buyer would see immediately that the goods have an expiry date.

The opening-course market is the only place where people pay for surprise and then blame time. The buyer of the disc is not buying a position; he is buying the interval before someone else reads that position to the end. No manufacturer prints an expiry date on the box, and nobody takes responsibility when the date arrives earlier than expected.

Who actually buys this disc

The blurb names no player at all — no author, no grandmaster endorsement, no sample game by anyone. For an instructional product in chess this is a significant credibility omission, because the genre norm is an author's name with a title and a handful of model games. That absence is not neutral; it is an information gap.

But the buyer archetype is legible even without a name. He is an improving player, somewhere between 2026 and 2300, with limited preparation time, playing the King's Indian as a weapon to win with rather than to draw with, and currently losing to opponents armed with the Bayonet. The product is positioned against preparation asymmetry, not against a specific opponent.

I split buyers into three bands, and this is inference, not data. Below 2026: medium-high practical utility, because opponents at this level rarely handle the Bayonet properly, so strategic guidance beats memorised theory. From 2026 to 2400: peak utility, because opponents here do prepare main lines, so a rare move produces the largest surprise dividend. Above 2400: medium-low utility, because opponents and their seconds will have the move in their files, the rarity advantage collapses, and 9...Nh5 reasserts itself.

The striking thing is that the band gaining most is the band with the least money and the fewest disc purchases. The band buying the most discs is the band for which the product has the least value. That is a familiar marketing paradox, and the blurb does not explain it.

One more variable the blurb leaves entirely blank: style fit. A player who likes direct attack and combinations will find 9...a5 a prison. A player who likes to hold, to accumulate small advantages, will find it a home. With no author name and no sample games, the buyer has no way to assess that fit before paying. This is the single most important information gap in the whole product packet.

Risk under a fast clock

One more technical point the blurb omits: strategy-first lines are more time-hungry than memorised ones. In rapid and blitz, a forcing line already learned lets the player move in three seconds. A strategic position forces the player to re-evaluate every move, and at thirty seconds a move, strategic judgement compresses to the point of collapse.

In other words, this product sells a weapon for slow chess and does not say that the weapon is heavier in fast chess. The buyer will discover that in the third round of a rapid event, when the clock shows two minutes and there is no move left to remember.

The contrarian angle

The most counter-intuitive thing here is that the blurb may be selling exactly the right thing while saying the wrong thing about it. It says: this line is objectively better. The technical truth: this line is not better, it is cheaper. And for a large share of players, cheaper is better.

I no longer believe in miracles on the board; I only believe in conversion rates. A move does not create chances; it exchanges one type of chance for another. 9...a5 trades tactical chances for strategic chances, trades win probability from preparation for win probability from judgement. For a player whose judgement is better than his opponent's, that is a good trade. For a player whose judgement is worse but whose memory is better, it is a bad trade, and the blurb does not tell him so.

The second contrarian angle: the very vagueness about data is the clearest signal of where the product sits in the market. A line that is genuinely strong objectively gets advertised with scores, with game counts, with engine evaluations. A line advertised with a feeling of freedom is selling the feeling of freedom, not the position. That does not make it useless. It only means the buyer needs to know what he is buying.

How I would verify this product

If I were asked to vet it on a buyer's behalf, I would ask five things. First: the total number of games in the author's database featuring 9...a5, split by rating band. Second: Black's score in those games against the score of 9...Nh5 in the same bands. Third: engine evaluation at a depth of at least thirty plies for White's three natural replies. Fourth: one complete sample game with annotations for every move, so I can check whether the author actually plays what he teaches. Fifth: a list of the opponents against whom the line was tested, with results.

Without those five, an opening DVD is a hypothesis in a box. The hypothesis may be correct. But the buyer is paying for a conclusion, and conclusions need a sample. The sample behind this product, as far as the blurb reveals, is small — monographs on rare sidelines of this type typically rest on a few dozen games plus the author's own analysis. That is why the blurb emphasises "very few concrete lines" rather than "extensively tested".

Conclusion

The King's Indian has survived seventy years on a paradox: Black concedes the centre in order to win the centre. 9...a5 is a smaller version of the same paradox — concede the certainty of theory in order to win the freedom of judgement. Whoever understands what he is conceding will use it well. Whoever buys it believing freedom is free will pay on move fourteen, when the clock has run four minutes and no book says what comes next.