9...a5 Against the Bayonet: A Strategic Escape Route in the King's Indian Defense

Last November, at an open tournament in Chennai, I stayed behind after the game to note down move nine. At board four, a young Indian player pushed his pawn to b4 — the Bayonet. His opponent, an Armenian grandmaster, did not think long. He pushed the pawn to a5. Four minutes on the clock. For a move the spectators around the board called strange, four minutes is the mark of someone who knows where he is going. I wrote in my notebook: 9...a5, the fourth such response in the four games with the same opening I had watched that week.

Most people will skip that detail. But pushed out into the press-room corridor in 2026, I learned to read games from what others leave behind. Move nine is not the climax. It is the blueprint.

Context

The King's Indian Defense runs on a counter-intuitive logic. Black concedes the centre to White, then strikes back at that very centre. The classic sequence: 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.0-0 Nc6. This is the Mar del Plata system, the sharpest branch of the Classical line.

White usually chooses the Bayonet: 9.b4. White's queenside advances, gains space, and pushes pawns toward Black's queenside. The intent is clear — White wants to attack the queenside faster than Black attacks the kingside. Both sides race on opposite wings. Whoever arrives first wins.

Black has many replies. The most popular is 9...Nh5 — the knight leaps to the rim, opens the kingside, and forces the position into concrete, calculation-heavy lines. That is the choice of most professionals. But there is another option few play: 9...a5.

Analysis

The move 9...a5 attacks the b4-pawn head-on. Black does not argue on the kingside at once. Black asks a question first: what is that b4-pawn doing there? If White holds it, Black gains time to prepare ...c5 and blow up the centre. If White pushes on with b5, the pawn structure changes, and Black gets weak squares to exploit.

The real value of 9...a5 is that it does not force matters. This is the core difference from 9...Nh5. 9...Nh5 opens the position, and both sides plunge into lines a computer has already calculated to move thirty. 9...a5 is different — it keeps the position relatively closed and shifts the emphasis toward plans rather than calculation.

In other words, it is a trade-off. Black gives up direct pressure in exchange for a low-theory position. One path is a forced line, where one wrong move loses the game. The other is a strategic position, where understanding the plan matters more than remembering the move.

I believe in the diagram, but I believe more in the space between diagrams. 9...a5 lives in that space. It is not in the orthodox theory books. It sits in the region where the player has to think for himself.

One thing must be said clearly: 9...a5 is not a new invention. It has appeared sporadically in top-level events as a space-gaining move that attacks the b4-pawn and prepares ...c5 or ...Be6. What a recent theory product did was to systematise it into a complete repertoire, turning an isolated move into a full plan. That is the real contribution.

Objectively, the analytical verdict is that the position is equal. White loses nothing. Black gains nothing. But practically, if Black understands the plans and White does not, the position becomes dangerous for White. That is the product's central argument: the player who understands plans beats the player who remembers lines.

Contrarian angle

Here I have to stop and speak plainly. That argument has a hole.

The theory product argues that computers often do not know how to handle these complex strategic positions. That argument sounded reasonable in the past. Today, Stockfish and Leela evaluate strategic King's Indian positions reliably. Modern engines are no longer confused by closed positions. They do not understand the position, but they do not need to understand it in order to find the right move.

That is the product's time-bound weakness. An argument that was true ten years ago may no longer be true today.

And there is a bigger problem. No data shows that 9...a5 is better than 9...Nh5. There is no win rate, no draw rate, no average centipawn loss, no sample size. Every claim of technical superiority rests on prose, not on a database. For a writer who always demands evidence, that is a hole that cannot be ignored.

In other words: 9...a5 is a legitimate choice, not a proven fact.

There is one more layer. The product format itself is obsolete. A static DVD cannot be updated after release. Meanwhile, opening knowledge is flowing to subscription platforms, where theory is corrected continuously. A line refuted by an engine six months after release will sit frozen on the disc, with no way to fix it.

The press room had no seat for me. Tactical history always does. And the tactical history of the King's Indian is the history of moves that were doubted and then accepted — or forgotten. 9...a5 stands between those two fates.

Takeaway

The lesson of this story is not the specific move 9...a5. It is the way a marginal move is packaged into a philosophy: understand the plan rather than memorise the line.

The person standing at the edge of the board sees the whole game, not only what lies inside the goal. 9...a5 is the move of the person at the edge — not arguing in the noisy centre, but asking a question where few look.

Will it survive the engine's verification, or is it just a short chapter in the history of forgotten escape routes? The answer is not on the DVD. It is in the games to come.