The Silent Antidote in the King's Indian: 9...a5, the Bayonet Attack, and a Headline That Does Not Add Up

Twenty-seven moves had already been played. 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 9.b4 — White's b-pawn steps two squares forward and points a finger at Black's c5. The move has a name: the Bayonet Attack. Normally Black answers 9...Nh5, then ...f5, a closed centre, and a race of attacks on opposite wings. But when I sit patiently with the board, my eye drifts to another square — a5. Because after 9...a5 Black stops racing White stride for stride; instead he takes the rhythm away. Empty chairs in a tournament hall whisper a private truth, and this move whispers the same thing: the Bayonet Attack cannot always be met head-on.

The Shadow of Mar del Plata

The structure was born at the 2026 Mar del Plata tournament, and it carries that city's name. Black seals the centre with ...e5, reroutes the knight to e7, and plays ...f5. With the centre locked on d5, the two sides fight separate wars on separate wings — White pushes on the kingside along the e4-f5 axis, Black squeezes White's e4 with ...f5, and both kingside pawn chains advance in a fixed rhythm. The sharpness Gligoric and Bronstein gave this structure in the 1950s reached another level in the Fischer–Spassky battles of the 1960s and 70s. Fischer played it beautifully; it also gave him some of his hardest tests.

From the 1990s the picture changed. Computers arrived, and White found 9.b4. The Bayonet Attack is essentially a practical weapon — White leaps to the wing without touching the centre, blocks Black's natural knight squares, and forces Black to answer White's threat first. Black's standard reply became 9...Nh5. But 9...Nh5 has a price: the game opens up, the structure sharpens, and the theory books write long forced sequences for almost every move. For a player with only a few hours of preparation a week, those long sequences are enormous risk. So the club-level question becomes simple: can you keep the structural tension without entering a memory contest? That is where 9...a5 walks in.

What 9...a5 Actually Does

First, the move speaks against White's latest step. White played b4 hoping to occupy Black's c5; Black hits that pawn directly with ...a5 and keeps the future threat of ...axb4 to open the file. If White saves the pawn with b5, Black plays ...c5 without hesitation and cracks the centre with a second pawn. If White accepts the exchange, Black gains the a-file and White's whole queenside structure softens a little.

In the language of theory the move is called practical, not revolutionary. Nobody claims a new discovery; the claim is that it creates a position where Black's counter-plan is clearer than White's attack. Black's real gain is not in the statistics but in the psychological balance — a player who has not memorised the long forced lines starts thinking from scratch in this structure. The whole confidence of the Bayonet Attack rested on White's comfort; 9...a5 slowly pulls that comfort away.

But there is a cost beside the gain, and hiding it would be dishonest. With ...a5 Black leaves his kingside structure temporarily incomplete. If Black later wants ...f5 — and in this structure he almost always does — then ...a5 and ...f5 together scatter the pawns far apart. If White coolly converts b5 into a knight on c5 and then rolls the kingside pawns at the king, Black's risk grows. In other words, 9...a5 borrows time from White, and Black must spend that time well. A loan unspent is repaid with interest.

In 2026, in Batumi, I annotated all forty-four games while sitting with the Bangladesh team, and the same structural imprint kept returning — a knight stranded on the rim behind a fixed pawn chain. In exactly these King's Indian positions the imprint is clearest. The computer will call the evaluation roughly equal, perhaps a shade one way. But for the human at the table, inside that small difference lie hours of calculation and the fear that one wrong move ignites White's attack.

What the Machine Knows, What the Human Feels

A claim is often heard here: in these strategic positions computers supposedly do not know what to do. That is a half-truth. Modern engines do calculate the numbers correctly, but they do not feel tournament pressure — the ticking clock, the fatigue in the third hour, the opponent's real intent. An engine can say the position is balanced; it cannot say whether that balance becomes a nightmare for a 2500-rated club player over the next fifteen moves. A machine gives an evaluation, not a risk — and practical chess decisions are made by risk, not by evaluation.

