King's Indian Defense and the 9...a5 Promise: Decoding an Anti-Bayonet Repertoire in the Engine Era

In a small office in Chengdu, I keep a notebook that never leaves my desk. It holds 2,400 games that I digitized over six months in 2026, when every tournament was suspended and I was forced to turn the pitch — and then the chessboard — into a spreadsheet. It was inside that raw data that a small move appeared often enough to make me stop: 9...a5, Black's reply to the Bayonet Attack in the King's Indian Defense.

What caught my attention was not the move itself. It was the way it is being sold. A recent DVD repertoire product advertises 9...a5 as an "antidote" for King's Indian players suffocating under modern opening theory. The problem is that after reading it closely, I could not find a single number. No engine match rate, no ACPL, no game database, not even an author's name. Only prose. And for someone who has spent 48 years measuring sport, prose is not enough.

Let us begin where nothing can be disputed.

The King's Indian Defense is one of the most famous hypermodern systems: Black concedes the center to White, then counterattacks it. After the moves 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.0-0 Nc6, we reach Mar del Plata — the sharpest variation of the KID, where both sides launch opposite-wing attacks: White on the queenside, Black on the kingside. It is a classic arms race, and like every arms race, it rewards whoever prepares better.

Then comes the Bayonet Attack, 9.b4. White advances a queenside pawn, gains space, and disrupts Black's rhythm. The move is not new — Garry Kasparov once turned it into a weapon in classic games of the 1990s — but it has become "fashionable" again, and it forces Black to produce a clear answer.

The mainstream answer is 9...Nh5. Black's knight leaps to the kingside rim, opens the position, and from here the game enters territory of forced, concrete lines that demand precise memorization. That is where players with an engine in their head — or on their screen — feel comfortable. It is also where theory is settled up to move thirty, and any error is punished immediately.

And 9...a5? That is a branch of a completely different nature. Black's pawn advances to attack White's b4-pawn, preparing ...c5 or ...Be6, and — according to the marketing — leading to positions that are "more strategic than tactical," with "very few forcing lines." The product claims that such complex strategic positions often confuse computers, and therefore favor the player who understands plans.

Sound familiar? That is the manifesto of an entire school: humans with plans, against a machine with a spreadsheet. And as someone who lives by spreadsheets, I have a few things to say.

First, this is a practical repertoire pitch, not a theoretical refutation. The distinction matters, and marketing tends to blur it. A theoretical refutation says: "I can prove that 9...Nh5 has a problem." A practical repertoire says: "I will take you to little-known territory where you can fend for yourself through your understanding of plans." Both have value, but they are two different products. The seller here is selling the second — while calling it by the name of the first.

There is nothing new in this. In football, it is the story of inflated tactics: "The 2026 World Cup did not create pressing; it merely stripped the mask off those pretending to press." Anything can become a product if people know how to package it. Chess is no exception.

But let us be fair. The move 9...a5 is not an invention out of nothing. It has appeared in top-level practice as a space-gaining, tempo-gaining move, attacking the b4-pawn and preparing to open the wing. Its value is real — for the right audience. For club players who want to understand plans rather than cram lines, and who face the Bayonet less often than professionals, it is a reasonable tool. The product does not deceive anyone on that point.

The problem lies elsewhere: the technical superiority of 9...a5 over 9...Nh5 cannot be verified from the material. There is no engine match rate, no ACPL, no database statistics. Only assertion. In my trade, an assertion without data is a promise, and a promise must be paid for with time.

This is where I apply my own rule: everything on the pitch is data waiting for a reader — if you are willing to sit down. In chess, the "pitch" is the board, and the data is the moves. I sat down. I searched. And I found something more interesting than the move itself: a split within the KID community.

On one side are the "engine warriors" — those who accept that at the highest level, chess has become a memorization race with computer assistance. To them, 9...Nh5 is the only serious choice, because that is where theory is established and every mistake is punished instantly.

