The 9...a5 Answer to the Bayonet: A Strategic Fix and a Data Void

In late 2026, I sat in a club room in Bengaluru, listening to a fourteen-year-old boy turn the pages of an opening book. He stayed silent for twenty minutes, then pushed his b-pawn to b4 and looked up smiling: "I've memorised thirty moves, uncle." I never forgot that sentence. In that room, a child's memory had become a weapon, and weapons carry a price.

Seven years later, I came across an opening product promoting the King's Indian Defence, Mar del Plata variation. It promised the opposite: a way out for those exhausted by memorisation. That is where this story begins.

Context: a variation that lives on memory

The Mar del Plata is the sharpest surviving branch of the classical King's Indian. Black concedes the centre early, builds the bishop on the long diagonal, then counterattacks with ...Nc6 and ...e5. White storms the kingside, Black storms the queenside, and both sides race to break through first. For decades this was a playground for people who enjoy calculating to the very end.

Then came the Bayonet. White's 9.b4 grabs queenside space and forces Black to choose: accept being squeezed, or walk into barbed wire. The orthodox reply is 9...Nh5, sending the knight to the rim, opening the position, and dragging both players into concrete lines. Memorise those lines and you live. Forget one move and you die.

The product I read proposes another road: 9...a5. Instead of opening the position, Black pushes the queenside pawn, challenges White's very b4-pawn, and gradually steers the game into a closed, manoeuvring struggle with few forced variations. The promotion says the chances are equal, and, most importantly, that the player need not remember much.

The analysis: promise and price

I read the product description carefully. There is not a single sample game. No move-quality metrics, no win-and-draw percentages from a database, no author name, no title, no record of the presenter in this very variation. There is only the claim.

In my trade as a documentary writer, a claim without data always deserves suspicion. It does not follow that the claim is wrong. The 9...a5 branch is known, not an invention. It has existed for a long time; it is simply chosen less often than 9...Nh5. Its appeal lies in practicality: it pushes the game into territory where both players must think for themselves.

This is where the marketing touches a real fear. At amateur level, the density of memory errors always exceeds the density of calculation errors. A club player may recall the first twenty moves and then stand at the board like a lost child. Selling them a low-memorisation system is selling them time.

The promise, however, goes further. The product asserts that computer engines "often do not know how to handle" such strategic positions. That phrasing is dated. Since the arrival of neural-network engines, machines handle long-term King's Indian positions far better than the sentence suggests. The wording belongs to the pre-AlphaZero era and has been kept on the shelf like old stock.

The contrarian angle: the flaw sits in the source itself

The thing that stopped me longest was not on the board.

The document I received has two parts. The headline reports the second round of a team event held in Samarkand, Uzbekistan, opening with the line "despite close calls, all top teams grab wins". The body discusses nothing but an opening product. Two unrelated texts have been spliced together.

For a reporter, this is the single most serious signal in the whole file. A document that contradicts itself in its first line means every fact after it must be re-verified. No result, no team, no pairing matches the headline. The Samarkand title most likely belongs to a different report, mixed in during aggregation.

Looking further into the body, another weakness appears. The phrase "very few concrete lines" is an overstatement. The Mar del Plata remains sharp by nature. No branch removes calculation. A well-prepared Bayonet player can still find play. The real value of 9...a5 most likely sits at club and international-master level, not at the level of the top grandmasters.

One more note on theoretical risk. An "antidote" move survives only until the other side has finished analysing it. In a branch as thoroughly trodden as the Bayonet, that life cycle tends to be short. Buyers should open a database and an engine first rather than trust the claim.

What remains

The home board has no grandstand, yet every move brings back a sky of memory. No grandstand applauds a correct move, and no grandstand remembers a wrong one. A player sits alone with thirty-two pieces and whatever they carry in their head.

Based on my experience following matches, products like this neither make nor break a variation. They only change how people study it. What I took home after reading: when a solution is sold as liberation from memorisation, is the buyer purchasing an opening system, or purchasing the feeling of no longer having to remember?

In a place with no grandstand, that feeling may be the most expensive item on the board.