9...a5: The Move of the Planner in a Game the Engine Is Steering

In a corner of a small Saigon café, the man sitting across from me placed a pawn on the board and then fell silent for a long moment. He did not look at me, only at the board. A half-finished King's Indian Defense position lay on the wood, and he was weighing exactly one move: pushing the pawn from a7 to a5. To an outsider it looks like an act so small it means nothing – a pawn advancing two squares, capturing nothing, checking nothing, making no sound at all. But in his head an entire plan was opening up: freeing the rook, aiming at the dark squares the white pieces were occupying, and pulling the game out of the territory the computer had long since memorized. I stayed quiet and did not ask. I have learned that in the world of chess, some moves are not born to win immediately, but to pull the player back toward the human part of the game.

The real story behind that move was just placed on the table again, through a new chess training product, and it lands squarely on the biggest crack in the contemporary chess world: the border between the player who understands plans and the player who merely memorizes the lines an engine has already charted. The people behind the product call it a practical offer, an escape route for thousands of club players whenever an opponent thrusts the Bayonet move into the middle of the board. To me it reads more like a letter sent in silence than like an advertisement. And like any letter written in silence, it must be read very slowly before you can hear what it is trying to say.

To understand why such a tiny move became a focal point, we have to step back and look at the whole forest of defense it belongs to.

The King's Indian Defense, known internationally by the abbreviation KID, is one of the oldest and most characterful defensive systems in modern chess. Beginning with 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5, Black voluntarily yields central space to White, gathers pieces at home, and waits for the chance to counterattack. It is the emblem of the hypermodern school: not to occupy but to control, not to press but to invite the opponent forward so he pours honey on his own feet. Its beauty lies in the fact that every piece has two jobs – to defend when needed, to strike when the moment ripens.

The Mar del Plata Variation is marked by the move 7...Nc6, opening one of the most violent two-wing battles chess has ever known. White piles pawns on the queenside, Black pushes pawns on the kingside, and both sides know that whoever arrives one tempo first will win. The peak of this variation is tied to the years when the world's leading players turned it into a familiar weapon, to the point that an entire generation of spectators knew every square on both wings by heart.

Then came the Bayonet Attack – White's thrust 9.b4. The name says everything: a bayonet fixed on the end of a rifle, driven straight into the queenside to seize space and smash the order Black was trying to build. The Bayonet is not an aesthetic move. It is a provocative one, and for years it forced Black to keep searching for answers. The most orthodox answer, the one played most at every level, is 9...Nh5 – pushing the knight to the rim to trade pieces, open the kingside, and drag the game into the whirlpool of forced variations that only an engine has enough patience to memorize.

That is exactly why a product launched recently deserves attention. Instead of following the mainstream current, it proposes a different medicine: 9...a5. A move strangely quiet for a system famous for noise. Instead of pushing the knight backward to defend, Black pushes the queenside pawn one square forward, right up against the white piece standing there, creating pressure on the opponent's pawn structure while preparing bigger things behind it: opening the rook's path, considering Bg4 or Be6, and above all – turning the game into a war of plans rather than a war of memory.

The product introduces itself with an argument that sounds very appealing to the café men like mine. It says 9...a5 creates positions that are objectively balanced but practically dangerous for White if the black player does not grasp the plans. It says this is a choice engines often do not know how to handle – because the position is strategic, with few forced lines and few technical anchors for an engine to cling to.

That is the sentence I want to linger on the longest.

After years of watching chess, I have learned that promises of this kind always have two halves. The first half is true: chess, at the human level, is mostly decided by who understands the position better, not by who has memorized more variations. A club player who has mastered the queenside plans can beat an opponent who knows a few engine lines by heart but loses his way the moment the position drifts off the page. The second half is the doubtful part: the claim that computers cannot grasp strategic positions has been steadily losing ground in the current era. Today's strongest engines evaluate King's Indian positions no worse than the most knowledgeable human, and sometimes point out hidden resources an entire generation of players overlooked.

