9...a5 — A Small Path Through a Sky the Machines Have Already Drawn Shut

In March 2026, at a café on Nguyễn Chí Thanh Street in Đà Nẵng, I sat behind a man of about sixty. He played Black, opened the King's Indian Defense, and on move nine he put his knight on h5 — exactly by the book, exactly as in every game he had ever replayed online. Thirty-two moves later he signed the scoresheet, stood up, and said something I have carried with me for two years: “I memorised everything, and I understood nothing.”

Outside, the city kept its noise. Inside, the click of the pieces on the wooden board sounded like studs on an empty stand. A boy of about twelve sat at the next table, phone in hand, replaying the old man's game on an analysis app. The screen glowed blue, a red arrow pointed at move eleven. The boy whispered: “Wrong, uncle.”

The old man did not hear. And I sat there, in a Đà Nẵng café, realising something I thought I had known for a very long time but had never put into words: the machine has taught us how to move the pieces without teaching us how to understand them.

Then I read a promotional document for a chess DVD. It promised something very human: an escape from memory. And that escape has a name: 9...a5.

Why one small pawn move becomes a large question

The King's Indian Defense is the strangest system in the chess opening library. Black plays 1...Nf6, 2...g6, 3...Bg7, voluntarily handing White a broad centre, and then spends thirty moves attacking that very centre. It is an opening of faith: faith that structure matters more than space, that tempo matters more than squares controlled.

The Mar del Plata Variation is the heart of the classical King's Indian. The name comes from a 2026 tournament in Argentina where South American and Soviet players put the system on the world map. After 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, the centre closes and both wings open. White pushes pawns on the kingside, Black pushes pawns on the queenside, and the only remaining question is who arrives first.

White's ninth move, 9.b4, carries a name as beautiful as a bayonet thrust: the Bayonet Attack. White does not wait. White pushes the b-pawn to b4, prepares b5 to freeze Black's knight on c6, and opens the way for the a-pawn and the rook on a1 to sweep down the queenside. It is an act of aggression before Black has finished developing.

Black's orthodox answer is 9...Nh5. The knight swings to the rim of the board, eyeing f4, eyeing White's g-pawn, and more importantly clearing the way for 10...f5 — a punch straight into the middle. From there the position becomes a race: White pushes c4-c5, Black pushes f5-f4, both sides charge at each other, and almost every move is forced.

That is why 9...Nh5 has become a club player's nightmare. To a machine those positions are beautiful: each side has a single sequence, and one tempo off means losing everything. To a human, it is an unending memory test.

Based on my experience following games at Đà Nẵng clubs and on online chess servers, most amateur players do not lose because they lack strategic ideas. They lose because they cannot remember move twenty-two.

An opening DVD steps into exactly that space. The document I read proposes 9...a5 as a substitute remedy for 9...Nh5. The promise sounds attractive: avoid the forced lines, enter sound positions with very few concrete lines, and — in the advertisement's own words — engines often do not know how to handle those positions.

Such a promise deserves dissection. Not because it is wrong, but because it is half right.

What 9...a5 actually does on the board

Start with geometry. After 9.b4, White has just opened a new front on the queenside. Black has three choices: charge the centre with 9...Nh5 and 10...f5 and accept the race; retreat by closing the queenside; or play a small pawn move on the edge of the board in exchange for structure.

9...a5 belongs to the third group. Black does not argue with White about speed. Black argues with White about space.

The core idea sits here: pushing the a7-pawn to a5 invites White to decide the shape of the queenside himself, and every choice White makes has a price.

If White stays quiet and continues on the kingside — 10.Nd2, 10.Bb2, 10.Be3 — Black will open the queenside with 10...a4 or prepare ...b6 with ...Ba6. White suddenly finds himself defending the very wing he had just announced he would attack.

If White pushes 10.b5, the queenside closes. The b5-pawn locks the c6 and c5 squares, but it also locks Black's own a5-pawn. Black loses space but gains something more valuable: a lock. After 10...Nh5, Black can still play ...f5, can still play ...Nf4, and the queenside becomes a quiet zone for regrouping without fear of a stab in the back.

If White trades with 10.bxa5, Black recaptures with ...Rxa5 and gets an open queenside with a lonely but hard-to-eat a5-pawn.

But pawn chains carry another question with them: who has to read ahead. In 9...Nh5, both sides are forced to remember. In 9...a5, Black switches modes: Black has to understand.

More concretely, Black needs to grasp four things in the follow-up plan.

