9...a5 and the Samarkand Headline: The Empty Square in Chess's Book of Accounts

My editor cut all 23 of my diagrams from the first match report I filed from the Purana Paltan press room and kept 400 words of result. That was 2026. In the ninth-round game, a single queenside pawn break at move 24 had created 14 squares of permanent space, and none of those squares made the paper. I learned the press room by counting what no one asked; that night nobody asked which square the win had come from.

Eight years later the same fracture has landed on my desk, this time in machine language.

A headline says: Samarkand, Round 7 (open) — Uzbekistan continue their dominating run and extend the lead. That sentence is the old squeeze of the tournament cycle, where six rounds of accounting dissolve into the adrenaline of round seven. Under the headline there is no scoreboard for Uzbekistan, no pairings, no names, no prize structure. There is a DVD: an opening repertoire for Black against the Bayonet Attack in the King's Indian Defense, built around one move — Black's 9...a5.

That distance between headline and body is the actual story. Where a result is claimed, evidence should sit beside it; where evidence is absent, the thing is not a result, it is a catalogue. And when the language of a tournament report becomes indistinguishable from the language of a product's back cover, the reader's only defence is the habit of opening the file.

What the board is actually doing

To understand any of this you have to sit at the board. 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, White's ninth move 9.b4 is the Bayonet Attack: sharp, direct, and deliberately one-sided. White takes space on the queenside, threatens to squeeze Black with b4-b5, keeps the c4-c5 break in reserve, and can sometimes add f2-f4 in the centre. Black's plan is just as deliberate in the other direction: ...f5, ...f4, ...Nh5, a pawn storm against the king. This is the Mar del Plata complex — both sides attack on opposite wings, and the 7...Nc6 setup is the contract that makes that race legal.

Black's main reply is 9...Nh5, which steers into forced lines. That is the problem. Forced lines belong to computers, and computer lines belong to memory: elite preparation today is substantially a test of who has memorised more branches and who reached which database file first. Esports taught me that a meta is a nervous system with patch notes; opening theory works exactly the same way, and every engine generation rewrites the notes.

The DVD argues that 9...a5 is the antidote: fewer forced lines, more strategic positions, less memorisation. It carries two further claims. One, that computers often do not handle these complex strategic positions well. Two, that objectively the chances are equal.

The first claim speaks to a real pain, and I know it. When stadiums emptied in March 2026 and the Moscow Olympiad was postponed, the chess calendar my year was built around simply stopped. I did what an INTP does with anxiety: I went into the archive. Over eleven weeks I replayed 480 games from Bangladesh's 2026–2026 era and built a 1,240-position database tagged by pawn structure. The empty stadium taught me that noise is a variable, not a truth. The memorisation burden is just as real, especially for club and IM-level players who spend nights learning branches that an opponent never touches.

The second claim is a different species. The second claim is accounting.

The numbers that were never written

Start with the board. 9...a5 is not a novelty, and the seller does not claim it is; it is an established sideline against 9...Nh5. Its job is simple: Black attacks the b4 pawn directly. If White supports it, Black can open the a-file and connect rooks, later strike at the base of the d5 chain with ...c5, or fall back on ...f5. White keeps time and space — b5 can smother Black's play, c4-c5 can break through, and White may buy time to build a storm on the kingside.

The trade-off is clear. Buying flexibility costs Black some firmness. Abandoning forced lines also means abandoning the engine's help; the position then rests on your own understanding and your own arithmetic. That is not a bad deal. For a club player it is precisely the deal that saves thirty memorised lines a night. The question moves elsewhere: how good is this, for whom, and for how long.

And here the catalogue stops. There was no engine evaluation — no number, no build, no depth, no date. There was no game: no moves, no player names, no event, no year. There was no accuracy measure that could show how precisely White was playing or how often Black was slipping. What remains is a sentence: objectively the chances are equal.

Every transfer rumour is a hypothesis; the medical is the p-value. Chess has the advantage that the medical is public — databases and engines are open to anyone. What is needed is which build, which depth, which date, and how many games. Without those four, the sentence is not analysis. It is assertion.

I do not say this from arrogance, because the same rule bites me. I tell my co-commentators, and I now write it: I do not put a raw engine evaluation into prose unless I can explain the move in words. Why is this move so strong, which square belongs to whom, which pawn structure puts the weight on whose feet — if I cannot open that in sentences, the move is not mine, it is the software's.

The claim that is ten years behind

The computer-can't-handle-it claim is the weakest of the lot, and specifically weak on timing. Its premise belongs to the pre-NNUE era of engine marketing. AlphaZero arrived in 2026; from 2026 the NNUE architecture changed how engines read the board. Closed, strategic, structure-driven positions are now evaluated far more reliably than they were. In 2026, writing that computers do not understand these positions means writing a sentence about a decade old. In sales copy that is forgivable. At the point of purchase it should be noticed.

