9...a5 Against the Bayonet Attack: An Antidote With No Data Behind It

An opening product for the King's Indian Defense is being sold with a striking promise: very few concrete lines, almost nothing to memorize, and enough firepower to neutralize the Bayonet Attack — the 9.b4 branch that has become White's fashionable choice in the Mar del Plata system. The notable part sits somewhere else. Across the entire sales copy there is no ACPL figure, no win rate, no sample game named, and no author. The seller commits to an outcome while offering not one piece of quantitative evidence a buyer could check. I have spent most of my career reading documents like this. Years ago, running a small betting-analysis blog in Chengdu, I settled on a rule I still use today: a claim with no number behind it is not a prediction, it is a pitch. And a pitch never has to answer for being wrong. Data never lies, but it does enjoy testing our patience. Some background is necessary. The King's Indian Defense begins 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. Black concedes the center early, then reclaims it with ...f5, ...f4 and a kingside attack. White answers by pushing b4 and opening the queenside. The structure is named Mar del Plata, after the 2026 tournament in Argentina where the first defining games were recorded. By my own classification, cross-checked against public databases, this branch sits in the ECO E97–E99 range. Those are opening codes, not figures published by the manufacturer. The move 9.b4 is the Bayonet Attack. White grabs queenside space and forces Black to decide immediately. The orthodox reply is 9...Nh5, dropping the knight toward f4 and accepting that the center will collapse within a few moves. That branch opens the position, forces the lines to be played out, and turns the opening into a memorization contest with computer assistance. The product on sale offers something else: 9...a5. It strikes directly at the b4-pawn, keeps ...f5 in reserve and steers the game toward a closed, maneuvering struggle. The advertising says so plainly: instead of the forced lines after 9...Nh5, the buyer gets sound positions with very few concrete lines. What is being sold is not a new move. 9...a5 has sat in databases for decades. What is being sold is a change in the character of the game — from an arena of memory to an arena of positional thought. In an earlier era the King's Indian was a strategic weapon for the very top. Garry Kasparov used it as an attacking tool, turning a conceded center into a long-range calculation exercise. Based on my own experience tracking games at open tournaments and in data-analysis sessions for betting clients, the system's share of elite classical games has shrunk considerably since computers changed how players prepare. It survives as a surprise weapon, not a permanent pillar. A promotional DVD does not alter that reality. More specifically on the move being sold. After White pushes b4, ...a5 puts the queenside pawn structure under immediate question. White has options: trade on a5 to open a file, push b5 to lock the wing, or keep the tension and shift to a central plan. Each choice leads to a different type of position, and they share one trait — a markedly slower tempo than 9...Nh5. Black is no longer dragged into a race. In exchange, Black accepts a long game where positional skill matters more than recall. That is a reasonable proposition. It is not a new one. The most interesting thing to dissect is the promise about memory. In the Mar del Plata, Black still has to play ...f5, still has to reckon with ...f4, ...g5, ...g4 and with queenside defensive resources once White opens the b-file. Choosing 9...a5 slows the central conflict; it does not erase it. A buyer expecting to escape the volume of theory will meet it again, in a different shape and later. The volume drops. The complexity does not. Then comes the more dubious claim: computers often do not know how to handle these complex strategic positions. That was once true. It belongs to the period before AlphaZero's results were published in December 2026, when search-based engines struggled with positions requiring long-horizon judgment. Since then, neural-network engines and later the NNUE evaluation generation have changed the picture entirely. The human-versus-machine argument, in this form, is dated marketing language. The missing data is not a small matter. An opening product earns its price only when the buyer can answer three questions: where has this line been played, how did it score, and what is the average centipawn loss. The sales copy names no game, no win or draw rate, no ACPL. With no author named, there is no way to look up that person's record in this very system. An opening course that does not name its teacher is a hole, not a detail. In an empty stadium, data is the only audience left. The format is easy to guess. The register — antidote, very few concrete lines, for players who like to think for themselves — matches the Fritztrainer-style training DVD common in the German-speaking market and distributed through major chess platforms. This is the most stable and least glamorous segment of the industry: selling answers to popular lines to a built-in customer base. Its revenue comes from structure and author reputation, not from new theory. And the target customer matters more than the technical content. The pitch aims at players who do not want to be filed under memorizers. That is identity positioning, not data positioning. Buyers are not purchasing a branch; they are purchasing an image of themselves. This sells well in the advanced amateur segment, where study time is limited and the fear of a better-prepared opponent never quite goes away. The counter-view. Valued by theoretical standards, this product is cheap. Valued by the real needs of club players and international masters, it has a place. 9...a5 does not solve the Bayonet Attack; it drags the game out of a garden the opponent has memorized. That is a genuine value, wrapped in overstated language. The claim that things can quickly become dangerous for White if Black knows what to do is unfalsifiable — it is true of every opening line in existence. The biggest risk for any antidote is obsolescence. A secondary branch already named in databases will be analyzed back by White. The cycle is short: months, sometimes weeks, are enough to neutralize a plan. Nothing in the copy suggests the producer considered that scenario. And there is a more troubling source-quality signal: the original document's headline refers to a team-event round in Samarkand, Uzbekistan, while the entire body is about an opening DVD. Two unrelated texts were spliced together. For anyone who works with data, that is a signal to stop and verify before citing. What to watch sits in a few places. Classical-game databases of the 2700 elite will show whether 9...a5 appears in reality or only in advertising. Opening trees on major platforms will show whether the Bayonet Attack remains fashionable. If the line is genuinely played, a White-side response will appear within one theory cycle. That is when we learn who was right. I bet on numbers before the world learns how to read them. Here there is no number to read. And in chess, a promise cannot be measured in centipawns.