First, the knight's placement. After 9...a5 and 10.b5, the knight on e7 no longer has a reason to jump to h5 immediately. It can stay put, go to g6 and back to e7, or redirect toward f5. In many games Black chooses ...Nd7 and ...f6, breaking the centre from the outside rather than striking at the kingside.

Second, the kingside. Black still needs ...f5, but the tempo is different. It is no longer about striking before White can breathe, but about striking once the centre is held. After ...f5 and ...f4, White may open the queenside later and change direction, so Black must keep enough pieces back to defend.

Third, the rooks. The f8-rook in this system often travels via f7 or f6 depending on the structure. If White has closed the queenside with b5, Black can use the a8-rook to pressurise the a-file.

Fourth, and here I think the advertisement is right without saying so: Black's opponent is both the person sitting opposite and his own memory. At club level, Black wins through structural understanding, not through memorised variations. Someone who understands why the Black knight wants f4, why the g6-pawn is strong when it stays put, will play 9...a5 far better than someone who has ground through the first twenty-five moves of 9...Nh5.

Now the reverse side. Why is 9...Nh5 still the main line?

Because in the main line White is tied to a precise sequence. Every Black move creates a concrete threat that forces White to respond. It is a race Black can calculate at home. In 9...a5, Black trades for freedom but shifts the burden onto understanding and on-table decision-making.

Put another way: 9...Nh5 buys you time at home with suffering at the board. 9...a5 buys you comfort at the board with extra hours of thinking at home.

There is no free lunch. Only two kinds of bill.

That is why the phrase “engines often do not know how to handle” needs to be read slowly. Modern engines handle almost everything. What the phrase actually says is: those positions have no single answer, so humans cannot swap understanding for a database. That is a plus for human chess. It is also a minus for the DVD buyer, because what is being sold is no longer variations but thinking — and thinking cannot be installed in three days.

Here I want to tell a story. In 2026, when I was still organising amateur tournaments in Đà Nẵng, I had a group of extremely hard-working sixteen-year-olds. One of them memorised forty moves of a Sämisch line. He won for three months straight. In the fourth month, his opponent deviated on move three, pushed the game off the book, and he lost in twenty moves. That evening I asked him what he feared most. He said: “What I fear most is when the board refuses to play along with me.”

9...a5 does not promise to free you from that fear. It only changes owners. It makes you fear other things: making a decision without a book, judging a closed structure with no arrows pointing the way.

The half of the truth the advertisement leaves out

At this point I have to talk about the document as a text, not just as a piece of theory.

Its title talks about Samarkand, about some round seven of an open event, and about Uzbekistan continuing to dominate and extend its lead. The entire body of content talks about a chess DVD and the move 9...a5. Not one fact mentions Samarkand, a round result, or any Uzbek player.

I thought I was reading a report from Samarkand; it turns out I was reading a hole.

That mismatch is a signal about source quality. In chess information, headlines and bodies are often merged from two different sources into one file and forwarded without anyone checking. The result is a very attractive headline glued onto very different content. A reader who skims the headline will think Uzbekistan is winning in Samarkand. A reader who finishes will find himself misled.

And here I must tell the other half of the Uzbekistan story, because it is a true story, and it is far bigger than a DVD.

Uzbek chess does not need a false headline to be noticed. In Warsaw in December 2026, Nodirbek Abdusattorov — born in 2026 — won the World Rapid Championship at seventeen, beating one of the strongest fields in the history of the format. At the 44th Chess Olympiad in Chennai, India, in 2026, the Uzbekistan team took team gold — a milestone Central Asian chess had never reached since the break-up of the Soviet Union. Javokhir Sindarov and Nodirbek Yakubboev were part of that generation.

That is the news. A headline about Samarkand could carry a whole story about a country rewriting the world chess map. But that headline, in the document I read, was hanging on an advertising billboard.

If a source can let through an error like that at the top level, how reliable are its other claims — specifically the claim that “objectively, chances are equal”?

The honest answer: reliable as a statement, and no more.

That document provides no ACPL figure. No win rate. No draw rate. No engine match rate. Not one master-game sample. Not one database citation. The word “equal” there is a salesman's judgement, packaged in an analyst's language.

That does not mean 9...a5 is wrong. It means: you are buying a hypothesis, not a proof.