But here I must be wary of myself. Since the core material of the analysis offers no engine data, no average centipawn loss, no win-draw-loss rates, the claim has no numerical proof behind it — only a sales argument. I do not dismiss the argument, but I do not accept it as truth without verification either. Likewise, 'very few concrete lines' does not mean the position is safe; peace in a King's Indian middlegame is almost never permanent. Anyone who thinks strategic means silent forgets the first rule of a two-wing fight — where the centre is closed, the bombs go off on the wings.

The Gap in the Headline

What stopped me most just before this analysis was not a move but a headline. The report says a high-level official meeting took place in Kazakhstan, attended by the country's President Kassym-Jomart Tokayev, and a man introduced as 'FIDE President', Timur Turlov. The headline and the body do not match: the headline speaks of diplomacy, the body of a commercial DVD's opening preparation. Two different things were folded into one envelope, and in that fold lies the work of verification.

I have to be clear here, because the Chennai experience of 2026 forced the lesson on me. That year I built a sixty-page preparation file on a 2540-rated Indian grandmaster and forgot to confirm his registration in time. He arrived two days late, played four games, scored 2.5. Since then I write the boring paragraph first — who played, when, the result, who paid. Then the analysis.

That same discipline says: the title 'FIDE President' cannot be used without verification. In widely available official records, the current FIDE President is Arkady Dvorkovich. Turlov is a known figure in chess administration and patronage, but if a headline calls him president outright, it is either a labelling error or a claim resting on incomplete information. An auditor's job is not only to criticise moves; it is to stand behind every word of a headline.

From this another question emerges, always relevant in chess politics. In an organisation like FIDE, the presidency is not merely a name but a centre of decisions. Who sits in the meeting, on what agenda, who funds it — without those answers a headline cannot be called analysis. A photograph of attendance and a structure of power are not the same thing.

This Structure at the Dhaka Table

The file that speaks the most truth to me is Bangladesh's. This country gave South Asia its first grandmaster — Niaz Murshed, in 2026, a year before Anand. Yet today the count is stuck at five, with no new grandmaster since 2026. The Bangladesh Chess Federation's annual government support is about eleven lakh taka. The Standard Chartered school chess programme ended in 2026, and the Premier League leans on Indian guest grandmasters for board one. These papers are the real game to me.

Through that lens I connect small preparation decisions like 9...a5 to a larger question. Where the pipeline loses a pawn at every stage — school to club, club to federation, federation to the senior team — where is the room for a young player to memorise brilliant theory? So the appeal of practical weapons like 9...a5 is even greater in a place like Bangladesh. Less memorisation, more thinking. But the fuel for that thinking also comes from a system — coaches, leagues, tournaments, practice partners. Without them, the best preparation strategy stays stuck on paper.

On 5 July 2026, in the middle of a league game in Dhaka, Ziaur Rahman died of a heart attack — fifteen-time national champion, holder of Bangladesh's highest-ever rating of 2570. I could not write for four days. Then I annotated forty-one of his games, every one of the fifteen title runs, eleven thousand words, one diagram at each critical moment. Those annotations taught me that a game is a document of a life before it is a diagram. When Niaz Murshed won the national title in 2026 at nearly fifty-nine, another line glowed beside the celebration — celebration without succession. And Manon Reja Neer's 2026 title at fourteen showed that talent does arrive; the question is not talent but the system that keeps it.

What to Look for at the Next Board

So I read this DVD's proposal on two levels. On the chess level, 9...a5 is an intelligent, honest answer — a route away from a long memory war against White's Bayonet Attack. But on the level of the announcement, two verification debts hang beside it: one, that strategic means safe — an equation that cannot be accepted without proof; two, that a headline places official diplomacy and a commercial product side by side — that too must be reconciled with numbers.

At the next tournament I will watch two things. First, in recent Bayonet Attack games, how often Black chooses 9...Nh5 and how often 9...a5 — that ratio will tell where the preparation market is heading. Second, the actual documents of the Turlov–Tokayev meeting — agenda, decisions, titles — because a correct move and a wrong headline can never stand on the same board. The page nobody commissioned is the only honest record.