Unfortunately, the product offers not a single number to back its promise. There is no engine match rate, no average centipawn loss per move, no statistics from a top-level game database. There is only prose, and prose is beautiful but cannot be verified. This is what any journalist watching the chess training market must say plainly: an offer without data may still be right, but the reader has a right to know it is being sold on belief rather than on evidence.

The most valuable thing in this product is not the move 9...a5, but the data gap surrounding it.

And when an idea is sold through a gap, people start looking at the gap more than at the idea.

Let us set the advertising aside for a moment and look at the move itself.

Using 9...a5 is not an invention out of thin air. Throughout the history of the variation, Black has repeatedly used queenside pawn thrusts to gain space and tempo, forcing White to split his attention between two wings. The underlying idea is classical: when White piles pawns on the queenside, that queenside becomes a target, and every empty square behind an advanced pawn is a potential weakness. The move 9...a5 turns White's advance into a debt, and Black simply collects the debt gradually by increasing pressure on that structure.

The novelty, if there is one, lies in the packaging. It is not sold as a theoretical variation for contesting an advantage, but as a practical tool for sidestepping forced lines. That is an important psychological difference. A theoretical variation promises truth; a practical tool promises safety. And in the club chess world, the second promise always sells more easily than the first.

Looking at the group this move targets, we see a familiar portrait: the strong club player, someone who does not have eight hours a day to memorize the long rivers of engine lines, someone who faces the Bayonet less often than professionals do. For them, the core value is not winning through theoretical superiority, but entering the game with a map of understanding rather than a dictionary of memorization. Seen that way, 9...a5 works like a map: it does not point to the exact square, but it sketches the terrain and lets the player walk it himself.

One interesting thing is that this move poses a small philosophical question in the heart of chess. If a move is chosen not because it is strongest, but because it makes the opponent think more than it makes us think, is that a technical choice or a psychological one? The answer probably lies in this: chess has never been only technique, even though at the highest level it is almost nothing but technique. At the human level, in small rooms, on old boards, the human factor still claims a not-insignificant share of the final result.

That is also why the division the product reflects is worth recording. Chess is splitting into two streams of people. One stream is the engine warriors – those who pursue the main, forced lines where every move has an optimal answer confirmed by the machine. The other stream is the human strategists – those who want positions with less theory and more ideas, where they can live on their own understanding instead of drowning in someone else's memory. The move 9...a5 leans toward the second.

The very existence of such a product is, in itself, a confession. It admits that a significant number of people feel squeezed by the wave of theory the computer blew up, that they are looking for an exit, and that they are willing to pay for a nominal escape route. There is a strange loneliness in this. The people who love chess most are sometimes the ones who feel their sport is slipping out of reach, turning into a contest between memory machines wearing human clothes.

I think about the stands. In chess, the stands are always different from football or athletics. People do not shout, do not clap in rhythm, do not see the decisive moment until it has passed. The chess stand is a space of held breath, of pencil scratching moves, of very slight nods when a beautiful blow appears. An empty stand still has an echo – you only need to be quiet enough to hear it. And products like this one, though sold to a small group, are still part of that echo.

The closer I look, the more I realize the real problem lies in the format, not only in the content. Chess training products have long been distributed as physical discs – a series of lectures packaged, recorded, and sold. That format has a structural flaw that cannot be fixed: it freezes knowledge at the moment of publication. If an engine later finds an improvement, or a top-level database records a game that changes how a variation is seen, the disc buyer has no way to update. He clutches a map that may have been obsolete from day one.

The chess training industry recognized this long ago and is shifting strongly toward subscription platforms and interactive databases. There, knowledge can be corrected, supplemented, and expanded in real time. A variation is rewritten the moment someone finds a hole. That is a fundamentally different model: instead of selling a static asset, you sell a dynamic relationship. And in the chess world, where everything changes weekly, the dynamic model always wins.