As someone who works as a commentator, I understand why people write this way. An opening DVD cannot be sold with a statistics table. It sells with a feeling: that you are about to escape a trap. And that feeling, in this case, is fairly accurate — the trap is real and the escape is real. It is just that the escape leads to another room that is also locked.

Here I have to put on the table a view I have held for a long time in this trade: numbers presented without provenance are not data, they are decoration. In chess, evaluation scores, move accuracy, engine match rates — all are useful when placed beside a real run of games with dates and names. Suspended in mid-air, they only make an advertisement look more scientific.

And one more thing about my trade. In 2026 an editor forty years my junior cut a piece of mine on the grounds that I was “writing literature in a sports report”. I left the newsroom after eighteen years, started my own blog, and rewrote everything from scratch. From that I learned one thing: a sports report must get the facts right, and must also say clearly where it does not know. If that document had written “we believe 9...a5 is equal, based on the author's experience”, I would have had nothing to complain about. But it wrote “objectively, chances are equal”. That word “objectively” is a promise nobody stands behind.

I thought the small move was an escape; it turns out it was a mirror held up to the place where I was too lazy to think.

The blind spot: when an advantage erodes itself

There is a paradox the document does not resolve, and I think it is the biggest blind spot.

If 9...a5 is genuinely good, it will become popular. If it becomes popular, White will prepare. If White prepares, the surprise value disappears. The advantage of a sideline lies in its rarity; it erodes itself through success. So anyone buying this product intending to make 9...a5 a main weapon is buying something with an expiry date.

The right way to use it, in my view, is to place 9...a5 inside a set. It is one option alongside 9...Nh5, 9...Nd7, 9...Ne8, depending on the opponent and on how prepared that opponent is. In club chess, opening preparation usually works reactively: people know what you like and look for a way to counter it. Variety is the weapon, not one particular move.

I also think there is a group of players who benefit most from this product, and it is not the strongest ones. It is players who have already played the King's Indian for a few years, understand the structures, but are tired of memorising lines they will never reach in a real game. For them, 9...a5 is a way to cool down. For a grandmaster with seconds preparing, it is one more line on a list.

And for a beginner, 9...a5 can be another trap: it gives a feeling of safety without giving a skill. You avoid a memory test and receive a judgement test. If you have never played a closed queenside position where you must find a plan yourself, you will get lost inside the very freedom you just won.

I also allow myself to look at this product through another lens, that of a youth coach. In chess as in football, there is a kind of overuse that is hard to see: pushing young people into a system far too mature for them. A thirteen-year-old is taught an adult's opening arsenal, then sent into games where one wrong move loses everything. His mind has not been prepared for that tempo. In chess, that tempo is memory.

9...a5, seen from this angle, is a healthy reminder: the board still has room for the person who understands slowly but understands firmly.

Signals worth tracking

In the end, a judgement about an opening line cannot be sealed inside one article. There are four signals I will be watching in the coming months.

One is the frequency of 9...a5 in real games, among players rated roughly 1500 to 2300. If the frequency rises clearly, the product's value rises, but its surprise value falls.

Two is White's results in those games. If White's win rate climbs, or if new refutations appear, the DVD's central claim weakens.

Three is whether the original source corrects the mismatch between headline and body. If it does, credibility is restored. If not, I will keep reading such documents with a question mark at the top.

Four is whether a top player uses 9...a5 in a classical game. That is the test every sideline must pass to step out of the twilight zone.

Of those four, the third has nothing to do with chess at all. It concerns us — the readers, the writers, the forwarders of information. And honestly, it is the one I care about most.

One move, two questions

There is something I did not tell you at the start of this piece.

The man in the café on Nguyễn Chí Thanh Street, the man who had memorised everything and understood nothing, played again. Three weeks later I met him in the same place. He did not put his knight on h5. He played a small pawn move on the queenside, slowly, and lost. He smiled and told me: “This game I understand why I lost. The last one I didn't.”

If 9...a5 has any real value, that value is not in the score. It is in forcing Black to understand what he is doing. And in an era when every opening line has been drawn shut by machines, forcing a person to understand is a rare gift.

But that is exactly why I do not believe the word “objectively” in the advertisement. Chess has no free objectivity. Every choice is a trade: memory for judgement, surprise for clarity, hours at home for hours at the board.

So the question is no longer whether 9...a5 is good. The question is which kind of bill you are willing to pay. And perhaps, for someone once exiled from the profession he loved simply for daring to write differently, that is the only question worth asking over a chessboard.

I thought I had been exiled from chess; it turns out chess led me deeper into myself.