9...a5 and the Trap of Belief: When an Old Move Is Sold as an Antidote to the King's Indian Defense

One November evening in Chengdu, with the city already deep in winter, I reopened the Mar del Plata database I had spent years compiling. On my second screen was a newly released opening course, selling the move 9...a5 as an "antidote" to the Bayonet Attack in the King's Indian Defense. The product page promised exactly what every amateur wants to hear: less memorization, few forced lines, and a balanced position where "if Black knows what to do, things quickly become dangerous for White." I read that line twice, then typed 9...a5 into my own search box. What came back left me sitting still for a while.

The King's Indian Defense is one of the most classic hypermodern systems in modern chess: Black willingly concedes the center early, only to counterattack it later with ...e5 and ...c5. Its sharpest and most famous branch is the Mar del Plata variation, where Black plays ...Nc6 and ...e5, producing a cross-board attacking race: White masses on the queenside, Black masses on the kingside, and both sides sprint to see who mates first. This is a land of pure calculation, classified within the ECO range E97 to E99.

The Bayonet Attack — 9.b4 — is how White grabs queenside space immediately, pushing the b-pawn to threaten opening lines for the pieces. It is the most fashionable and irritating choice against Black players of this system, forcing them to produce a precise, prepared answer. The orthodox reply, regarded by the theory community as the "main move," is 9...Nh5 — it opens the position, forces concrete lines, and returns the game to the terrain computers analyze best.

The Mar del Plata was once the weapon of world champions. In the engine era, the King's Indian's share of elite classical games has declined structurally. It survives at the top level mainly as a "surprise weapon" rather than a mainstay, because it concedes space and invites sharp play that has been theorized move by move.

The course I was watching proposed another route: 9...a5. This is not a novelty. It is a branch long known in Western databases, ranked secondary to 9...Nh5. The product's selling point is not freshness but practicality. It does not claim 9...a5 is theoretically stronger. It claims 9...a5 changes the character of the game — from forced lines that computers punish, to a strategic struggle with little to memorize, where both sides can keep the chances balanced.

This is familiar marketing in the opening-theory market. Sellers don't promise you will win; they promise you won't have to memorize. That is a claim about psychology and practice, not objective superiority. And for a segment of players — those tired of memorizing dozens of lines while engines hunt for holes — that promise is more attractive than victory. The product targets a specific identity segment: players who see themselves as "independent thinkers," not "memorizers." It is an identity-based sales angle — it sells not a move, but an image of you.

But there is a problem. The entire course supplies not a single verifiable number. No ACPL (average centipawn loss against the engine's best move), no win rate, no draw rate, no sample games cited. It speaks of an "antidote" without evidence that the poison has been neutralized. A tactical claim without accompanying data is just a belief, carefully packaged.

In forty-eight years of watching this industry, I have learned one principle. If a seller says a move "keeps the chances balanced," ask how many grandmaster-level games they checked. If they cannot answer, that is not theory — that is desire presented as fact.

I recall the winter of 2026, when every tournament was suspended and I sat digitizing thousands of old games into spreadsheets. I discovered that winners often do not win because they do something new, but because they understand their chosen plan more deeply than the opponent understands it. The empty stadiums of 2026 were the most perfect laboratory sport ever accidentally created, and they taught me that faith in a plan is only worth something once the plan has passed verification.

Now to the counterintuitive part. The course builds its argument on an outdated premise: that "computers often do not know how to handle these complex strategic positions." This sounds reasonable to amateurs, but it belongs to the pre-AlphaZero era. Today's neural-network engines handle the long-term strategic positions of the King's Indian far better than the ad implies. Selling a move with the argument "computers don't understand it" is a sign of a dated mindset. It may appeal emotionally, but it is no longer technically true.

The more interesting point lies elsewhere. The promise of "few forced lines" may be partly right, but the Mar del Plata is inherently sharp ground. No branch removes calculation entirely. A well-prepared White Bayonet player can still find play, and once 9...a5 becomes popular, a White-side response will almost certainly appear within one theory cycle. That is the law of the board: every "antidote" breeds an "antidote to the antidote."

The product's real value likely lies at the club and IM level, not at the top. For club players, having a clear plan for Black that depends less on memorizing dozens of lines is a genuine benefit. For a 2700 grandmaster, it is just another entry in the database — and one his opponent has already read.

I also note a small but telling detail about source quality. The original text's headline refers to a team event in Samarkand — "Round 2 (open)," with "all top teams grabbing wins." But the entire body below discusses an opening course. The two texts are unrelated. For someone in my line of work, who verifies by data, this mismatch is a red flag. No one should cite a source that does not know what it is talking about. And if you read this product elsewhere, check which part you are actually reading.

Broadly, this course sits in one of chess's most stable markets: opening theory. It sells answers to popular lines. It survives because there is always a player who wants a "weapon" in the bag, something to bring to the board that the opponent has never seen. Everything on the board is data waiting for a reader — if you take the time to sit down. But this product competes directly with free databases and free engines. Its survival depends on the written narrative and curated structure, not the raw lines. Tellingly, the very argument "computers don't understand" becomes a point of differentiation — a marketing asset, not a technical fact.

So where is the problem? Not in the existence of a course. The problem is whether the buyer can tell "a legitimate branch" from "an antidote." The move 9...a5 is a legitimate branch. It is not a solution. That difference — between a choice and a promise — is where many players lose hundreds of study hours.

I no longer believe in miracles on the board; I only believe in the conversion rate of advantage, measured across thousands of real games. If someone wants to sell you a move, demand data. If there is no data, you are buying belief, not theory. And belief, in chess, does not count for points.