9...a5 Against the Bayonet: Auditing the King's Indian Branch Where the Engine's Claim Loosens

A team-event scoresheet landed on my desk last year: White pushed the b-pawn to b4 on move nine — the Bayonet Attack, the fastest door into the King's Indian Defence's Mar del Plata complex. Black played ...Nh5 without hesitation, then ...f5, and for the next twenty moves the board held not chess but a memory contest. White burned forty-two minutes on the clock; Black thirty-five. By move fifteen both men were still walking a road they had carried from home, and nobody in the hall could say who stood better. The first clue was a scoresheet nobody thought worth keeping.

That evening a review copy arrived — a DVD in the 'Svitlana's Smart Moves' series, promising a quieter answer to 9.b4: 9...a5, a line where you need less memorisation and more thought. The question is not whether that promise is true. The question is whether there is a single number behind it.

Context: what the Mar del Plata is, and why the Bayonet unsettled it

The King's Indian bargain has not changed in four decades. Black concedes the centre, strikes with ...e5, then throws pawns at the kingside. White gets space; Black gets time on the other side of the clock. The Mar del Plata — named for the 2026 Argentine tournament where the structure crystallised, filed under ECO E97–E99 — is the sharpest version of that bargain: ...Nc6, ...Ne7, ...f5, and mutual attacks on opposite wings. From the 1950s through Kasparov's era it belonged to players who wanted fire on the board.

Then came the Bayonet, 9.b4. White throws the b-pawn forward to slow Black's kingside schedule, gains queenside space, and trains c5 on Black's centre. The Bayonet's real strength is not attack but delay — it pulls Black off the timetable on which the Mar del Plata's beauty depends. Since the 1990s it has been White's most popular weapon here: less book, more pressure.

Black's main reply is 9...Nh5, the principled route to ...f5 and ...Nf4. Its price is explicit: long forced sequences, little space, and a demand for precision on every move. And one structural fact frames this entire debate: at elite level the King's Indian has contracted over the past decade. After neural-network evaluation became standard around 2026, White began extracting edges even in the quiet, semi-closed structures once considered engine-blind. The opening did not die — it lives at club level, in team events, in the hands of players who want to win from equality. It is no longer a daily visitor at 2700.

That contraction is exactly where a product like this finds its market. An opening that leaves the top trickles down — and the trickle needs explanation. The Olympiad supplies the raw material: a biennial event with well over 150 national teams, public scoresheets, and a prestige word for the title. But Olympiad games carry a selection bias few reviewers mention. In team matches a draw is often sufficient, and result-discipline is the worst thing that can happen to opening material — the games where both sides choose safety early cannot tell you who is better in the line.

Core: what 9...a5 actually does

Three jobs, in order. First, space on the queenside: the b4 pawn has left its base, and ...a5 targets it while threatening ...a4 to push it further and break the rhythm of White's queenside play. Second, a different knight route: instead of ...Nh5, Black heads for ...Na6–...Nc5 or ...Nd7–...b6, which is a genuine departure from the Mar del Plata's kingside template. Third, delaying the ...f5 decision until Black chooses it rather than having it forced. The move buys Black time — and that is both its strength and its weakness. Strength, because at club level time is the scarcest currency; a player who cannot recall twenty moves of the main line can still play ...a5 on understanding. Weakness, because White gets the same time, and a Bayonet player will spend it on c5, Be3 and queenside pressure.

The product's framing is that the Mar del Plata produces rich strategic positions where engines do not see far enough and human understanding gets work. That sentence is a century old and still arguable. What is not arguable is that the picture changed. Since 2026, NNUE evaluation has made engines substantially better in closed, manoeuvring structures; 'the computer doesn't understand this' was a 2026–2026 heuristic, not a 2026 fact. More tellingly, the product's own description undercuts its claim inside a single paragraph: it says engines cannot handle these positions, then concedes that after 9...Nh5 'computer analysis becomes important again.' That is not a contradiction but an admission — the product knows the engine is competent here, and sells discomfort rather than incapacity.

Now the part every review should lead with and this one buries. There is not a single number in the material: no engine evaluation, no average centipawn loss, no engine match rate, no database frequency for 9...a5, no results split, no named model game. A product making technical claims without technical evidence is the story here. 'Objectively equal' plus 'dangerous for White' is a familiar pairing in opening literature — reassure the buyer that the line is not refuted, then promise him chances. Both can be true. Without data they are assurances, not claims.

I read the description three times looking for the presenter's credentials. The series title gives only 'Svitlana' — no full name, no title, no indication whether she has played 9...a5 in serious classical chess. For a product whose entire pitch is 'learn from my experience,' that is the evidence that matters most. Where verification fails, a reporter records the blank rather than filling it.

There is also a terminology error in the title itself. The King's Indian Attack (1.Nf3, 2.g3, 3.Bg2, d3) is a different opening from the King's Indian Defence described here. In a search-driven market this cuts both ways: it attracts Attack players who will find nothing they wanted, and it misses Defence players who were the buyers.

