A Samarkand Headline, a Bayonet Attack DVD, and the Audit Nobody Ran

It is half past midnight. On the table in my small Penang flat sits a cup of tea whose contents went cold a long time ago. One browser tab stays open on the laptop for a single headline: Round Seven in Samarkand, Uzbekistan pressing on and extending its lead. Nine years of cricket coverage has worked itself into the muscle — if the headline carries a result, you click through and go looking for the scorecard.

I clicked. There is no Samarkand inside. No round seven, no tables, no player's name. What is there is the back-cover copy of a King's Indian Defence course — Black's 9...a5 against the Bayonet Attack, and, underneath it, the sentence: computers often do not know how to handle these complex strategic positions.

My notebook is called 'What the Scoreboard Cannot See.' Tonight, in the scoreboard's place, sits an advertisement for a DVD. The headline is news; the interior is retail — both without a single number, both on the same screen.

First, get the board's language right

After 1.d4 Nf6 2.c4 g6, Black posts a bishop toward the king and then punches the centre with ...e5: the King's Indian Defence. When White enters the Classical channel and lets Black play 7...Nc6, the version that appears is the Mar del Plata — both sides attacking on opposite wings, the clock running fast, the theory bottomless. White's chief weapon is 9.b4, the Bayonet Attack: take space on the queenside, press on c5, threaten to break Black's centre.

Black normally has two replies. 9...Nh5 — sharp, almost forcing, a line you memorise; one slip and the position is often gone. And 9...a5 — older, slower, a fight about structure. Play the first and you must memorise; play the second and you must think. The gap is a single move, yet the entire advertisement rests exactly on that gap.

One thing needs saying plainly, because without it the arithmetic goes wrong: 9...a5 is not a new discovery. It is an established sideline, part of recorded theory. The move being sold as an 'antidote' is an old horse wearing new paint. Antidote here is not a chess word, it is a shop word.

And yet the seller's pain is real. At club and IM level, memorisation fatigue is a genuine problem. A White player sits at home, asks an engine exactly where Black goes wrong after 9...Nh5, then sits at the board trying to pull that answer out of his own head while the clock runs. Whoever builds training material around that pain has grasped a real problem.

The trouble is that they are pointing at the wrong enemy. Because the two claims doing the selling are both made without evidence.

Claim one: the computer cannot handle it

'Computers often do not know how to handle these complex strategic positions' is a sentence roughly eight years old. In 2026 AlphaZero showed what a program could do in closed structures. In September 2026, NNUE — the efficiently updatable neural network — entered Stockfish 12. From that moment, evaluation in closed, strategic, slowly-frozen positions stopped being blind. Leela-based engines, the next generation of Stockfish, cloud engines: they recognise these structures instantly, because these are the very structures they have trained on millions of games.

In other words, 'engines cannot play strategically' is a pre-NNUE marketing slogan that is technically dead. The slogan survives in the training market. The position does not.

Claim two: objectively, chances are equal

'Objectively' is not a feeling. It is a measurement. To speak objectively about a move you need three things. One: a current engine evaluation — where does the position sit after 9...a5? Two: a game-database sample — at what level, how many games, how many of them stayed equal? Three: a picture of human accuracy, an ACPL-style figure that tells you how badly the move breaks in human hands.

None of the three is here. No evaluation, no game sample, not even a single game reference. A sentence that says 'equal' without a number is not an assessment, it is a statement. And a claim made without proof returns without belief.

Who wrote the course, what title they hold, what their own record in the King's Indian looks like — not one word of that is here either. To argue credibly for 9...a5 demands theoretical depth, and the only proof of that depth is the author's own games, their coaching record, their name. What is an antidote without a name? Simple question, missing answer.

This is where I have to pull in the arithmetic from my own table. In 2026, writing about Dhaka's club chess, I dug out the Bangladesh Chess Federation's annual government grant — roughly eleven lakh taka, for a full year. Less than a single franchise cricketer's bat sponsorship. Two years later, in late 2026, I spent two weeks at the Dhaka Premier Division Chess League, where Police, Navy, Biman and the bank teams put Indian grandmasters on guest boards. One of the signings I tracked was a Kolkata grandmaster joining a bank side for a reported fee of about one and a half lakh taka.

