King's Indian, Mar del Plata: Black's 9...a5 Against the 9.b4 Bayonet — a DVD Proposal, Engine Silence, and a Headline Still Awaiting Verification

"I laid the board out, and the pattern had been waiting." Sitting down with an analytical document, I spotted the crack before anything else. The headline announces a high-level official reception in Kazakhstan — President Kassym-Jomart Tokayev present, and beside him a name tagged "FIDE President": Timur Turlov. But scroll down, and the substance is not diplomacy at all. It is an opening DVD for Black against White's Bayonet Attack 9.b4 in the Mar del Plata branch of the King's Indian Defense, whose central proposal is 9...a5. Two worlds sit side by side in one file — an official handshake on one page, pawns, squares and a chessboard on the other. The headline half still needs its sourcing checked; so I take the DVD as this article's technical object, and I file the headline away in a separate column of the ledger.

The name comes from Argentina's 2026 Mar del Plata tournament, where this structure first stood on a big stage. The main line is familiar: 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. The centre closes here. White's space is on the queenside, Black's counterattack on the kingside — two wings, two clocks, one decision. Through the 1990s, 9.b4 — the Bayonet Attack — became White's sharpest tool: it pushes the c4-pawn forward, opens room on the b-file, and keeps the c5 break and Nd2-b3 in mind. Black's best-known reply is 9...Nh5; after 10.Re1 Nf4 11.Bf1 the knight plants itself on f4, Black plays ...f5, fire builds in front of the king, and both sides sprint at their own attacks. These lines have been a memorisation war for years; thousands of games in the database, an engine's best move on every square.

The DVD's proposal is simple. When White plays 9.b4, Black does not enter the forced complications of 9...Nh5; Black plays 9...a5. The argument: this path contains "very few concrete lines", so the memory load drops, and the game drifts into a strategic struggle. The document further claims computers often do not know how to handle such positions. Here is my first yellow light. The claim is attractive, but behind it there is no engine evaluation, no average centipawn loss, no win-draw-loss rate, no sample games.

The real job of 9...a5 is to put the question to the b4-pawn. Black pushes at White's most advanced pawn and sets White before two unequal paths. One, 10.bxa5 — White opens the a-file; Black's rook on a8 becomes active, but the b7-pawn falls behind and the support of d5 loosens slightly. Two, 10.Ba3 — White holds the tension; Black then chooses the order of ...axb4 or ...Nh5, and exactly when the structure changes becomes the question of the game. Neither is new. 9...a5 is not a novelty, it is an old practical tool, an alternative door set beside 9...Ne8 and 9...Nh5. The DVD therefore claims no novelty — it claims to open a road to survival.

Strategically the picture is clear. Black's plan is to fix the queenside pawns and raise a storm on the kingside: ...f5, ...f4, the g5 push, opening lines for the queen's raid. White's plan is the mirror: the c5 or b5 break on the queenside, Nd2-b3, a4-a5 to loosen the squares in front of Black's king. This is the race of opposite wings: whoever arrives first wins. 9...a5 buys Black some time in that race, because White's memorised lines go dead. Here lies the DVD's real value: it does not return White from memory to plan, it returns Black. For a club-level player that is the true benefit — one idea instead of ten memorised moves.

My objection to the engine claim has two layers. First, "modern engines do not know" is the wrong sentence. An engine evaluates every position in numbers; after 9...a5 it will give either equality or a slight edge to one side, because material balance in this structure is genuinely even. What the engine cannot measure is human practical difficulty — the memory load, the clock, the fear of the first mistake. Second, the claim has no engine example behind it; no concrete position is shown to say "here the computer makes this error". Without evidence the claim is not false, it is half — and you cannot build a repertoire on half a claim.

"Very few concrete lines" equals safe — that equation needs breaking. Even with a closed centre in the King's Indian, the pawns never go quiet; when Black plays ...f5 and White breaks with c5, both kings come under attack on the same board at the same time. A supposedly "strategic" position therefore turns sharp at any moment in practical play. The document's own risk list admits this: the technical claims lack data support, and the sample is not enough for the "soundness" claim. Only one real protection exists — practice. The player who plays 9...a5 on the strength of memory alone will see fire in front of his own king ten moves later.

"The lament was loud; the ledger was quiet; the ledger won." The crack between headline and body follows the same shape. The headline paints a certain picture of a diplomatic meeting; the body presents the probable claim of a commercial DVD. Timur Turlov is written as "FIDE President" — by the most recent official records in my hands, the office of FIDE President is held by Arkady Dvorkovich. If Turlov holds some other FIDE role, there is no problem; but if the office is written literally, that is a fact-check flag, not a player-related data point. Here the document admits its own gap: the first-stage analysis did not assess source quality or time sensitivity, so several conclusions carry a reduced-confidence tag. "Empty stadiums taught me which noises were ever real" — the same lesson again. The headline talks loudly, the body sits quietly, and the truth of the game hides in the body.

Looking forward, three things deserve watching. One, whether 9.b4 use is rising at club and online level — if it rises, demand for practical doors like 9...a5 will rise too. Two, whether the DVD's line holds up in real games — the only proof of holding is a concrete game, not the language of marketing. Three, whether the headline gets corrected. My watch-list once held the account of a ten-year-old who played plans instead of memory; that child taught me where my model had gone lazy. The DVD is selling exactly that lesson — the only question is who will prove the rest.