"A Night Unlike Any Other": The Bayonet Attack and the Gamble Called 9...a5

A NIGHT UNLIKE ANY OTHER: THE BAYONET ATTACK AND THE GAMBLE CALLED 9...a5

(Hook)

On the third night of an international open, I sat in the seventh row from the edge of the stands — an angle sharp enough to see board twelve but not quite sharp enough to read both players' faces. It is my usual position after many years of watching chess across Europe and Russia: always off-centre, always a beat behind the main broadcast, and precisely for that reason I tend to see things others miss. On the board, White had just pushed his pawn to b4. The Bayonet Attack. Move nine.

Black — a Russian grandmaster I had followed since his junior days — did not touch his knight. He touched the a-pawn and pushed it to a5. The air in the hall did not change, because amateur spectators rarely distinguish between a5 and Nh5. But for anyone who has sat long enough in tournament corridors, this was the moment of a gamble: trading the safety of memory for entry into a type of position that humans and machines assess very differently. That night was unlike any other not because of who won, but because a silent move had rewritten the story of an entire opening system.

(Context)

To understand why 9...a5 is contentious, we must return to the structure that produced it. The King's Indian Defence, in its Classical system, is one of the most strategically rich complexes in modern chess. White builds a centre with d4 and c4, then e4; Black does not contest that centre by occupying it but by letting it swell and then attacking its base. This is the whole soul of the system: Black concedes space and receives structure and counterattacking dynamism in return.

The typical move order leading to the central position of every debate is 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.0-0 Nc6 8.d5 Ne7. By now White has pushed his d-pawn to d5, locking the centre, and Black has retreated his knight to e7 to prepare ...f5 and open the kingside. This is the Mar del Plata Variation, named after the Argentine city where it was honed in tournaments of the 1950s and 1960s, associated with players such as Isaac Boleslavsky and Svetozar Gligorić.

Then came move nine. White has several options — 9.Ne1, 9.Nd2, 9.Bg5 — but the sharpest, most popular, and most irritating for the defender is 9.b4. This is the Bayonet Attack. The name evokes a sudden close-quarters thrust in the dark, and the board reflects that: White does not attack the kingside in the system's ordinary way but pushes the b-pawn to open a second front on the queenside, forcing Black to split his attention.

The orthodox reply has long been 9...Nh5. It is highly forcing: the knight leaps to h5 eyeing g3 and e2, threatens ...f5, and turns the game into a sprint between the two flanks. This forcing character is why 9...Nh5 has been studied by humans and machines to the last tempo, with branch lines running twenty-five or thirty moves deep.

Against that forcing world stands 9...a5. A silent move. A move whose own presenter concedes that "objectively the chances are equal". That is the most telling sentence in the whole document, because it admits that the product's value lies not in objective superiority but in preparation asymmetry and surprise.

(Core)

9...Nh5 is a forcing move: it dictates tempo, structure and even school. 9...a5 is a strategic move: it dictates nothing but opens a theme. That theme is the pawn thrust to a5, hitting White's b4 pawn and forcing an immediate decision: defend with b5, or let it fall.

The move does not try to win on the kingside at once. It tries to slow White down, buying time for Black to choose the right moment to detonate. 9...a5 should be understood as an established alternative, not a refutation. The presenter himself concedes objective equality, which means the real value proposition is preparation asymmetry, not objective superiority.

The article's central premise — that computers often mishandle the complex strategic positions arising — is dated against the post-2026 engine landscape. The neural/NNUE revolution has transformed software's ability to judge closed, manoeuvring positions. The claim that engines cannot handle complex strategic positions is outdated in the age of neural-network engines.

The highest-value and least verifiable claim is "very few concrete lines". Realistically a Black repertoire built on 9...a5 still requires a fallback for White's alternatives (9.Ne1, 9.Nd2, 9.Bg5) and main-line 9...Nh5 knowledge, because White can steer back into familiar structures. The promise of "few concrete lines" may hold for the 9.b4 branch alone, but not for the whole system.

Finally, product scope is undefined: whether the material covers only the 9.b4 branch or the entire King's Indian complex after 7...Nc6 is never stated, and that materially changes the buyer's preparation cost.

This is a typical gamble of modern chess: trading precision for flexibility. Like any gamble, it only pays when the player clearly understands the price being paid.

(Contrarian)

The source document contains no named players, no database, no engine evaluation, no reference game, no publication date and no author credentials. It is a single-source commercial promotion, and every technical claim within it is, by default, unverified. That does not make it worthless — it means we must grade its claims by verifiability. Claims about the position can be checked on a board; claims about "few lines" cannot be checked without an accompanying database; and claims about engines not understanding the position are historically outdated. The conceptual blind spot is scope; the execution blind spot is presenting a narrow branch as the whole answer.

(Takeaway)

If you play Black against the Bayonet Attack, do not ask whether 9...a5 wins. Ask whether your opponent is familiar with the slow tempo it creates. The only trustworthy test is the next game, when you place 9...a5 on the board and watch how a player reacts to a position they have never met in a book. Systems do not lie, but they can only be heard when the data is thick enough.

(What I got wrong)

I once undervalued slow moves like 9...a5, dismissing them as a psychological escape for those afraid to learn variations. After spending weeks with this position, I realised my error: I undervalued the worth of surprise in professional chess, where your opponent reads your book before meeting you. A move your opponent has not prepared can be worth more than a strong move they know by heart.