If we place the new product in that context, we see it rowing against the current. It defends a correct philosophy – understanding plans matters more than memorizing lines – but wraps that philosophy in a fading format. It is a beautiful and sad paradox: the content speaks of human flexibility, while the form is as rigid as a page already printed.

There is one more thing that made me stop, and it belongs to journalism, not to chess.

News of this product circulated with a strange headline. It was attached to a European schools chess event, to a place name in Armenia, to images of young champions and a cultural celebration. But when you open the body, all it contains is an opening-repertoire product for the King's Indian Defense. Two unrelated subjects, stitched into a single cover. Any editor who saw that would have to frown.

I say this not to catch a product out, but to point to a disease of the information age. When speed of circulation matters more than accuracy, people join headlines and content like two puzzle pieces from different pictures, trusting that readers will not notice. Most indeed do not notice. But in chess, where everyone is in the habit of checking every move again, that inconsistency leaves a small scratch on the publisher's credibility. And in a small community, where good news travels as fast as bad, a small scratch can widen into a crack.

Now let us return to what I consider most important, and also most easily overlooked.

The product's central argument – that computers cannot grasp complex strategic positions – was once true, but is losing weight year by year. Today's strongest engines are no longer machines specializing in calculating forced sequences; they have learned to evaluate static positions, to understand pawn structures, and even to propose long-term plans that humans once considered their own domain. When the foundation of a claim erodes over time, the claim does not become false immediately, but it grows more fragile every day.

This leads to a strategic consequence for the seller himself. If you position your product as something that fights the computer, you are betting that the computer will stop improving. That is a bet no one has ever won. Conversely, if you position it as a relief depot for the club player who wants to be practical, to save time, to have a playable line that is easy to remember and hard to beat, then the product stands on much firmer ground – because the practical needs of humans do not depend on the development speed of any machine.

The true value of a practical move lies not in beating the engine, but in helping a human stand firm against other humans.

After all, most games in a life are played not by two engines but by two people. Two people with limits on time, memory, concentration, and fear of a position they do not understand. And in that setting, a move that is objectively balanced but practically annoying can be stronger than a move that is objectively strong but easily misunderstood.

When I told this story to a friend who coaches young chess players, he stayed silent for a while and then said something I never forgot. He said what he teaches his students is not moves, but how to sit at the board when they do not know what to do. Products like this one, in the end, are trying to do the same thing on a larger scale: to teach a posture, not just a sequence of moves.

And I realized this is why I still sit with such products, even though they rarely provide the evidence I want. They say something about how humans are trying to keep their share in a game where machines have grown too good. Generations will pass, but the poem about them keeps being written. The club players, the café men weighing 9...a5, the people who never appear on the news feed but still patiently play move by move – they too are a generation, even if no one names them.

I do not know whether 9...a5 will stand the test of time. Perhaps in a few years the database will show it neutralized by some White improvement. Perhaps it will quietly disappear from theoretical tables, like so many other ideas once praised and then forgotten. But to me, that matters less than the question it raises: in a sport where machines keep getting better than humans, where will humans choose to stand?

The man in the café finally placed the pawn on a5. He looked at me, smiled very slightly, and said: this is the move I play when I do not want to think too much. I did not ask more. I wrote the sentence down, because I knew it would remain true longer than any line of theory written after it.

Saigon at night still hums with traffic. The man packed up the board, placing each piece into a cloth bag, careful as if storing something more fragile than a game. I stayed alone, looking at the empty board, and I heard – very clearly – the echo of games without spectators, of players whose names are never called, of a tiny move that may not even be right but was certainly thought through. Perhaps that is what chess always keeps, no matter how strong the engine becomes: a seat for the person who listens to the position rather than obeying the numbers.

And I left the café carrying a small, stubborn belief that the day the computer understands everything on the board will still not stop humans from sitting down, placing a hand on a pawn, and asking themselves what story the next move should tell.