Scope: one node of a large tree

The material answers 9.b4 only. In the Mar del Plata, White retains independent and still popular routes at move nine — 9.Ne1 and 9.Nd2 — plus earlier deviations such as 8.Be3 and 8.Re1. An anti-Bayonet module is not a Mar del Plata repertoire, and the gap between those two things is where buyers lose money. The product is not at fault; the perception the title creates is.

This matters in India specifically. The Tamil Nadu pipeline — the Tal Chess Club lineage, the Velammal schedule, the state's cash awards — has produced players who learn to memorise enormous opening files very young. The real problem in Indian junior chess is not a shortage of talent but a surplus of memory. A 'think, don't memorise' pitch touches a genuine need, whether or not this particular product fills it.

And that is where the easy promise becomes hard. Who struggles most in a strategic, non-forced position? The player who has stopped memorising but has not learned to decide. Sell 'few concrete lines' and you are really selling one decision per move — and deciding under pressure is always harder than recalling a rehearsed move. The promotional text itself says these positions demand long-term planning and tactical sharpness, which sits awkwardly beside the promise of less calculation.

Contrarian: this product sells permission, not a line

The buyer is not purchasing 9...a5. He is purchasing permission to stop memorising and still play the King's Indian. The real merchandise here is psychological relief, not skill — it is aimed at the player who has memorised thousands of lines and still gets stuck at the same club tournament every year.

So the important question is not whether 9...a5 is theoretically sound. It is whether the buyer can handle the position it creates. And the 'antidote' framing is itself the trap: once a line is published, it is available to the opponent. The moment 9...a5 spreads at club level, its chief selling point — few concrete lines, therefore less preparation — evaporates, because White can now prepare against it specifically. A secret weapon becomes a gift the day it is printed.

There is a charitable reading of the engine claim, and it is stronger than the literal one. Engines may evaluate these positions accurately now, but reading an evaluation and executing the plan over the board are different skills. The evaluation belongs to the machine; the plan still belongs to the human — and that gap is what actually rescues the product's argument.

Then there is the angle I find most interesting. Chess literature has always marketed the Mar del Plata in the language of attack — fire, storm, king hunt — as though the opening belonged to the brave and strategy were a second-class citizen. A product that opens the same opening in the language of strategy, presented by a woman, inverts an old classification — the one that told generations of women players they were 'positional' while sharpness belonged to men. Here the sharpest system in chess is taught positionally, and that is the selling point.

There is an economic layer too. Instructional content has become a modest but real livelihood channel for titled women in a sport where the patronage economy — PSU jobs, state grants, board medals — has historically done that work. Buying an opening DVD is not only buying moves; it is placing a small stake in a titled player's living. That does not change the review's direction, but it changes its accounting.

The largest risk is not refutation. It is that the buyer purchases the module and plays exactly as before. Chess content sells the way gym memberships sell: the moment of purchase is the moment of hope, and the moment of practice never arrives. The second structural risk is format: static media competing with subscription platforms that update weekly and bundle live databases. This product's shelf life will be ended by a calendar, not an engine.

And the industry signal is simple. Olympiad games cost nothing and carry prestige, so the value of any such product lies in explanation, not discovery. 9...a5 is not new; explaining it so a club player can use it is. If you already own a complete King's Indian repertoire, face 9.b4 twice a month, and your problem is recall, this is useful. If you are looking for a complete Mar del Plata answer, it is not — and that distinction belongs to the buyer, not the product.

Takeaway: what to watch

I went looking for a result and found a whole life — that of the player who meets 9.b4 every week and opens a database every next morning to see whether anything has changed. For him this is a small, cheap, possibly useful door. Not for the player who mistakes one module for a repertoire.

Four things will settle it. Engine evaluation: a clear improvement for White and this becomes another ageing line in a catalogue. Elite uptake: one serious classical game with 9...a5 and a good Black result restores interest. White's move-nine preference: a shift from 9.b4 toward 9.Ne1 or 9.Nd2 shrinks the addressable market. And publisher disclosure: the presenter's name, title, model-game list and runtime.

The next Olympiad's game corpus will reopen the argument — every cycle hands the industry free raw material, and every cycle a publisher must decide which story to build from it. If 9...a5 appears two or three times looking the way this product says it looks, the move earns a place at club level. If it looks merely balanced, the product stays a module: useful, but smaller than its promise.

Thirty years later the notebook still knows the sound — of that hall, of those clocks, of the man who said on move fifteen, 'I know this line.' The only question worth asking of any product is whether it takes that man past the line or deeper into it. A product that hands you moves leaves you where you stood. One that teaches you what to do when the moves run out takes you forward. The claim here is large and the data is small — but if it teaches one player that sitting at the board and thinking is a legitimate answer to the Bayonet, it has done its job. The question is whether the next scoresheet that reaches my desk will have anything new written on it.