There is also an uncomfortable truth folded inside the claim. Engines can play those lines well; the problem is me. In a strategic position, the duty of choosing the right move falls on the human. The real market is not the engine's blindness but my own limits of accuracy. I trust a system that can explain its own failure in the 89th minute. A product that explains only the opponent's limits, and never its own, deserves to have its fine print read carefully. Tactics are spells with pressing triggers; the wizard reads the fine print rather than the headline.

One more practical risk goes unwritten. Strategic positions leave more doors open under time pressure. A line that rests on understanding breaks faster in rapid or tiebreaks, where a forced line would at least have been protected by memory. The player who trades memory for understanding is not buying time. He is borrowing it.

Where a headline is doing a table's job

Back to Samarkand. To write about Uzbekistan I need round seven's pairings, the standings, who is on which board, what each team's score is. Instead of those I was handed a DVD. Uzbekistan's chess rise is a real item in the international record — a dense cohort, state investment, the pressure of hosting opens at home. But to place that sentence in an analysis I need a table, not a guess. You cannot seat a scoreboard on a guess.

A headline is trying to do a table's job: give a fast, precise, final number — who leads, by how much. In a tournament cycle, this kind of sentence gives emotion the appearance of accounting. Lead extended — there is heat in it and no verification. In a tournament week, the audience wants a result and the media supplies the taste of a result; the wider the gap between the two, the more room a catalogue has to sit down in.

For comparison, one verifiable line. At the Chennai Olympiad in 2026, 188 open teams and 162 women's teams played; for Bangladesh, 78-year-old Rani Hamid held board one and scored 3/7 against players four decades younger. That line carries a name, a score, an event, an age, and the age gap against her opponents. It weighs more with me than a headline. At the 43rd Chess Olympiad in Batumi in 2026, Bangladesh's open team played 44 games across 11 rounds inside a 180-team field. In those two weeks I filled 212 notebook pages and filed nine dispatches nobody in Dhaka had commissioned. A notebook in Batumi can hold a World Cup if you fold it right; a report written without care cannot hold a single round.

Which brings in Bangladesh, and this is an accounting story rather than a sentimental one. In 2026 Niaz Murshed became South Asia's first grandmaster, a year before Anand. That file stayed open; the pipeline did not. India's count of grandmasters is now past eighty. In Bangladesh, the wait for a new grandmaster has run year after year since 2026. With the federation's annual budget hovering around Tk 11 lakh, and with Standard Chartered's school chess programme closing in 2026, nothing filled the gap: no permanent school line, no corporate patron, no published candidate list. Absence never appears as a line item, because nobody keeps the ledger.

The gap nobody is watching

The argument in the room is whether 9...a5 is good and whether engines can handle it. The real gap is not in the play but in the framing. A seller says computers do not understand; a headline says Uzbekistan extended the lead. In both cases the profit sits in the same place: if the reader does not verify, the profit is locked in. The training market sells the comfort of less memorisation; the news market sells the comfort of a trending word. The same reader buys both, and both times his only defence is the same: the habit of opening the file.

Here is the second uncomfortable point. A federation that does not publish its players, budgets, and programmes cannot argue with a headline, because it has nothing to lay beside it. You do not need to be a grandmaster to disprove a product's claim. You need a database, a date, a name, a budget line. Where a system gives you no candidate list, no school-programme statistics, no year's games, no declared budget, the language of the catalogue and the language of the federation's press release start to sound alike. The headline joins the chorus.

The quiet joke of the academy follows. One party sells the comfort of disengagement, another announces the void, a third summarises the trend. None of the three can show a game bundle, a pairing sheet, or a budget document. All three can show a confident sentence.

After eighteen years in this work I hold to one rule that came from the beginning: write the evidence, not the result. I learned the press room by counting what no one asked. Today the familiar question is back on my desk: where was this thing actually played? On which board, in which square? And who placed the name Samarkand inside the catalogue?

What to watch in the coming months

So what I will sit and watch is this: whether 9...a5 appears in the regular repertoire of high-rated players, or stays a club-level sideline while the elite keep playing 9...Nh5; whether an independent, verifiable file on Samarkand round seven is published somewhere, with names, a score table, pairings, and a prize list; and whether anyone opens Bangladesh's ledger in public — how many players stalled at which step since 2026, who received how much in which financial year, and what the training budget has put toward bringing chess back into schools.

If even one of those three sees daylight, the next lead will not have to be written from a guess. When the chess calendar freezes I keep moving pieces in the margins, and I treat the roar of a tournament cycle as a variable rather than a truth. Who will keep that ledger?