My standing question follows me there: who paid for this board? Chess in Bangladesh runs on favours and personal ties, but those ties are traceable — who knows whom, who phones whom, which club honours which manager's word. Chess in the world market runs on courses and licences. Two ledgers, two languages. One counts years and thinks slow; the other counts overnight sales. The DVD speaks the second language.

The real opponent

Here is my main disagreement. For the player who wants to use 9...a5, the enemy is not the engine. The enemy is the clock, and his own half-memorised memory.

Picture the actual scene. A club player rated somewhere between 2026 and 2100, opposite a prepared opponent. Playing 9...Nh5 means walking fifteen to twenty moves along a mountain path; one slip and the position is gone, while the clock keeps beating. He remembers twelve or thirteen moves accurately, then doubt arrives, he stops to think, time drains. In the last ten moves the structure is broken, and across the table sits a less error-prone opponent.

9...a5 performs a small piece of contractor's work here. Objectively Black may drift into a slightly inferior position, but in exchange he buys the right to think. This is not an antidote to the engine — it is an antidote to the clock and to human memory. The seller has declared war on the engine. The real war is with the clock, and admitting that would not sell a product, because nobody buys a story about beating a clock.

What nine years of watching boards has taught me is this: positions are equal in the engine, and unequal on the score sheet. On the score sheet, the winner is the one who, before the last move, still has time left to count one more.

What everyone is missing

Shouting at the DVD seller leaves the story half-written. The mismatch between headline and interior is an economics story, not merely a dishonesty story. Chess media traffic festivals have gradually been taken over by course sellers. Touch a trending name like Samarkand and you get the search; inside the search you can sell a DVD. That is not a matter of contempt, it is a matter of structure — the way cricket analysis was once written in the language of its sponsors.

An older habit is complicit too. We love high-level opening theory a little too much, and that is exactly where our arithmetic slips. The famous wins after 9...Nh5 are carved into memory; the jewel lines of 2700-level preparation are archived. We forget that below 2026, most King's Indian defeats do not come from theoretical depth — they come in the last ten moves, from shortage of time, from stopping one move too early.

More uncomfortable still is the distance between the price of knowledge and the price of entry. In 2026, Bangladesh's Niaz Murshed became South Asia's first grandmaster, a year before Anand. Today India counts more than eighty grandmasters; Bangladesh counts five. Engines, databases, online courses — all within reach of a hand, and yet the grandmaster line advances in generations, not in decades.

That is my second disagreement: the false slogan is not mere advertising noise, it is a symptom of an empty entry path. Where nobody builds the lower steps, somebody builds a course. Where no board sits on a school table, a DVD sells.

In July 2026, the country's most decorated player, Ziaur Rahman, died at the board, and I stood outside the hall writing field notes. Five months later, in December, fourteen-year-old Manon Reja Neer won the National Championship undefeated. Holding both facts at once, I have never written a 'next Niaz' headline and never will. Instead I print an audit every quarter: who crosses 2200, at which event, on whose money. It is called Pipeline Watch.

Names matter here. A move may be called an 'antidote' and no harm done. A generation should not be named at the price of one course.

What I will be watching

One thing is worth keeping an eye on in the coming months: whether 9...a5 returns to elite boards. If the line lives in the preparation files of 2700s, it stands on its own feet; if it does not, it remains a product of the course market. The database will tell us. The back cover will not.

Alongside that I will watch the demand in the training market — 'understanding, not memorisation.' A DVD may be sold on a flawed argument and still answer a real need, and where the need is real, the supplier eventually changes.

And closest to home, I will watch my own country's entry path. If Neer and his contemporaries climb past 2200 within fifteen years of age, we will know the picture is turning. If they do not, another course will sell, and we will click another headline.

The last word belongs to the clock. In a chess hall it is not the engine's numbers that win, it is the clock's numbers; the ledger that carries the true account is not the same paper as the advertisement. The best columns are built like a rehab plan: small steps, honest pain, no shortcuts. And what the clock says before the last move, the score sheet never says at all.