On the other side are the "human strategists" — those who feel suffocated by forced theory and seek territory where plans matter more than memory. To them, 9...a5 is an escape, a way of saying: "I do not want to play the computer's game."

The very existence of a product like this is itself data. It proves that there is demand — a group of KID players who feel cornered and are willing to pay to get out. That is a market signal, not a technical one. And the two should not be confused. A best-selling product does not prove a move is strong; it only proves the fear exists.

Now comes the part I must say plainly, even if it is not pleasant for the seller.

The product's central claim — that computers often cannot handle complex strategic positions — is weakening over time, not strengthening. This is the paradox of every "anti-computer" manifesto. Stockfish, Leela, and modern engines do not merely calculate tactics; they evaluate strategic KID positions reliably. Every year, the gap between "human understanding" and "machine understanding" in this type of position narrows. What was true in 2026 has become less true in 2026.

In other words, the seller is relying on an advantage with an expiration date. And that date is running out. I have seen the same pattern in other sports: a claim that once carried weight, then was overtaken by technology, then became a nostalgic belief sold to those who have not yet caught up.

There is one more issue, structural rather than technical: format. A static-disc repertoire product has a fatal weakness — it cannot be updated. If, after publication, a line is refuted by an engine, the buyer still holds the old disc while theory has moved on. In an era when opening knowledge flows through continuously updated subscription platforms, a frozen DVD is obsolete from the day it is printed.

I have seen this in other disciplines. An analysis correct at the moment of publication can become an error after a single season. People do not pay for the truth; they pay for the truth that is still valid. And a valid truth needs to be kept alive, not canned.

So, should we throw 9...a5 away?

No. That is a lazy conclusion, and I do not do this job to be lazy.

Here is what I believe: a move does not have to be the strongest theoretically in order to be the most useful practically. 9...a5 has real value for the right person, the right level, the right goal. The problem is only in how it is positioned — as a universal "antidote," when it is in fact a specialized tool for a small segment. If the seller were honest about that, the product would be better, and buyers would be less disappointed. This is a lesson I have learned many times over 48 years: selling the right expectation matters more than selling the right product.

But I keep returning to the biggest gap. Throughout the material, there is no author's name. For a repertoire product, the author's credibility is the greatest asset. A DVD signed by a KID-specialist grandmaster carries entirely different weight from an anonymous product. The silence about identity is not a minor detail; it is part of the story, and usually the most suspicious part.

And I must admit something about the very document I am analyzing: it has a serious data flaw. Its title speaks of a European schools chess championship, while the entire body speaks of an opening product. These are two unrelated subjects. When a source contradicts itself at the level of the headline, I lower the credibility of the entire source — not out of malice, but out of principle. Data does not lie; the reader of data lies. And a misleading headline is the gentlest form of lying, and the most dangerous, because it is the hardest to detect.

Back to the larger question. Will 9...a5 actually become a mainstream choice at the highest level? The answer lies in the database, not in advertising. I will track three signals: the actual scoring of 9...a5 against the Bayonet at the 2600+ level, its appearance in recent elite games, and the overall frequency of the KID at the top. If all three move in the same direction, I will believe it. If not, I will continue to treat it as an interesting but unproven sideline.

This is how I work, and I have no intention of changing. I once told a young colleague that "I no longer believe in miracles on the pitch; I only believe in the conversion rate of chances." In chess, the equivalent miracle is a move advertised as a "secret." The equivalent conversion rate is game data. I still stand with the numbers.

But I do not want to end on skepticism either. Because behind all the advertising, there is something to respect: someone is trying to keep chess a human game. In an era when young players learn by copying engines, someone reminding us that plans matter more than memorization is a contribution, even if it is sold with an overreaching promise. Good intentions do not make an assertion true, but they do make the debate worthwhile.

I will not buy that disc. But I will reread my notes on 9...a5, and I will keep it in my database a while longer. Not because of the advertising. But because that is the only way I know to answer the real question: can an old move learn something new in a world that has changed?

The chessboard, like the pitch, never stops asking questions. Only its readers stop sitting down.