The a5 Move and the Bayonet: The King's Indian's Change of Direction, July 2026

The white pawn stepped to b4, and the small room where I sat went silent. A July afternoon in 2026. On the table lay a theory bulletin that had just arrived by post, its paper still smelling of ink. The move carried a fierce name: the bayonet.

White pushes the pawn to b4, driving a spike into Black's queenside. For anyone used to garrisoning the King's Indian Defence — Mar del Plata, this position is like a knock at the door at midnight. An entire defensive system, built on faith in the kingside, in the g-file, in the flood of pawns on the right side of the board, suddenly has to look over its shoulder.

I remember sitting for a long time over that position without making a move. Some positions teach you that a chessboard is not a place for remembering, but a place for thinking.

The 21st Chess Olympiad in Nice, France, had just closed on 30 June 2026, and every dispatch said the same thing: the games there were decided in the opening. The King's Indian Defence was among the most discussed systems, and inside it, the Mar del Plata branch was where the arguing ran fiercest.

The King's Indian was born from a contrarian idea: Black willingly concedes the centre to White, then attacks that very centre. Mar del Plata is the sharpest branch of that idea — Black puts a knight on c6, pushes the e-pawn to e5, builds the d5–e5 pawn wall in front of White and opens a two-wing race: White attacks the queenside, Black attacks the kingside. Whoever arrives first wins. There is no defending here, only speed.

For years, Vietnamese players studied Mar del Plata like scripture. The names attached to the system — Svetozar Gligorić, Wolfgang Uhlmann — were copied into almost every young player's notebook. We knew moves nine through twenty by heart and believed that knowing the line was enough to survive. The trap lay there: the more you memorise, the less you think.

Then the bayonet appeared. White plays b4 on move nine, without haste, without fury, simply sealing the queenside before opening the front on the other side. Black must answer. The familiar reply is Nh5 — bringing the knight to the rim, opening the position, forcing both sides into long strings of concrete moves. That is the answer of the age: strong, head-on, and demanding a monstrous memory.

The bulletin in my hand proposed another answer: a5.

Black pushes the pawn to a5, striking directly at the b4-pawn that has just advanced. It sounds simple; the consequences are not. When White decides to close the position with b5, the queenside is sealed at once. A closed queenside means White loses his most natural direction of attack, while Black gains time to rearrange his pieces to an entirely different rhythm: the knight to e8 and then g6 or c7, the f-pawn advancing once the wall is firm, and the kingside becoming the only road left for either side.

The core point is this: a5 is not stronger than Nh5 in theoretical terms; it changes the kind of question being asked. Nh5 forces the opponent to answer from memory. a5 forces the opponent to answer with judgement. For an ordinary player, without an analysis team behind him, the second kind of question is far easier to breathe through.

Structurally, a5 turns the queenside from a battlefront into a wall. White loses his target, loses his direction of attack, and must switch to a central or kingside plan — plans that demand patience, which players used to winning quickly rarely have. Black trades space for clarity. It is a classical chess exchange, and a very modern one.

I set that position up again many nights in a row. Each time, I saw the same thing: Black does not need to win on move 10, 11 or 15. Black only needs not to lose on those moves. That modesty is the strength.

But it must be said plainly: the label of "antidote" attached to a5 is an overstatement.

a5 is a known sideline. There is no discovery here. Any player who bothers to leaf through a few months of literature will meet it, and anyone who has played the bayonet long enough has already prepared an answer. The collective memory of the chess world always carries a blind spot: it believes one new move is enough to solve an old system.

The promise of "very few concrete lines" has also been softened. Mar del Plata remains Mar del Plata. Both sides still push pawns on both wings, still race for speed, and the line between a correct move and a wrong one is still thin as thread. If White keeps the position open and lands his kingside blow first, that wall becomes a fence imprisoning Black himself.

More subtly, a5 aims at a certain kind of player. It tells him, "you are a thinker, not a memoriser." That is a seductive invitation, and one that makes people misjudge themselves. The bulletin circulating in Saigon and Da Nang this summer names no author, cites no practical game, and offers not a single piece of data to verify. Based on my experience following games at domestic events and through bulletins sent from abroad, I am obliged to read it as an introduction, not a conclusion.

The chessboard does not reward the man hunting for a miraculous move. It rewards the one who sits one beat longer after the homework is done. I once thought I had been exiled from the board when I lost my youth to dry pages of theory; it turned out the board was leading me deeper into myself.

And the question I took home after that July afternoon remains whole: is the chess player winning with someone else's memory, or with his own understanding?