9...a5 against the Bayonet Attack: A Strategic Path in the King's Indian Defense

9...a5 against the Bayonet Attack: A Strategic Path in the King's Indian Defense

In November, a reader in Kazan sent me a text file. The header read: "Samarkand R2 (open): Despite close calls, all top teams grab wins" — a team-event round report. The body, once I scrolled down, was a product blurb for an instructional DVD promoting Black's ninth move 9...a5 in the King's Indian Defense. The two documents have nothing to do with each other: one headline about a team event in Uzbekistan, one body about opening theory. Somebody stitched two unrelated sheets together.

I read the file three times. The first pass, to find the fault. The second, to see whether anything was worth learning. The third time, I pulled the board off the shelf. Because behind the broken label, the real question was still sitting there: can a ninth move filed under the secondary branch change the character of an entire opening line, and if so, what is the price.

Context: Mar del Plata and the Bayonet blade

The King's Indian Defense is a classic hypermodern system: Black concedes the centre early, builds a bishop on the long diagonal, and strikes back with ...e5. The Mar del Plata variation — where Black plays ...Nc6 and ...e5, with the knight retreating to e7 — is the sharpest branch of the whole system, sitting roughly in the ECO E97 to E99 range. Both sides push pawns on opposite wings, White on the queenside, Black on the kingside, and both know that being one tempo late is fatal.

For decades, Black solved that position with ...Nh5. The knight jumps to the rim, eyes f4, and forces White into concrete calculation immediately. That is the main line, where leading players have poured thousands of hours of analysis. Garry Kasparov used it as a strategic weapon in the 1990s; Teimour Radjabov and Hikaru Nakamura kept the branch alive in top events afterwards.

Then came the Bayonet. White's 9.b4 pushes a pawn at the queenside structure, gains space, and turns the game into a race White controls. The Bayonet became fashionable, and every King's Indian player needed an answer. The move 9...Nh5 answers with violence. The move 9...a5 answers with an entirely different question.

A ninth move against the current

Black pushes ...a5, attacking the b4-pawn that has just advanced. The gesture breaks White's comfort on the queenside, but not by opening the battle at once. It forces White to decide before completing the plan: push b4-b5 and lock the wing, trade on a5 to open the a-file, or keep the tension with a developing move such as Ba3. Three choices, three different types of position, and none of them hands White an automatic advantage.

Trading on a5 opens the a-file for White's rook, but by structural logic it also gives Black the tempo to push ...c5, opening the centre in a direction that suits the g7 bishop. Pushing b4-b5 keeps the space but turns the queenside into a dead wall, letting Black mass forces on the kingside. Keeping the tension means White lives with a hanging threat, and every subsequent move must account for a sudden ...axb4.

This is the real attraction of 9...a5. It changes the nature of the question. After 9...Nh5, both sides enter a narrow corridor where every move has a best answer and whoever forgets the book dies. After 9...a5, both sides enter a wide room where no single best move exists, only long-term plans that are better or worse. For a Black player short on time, that is a proposal worth weighing.

Placed side by side, the two branches differ along three axes. Preparation time: 9...Nh5 demands hundreds of variations, 9...a5 demands an understanding of structure. Position type: 9...Nh5 produces an open fight, 9...a5 produces a closed game where both sides improve their pieces in silence. Risk profile: 9...Nh5 loses to forgotten theory, 9...a5 loses to playing worse in a position with no theoretical life raft.

But changing the character of a game always carries a price. Black gives up the immediate attacking initiative. White keeps queenside space, and space is an asset that does not vanish simply because Black would like it to. If White plays correctly, Black's queenside can be squeezed for the first thirty moves, forcing Black to endure an uncomfortable position without any clear counterattack to relieve it.

The blind spot sits in the promise, not in the move

The promotional text accompanying the DVD promises three things. First, that 9...a5 leads to an objectively balanced position. Second, that it contains very few concrete lines, meaning little memorisation. Third, that it suits players who like to think for themselves. The third sounds reasonable. The first may be true. The second needs a simple test: where is the evidence.

There is no chart, no move-quality metric against an engine, no win-draw-loss rate from a database, no sample game cited, and the author of the product is not even named. A technical claim without data can still be true, but it cannot yet be verified. And when the sales copy states that once Black knows what to do, things quickly become dangerous for White, that is advertising language in absolute form, not the conclusion of an analysis.

The second promise has its own problem. The argument holds that computers often do not know how to handle such complex strategic positions. That sentence was correct in the pre-AlphaZero era. Today's neural-network engines handle the long-term, closed positions of the King's Indian far better than the copy implies. Keeping that argument in a new product is a sign of dated language, and it is also a commercial positioning trick: separating the product from the free tools a buyer could use instead.

And 9...a5 itself? It is a legitimate branch, known for a long time, and not an invention. A well-prepared White player can neutralise it, and once the branch is widely played, a White-side response will appear within one theory cycle. The real value of the product sits at club level, where players do not have hundreds of hours to memorise, rather than at grandmaster level.

What I got wrong

In 2026, my first analysis of RB Leipzig's gegenpressing was torn apart. I argued the model collapses against a deep defensive block. The online community called me a conservative. I did not argue back. I spent six weeks rewatching footage, logging 412 failed pressing situations, then published a correction with data. Since then I insert self-drawn charts and source notes into every piece. Data never takes offence.

Applying the same standard to this text file, I have to write two lines in the notebook. Line one: the technical content on 9...a5 has some value, worth testing at club level. Line two: the headline and the body do not match, the author is unnamed, the data is absent. A source like that can only be used after cross-checking against the original database and against games actually played.

Systems do not lie, but they can only be heard when the data is thick enough.

Takeaway

I will watch one signal in the coming months: whether 9...a5 appears in classical games at 2700 level or above. If it does, the promise of that DVD is confirmed by practice. If it does not, it is only the echo of a branch known for a long time, wearing an advertising coat. Patience is not stillness. Patience is waiting for the opponent's counterattack to arrive on the right beat.