King's Indian at the Crossroads: 9...Nh5 or 9...a5 Against the Bayonet?

In the summer of 2026, in a small study room, I set the pieces back up on the wooden board. White's ninth move was 9.b4 — the Bayonet Attack, a bayonet thrust straight into Black's queenside. White's pawns rolled forward, claiming square by square, and opened a thickly documented chain of variations. Across from me, an old friend who once taught chess in the Soviet school laid his finger on Black's a-pawn and pushed it gently: "9...a5." He did not lecture. He said one thing: "Don't fight where they are waiting."

For years afterward I repeated that whenever I sat down before an opening board. That summer I learned an old lesson: a scheme looks beautiful only when the opponent agrees to stand still. And the story of 9...a5 in the Mar del Plata system is a clean demonstration of that principle.

The King's Indian Defense, in its Mar del Plata form, is one of the most classical opening systems of this century. Black deliberately concedes the center and concedes space, then strikes back with ...e5 and ...Nc6. Its beauty lies in the fact that Black does not seek immediate equality but builds strength for a blow on the kingside. The name Mar del Plata comes from the 2026 tournament in Argentina where the system took shape; ever since, it has been the weapon of players who like to swim against the current.

By the 1990s, White found the answer. The Bayonet, 9.b4, turns the queenside into a battlefield before Black can organize ...e5. White advances the pawn, gains space, and forces Black to calculate precisely from move nine. Against the Bayonet, the orthodox reply is 9...Nh5 — the knight hops to the rim of the board, opens the position, drives the variations into narrow channels, and makes computer analysis important again. That is the road for those who memorize.

9...a5 is the other road. Theoretically it is not stronger, and it is not remotely new — it is a well-known secondary branch. What it offers is a change in the character of the game: from concrete, easily memorized lines to a strategic struggle with few forced moves. The move ...a5 attacks the b4-pawn, opens the way for Black to reclaim queenside control, and steers the game toward a closed structure where both sides must play quiet moves — not delivering a blow at once, but accumulating small advantages.

What matters here is that the competitive essence is not which move is stronger. The two sides are considered balanced. The real value is practical and psychological: it is for players who want to think for themselves rather than sit and memorize. In other words, it sells an idea, not an advantage.

A move that opens the position and forces variations — that is 9...Nh5. A move that keeps the position closed with almost no concrete lines — that is 9...a5. In an era when the average player already knows thousands of moves, the difference between these two roads is not on the board; it is in the opponent's head.

There is a blind spot here that must be stated plainly. People call 9...a5 the "antidote" to the Bayonet, and that phrasing goes a little too far. An analysis sheet can lie, but a move that others have dissected down to the bone cannot. This is a known branch. Any well-prepared White player can find a rebuttal; and once the opponent knows in advance what is coming, the so-called "surprise" vanishes instantly.

The second blind spot is subtler. It is argued that computers often do not know how to handle the complex strategic positions arising from 9...a5. Written in 2026, that sentence still sounded reasonable — especially after Deep Blue beat Kasparov in 2026, when the whole chess world was shaken. But it was already going stale at the moment it was written. Machines were weak in long-range strategic positions precisely because they were not yet strong enough; as they grew stronger, that weakness would disappear. Choosing an opening because "the machine doesn't understand it" is betting on a gap that is steadily narrowing.

And the third blind spot: the word "antidote" does not exist in chess. No tactic is ever old; only the way of reading a position goes out of date. A move only solves one specific opponent, at one specific moment, with one specific preparation. Selling it as a universal formula is selling expectation, not selling the move.

So is 9...a5 worth adding to a player's opening repertoire? I think so — but as a knife in the drawer, not as a shield. It is useful for club players, for someone preparing against a specific opponent who already knows the Bayonet. It is not useful for anyone looking for a formula that requires no thought.

The opening-theory market is chess, not a card game — yet the sellers like to flip cards. And here is what I want to leave behind: watch whether, in the coming months, ...a5 steps out of the pages and appears in top-level games. If it reaches the board, the next question will be: what will White answer with? That is when the story truly begins.