The File That Called Itself Tennis: One Wrong Label and Bangladesh Tennis's Three Silent Decades

Row 1511. The column headers read: Player, Surface, First-Serve Percentage, Break-Point Conversion. The values inside read: Brent at $105.52. WTI at $92.93. A benchmark spread of $12.83. Daily flows of 33.7 million barrels through the Strait of Hormuz.

I scroll. Where the surface field should say clay, grass, or hard court, it says diesel, refinery, political unrest. There is no player name. There is a head of state, two financial market analysts, and a Gulf military coalition. And stuck to the top of the file, stubborn and declarative: Domain Label — tennis.

My job is to read these files. For nine years I have built the invisible layer beneath the sport — the dataset that sits under whatever the scoreboard happens to show. The shoulder injury taught me that pain is just unstructured data waiting for a schema. That lesson is doing work today, but not on a match. Today I am not analysing tennis. I am analysing the failure of an analysis.

Context: nine dimensions, one lab, one stubborn label

My working framework has nine dimensions. Technical and tactical. Data and form. Tournament structure and scheduling. Tour landscape and player positioning. Rules and governance. Team and management. Risk. Media narrative. Industry transmission. Each dimension has its own table, its own indicators, its own declared information basis.

I built it that way because memory alone cannot carry the weight of a season — write it down or the seasons blur into story.

My lab is small. In 2026, at sixteen, a rotator cuff injury ended my junior career at the Barishal divisional training centre. I stayed in the sport by hand-logging serve percentage, unforced errors and break-point conversion for all 32 matches of the 2026 National Tennis Championship at the Ramna complex. One post showed the champion won only 54% of baseline rallies but 78% of net approaches. It circulated through Dhaka's tennis clubs. In 2026 I tracked xG and PPDA across all 64 World Cup matches and wrote before the final that France's 0.7 xGA per match — not Mbappé's speed — was the real story. France won 4-2. In 2026 I built a database of more than 500 matches played behind closed doors and found home advantage in football drops 32% without crowds while tennis serve percentages stay flat. In 2026 I profiled a 14-year-old named Zarif Abrar and predicted his breakthrough.

Today the same pipeline that carried all of that handed me a broken marble.

What the file actually contained

Information points one through nineteen, not one of them tennis. Brent $105.52. WTI $92.93. A $12.83 spread. US diesel at $6.528 a gallon — record territory. 33.7 million barrels a day through Hormuz. Houthi missile attacks on Saudi Arabia. A prospective US–Iran truce, sourced to the conventional wire formula of unnamed people close to the talks. Masoud Pezeshkian. Erik Meyersson of SEB Research. Tim Waterer of KCM Trade. Kpler. A possible US diesel export ban. A conflict running since "the end of February." A naval blockade. A closed strait.

And the label: tennis.

The first wound: where the label came from

Classifiers lean on keywords and batch processing. English news vocabulary reuses the same tokens across domains — strike, rally, service, break, breach. A keyword router is not stupid; it is fast, and speed is how this error gets made. In a production data pipeline, misclassification is not an if. It is a when. The real question is what the pipeline does after the label goes wrong.

The second gap: half-filled cells

The 'Entities Involved' field was never populated. It contains the instruction "identify from the information points above" — direction, not data. 'Time Sensitivity' is worse: "not assessed in Stage 1." A dataset that cannot tell you who is involved or how time-sensitive the item is cannot be a basis for any decision. In tennis or in oil, the rule holds. I learned it in transfer windows: every transfer rumour is a missing value that everyone reads as a price.

The third gap: the smell of the source

London dateline. No named outlet. And a scenario — a US–Iran war running since late February, a blockade, a closed Hormuz, record diesel — that I could not match against mainstream reporting. That is not proof the event did not happen. It is proof I could not confirm it. Confidence: medium, inferred from internal consistency without independent verification. Until provenance is confirmed, this document must not enter any factual dataset. A claim is not a record. Turning a claim into a record requires chain of custody.

Why this contamination hurts tennis more than anywhere else

The verifiable Bangladeshi player pool is roughly six names: Khaled Salahuddin, Sree-Amol Roy, Shibu Lal, Ranjan Ram, Jonathan Mridha, Zarif Abrar. In a six-name dataset, a single spurious association can rewrite the whole picture. In a large population one bad point nudges an average. Here it is a 16% rewrite. Zarif Abrar's 2026 J30 title — the first ITF junior title won by a Bangladeshi — is a generational signal, not a trophy. Put one unrelated geopolitical variable beside it and it will parade as a pattern for three years.

I will name my own trap here. 'Counter-intuitive discovery' is my brand, so my brain reaches for the contrarian read before the data has spoken — dismissing a first-ever ITF junior title as statistically trivial. The rule I hold: state the null hypothesis in one plain sentence, and only flip it when the evidence forces the flip. The null here was 'the pipeline is broken, delete the file, move on.' The data did not force a flip. So I am publishing the obvious read.

The cross-sport metric smuggling trap

My origin story is a World Cup xG experiment, so football-analytics vocabulary arrives pre-loaded. In 2026 I saw why it cannot simply be transplanted: crowds changed football's home advantage by 32%, and changed nothing about tennis serve percentages, because tennis home advantage lives in court speed, bounce and organised club habit, not in the stands. Cross-sport comparison is legitimate when you measure both games separately. Cross-sport metric borrowing is not. What is genuinely collectible at Bangladeshi domestic level is a short list: service hold rates at J30 events, tie-level Davis Cup records, junior draw entry lists — and for each, the name of whoever collected it.

The baseline: 2026 to 2026, then what

The National Championship launched in 2026. The 2026 Davis Cup Asia/Oceania semi-final followed. Then three decades of near-total absence from the record. I do not call those decades a talent gap. I call them a data gap. There is no evidence of missing ability, because the instrument was switched off. The schema was broken, not the players. Which is exactly why a mislabelled file worries me: I know how a broken schema makes a generation invisible. Which is also why micro-breakthroughs matter — Abrar's J30 title, BKSP girls sweeping domestic events, Mridha's roughly 508 career high, home Davis Cup ties in Dhaka that moved local tennis more than any talent hunt. With an explicit ceiling attached: no Grand Slam main draw, no top-100, no ATP title.

Contrarian: the instinct to delete

The mislabel is not the real finding. The real finding is that a mislabel is a symptom. A wrongly labelled file at least means someone tried to index something. Total absence means nobody tried. The instinct to delete is the same instinct that left three decades unrecorded. And the silent contamination risk is real, but the deeper harm is in the user's head: every unexamined number we accept builds a habit, and that habit will one day make us slow to recognise our own genuine signals. I also flag my own positional risk — born in Canada, working in Dhaka, I could audit a broken federation from above and flatten the club culture that actually keeps the sport alive. Ramna, Gulshan, Officers Club. The BTF-era figures who kept it running deserve quotation, not verdict.

Takeaway

Three fixes. A domain-confidence gate and keyword-consistency check before any Domain Label is committed. Treat unfilled 'Entities Involved' and 'Time Sensitivity' as failed extractions, exempt from scoring. And record who collected every number. Over the next six to eighteen months I am watching three signals: the service-hold range of Bangladeshi juniors at J30 level, whether the BKSP effect on domestic women's draws is structural or a one-season flash, and whether the sponsor-TV loop around home Davis Cup ties in Dhaka finally moves. If we cannot this week identify a file whose label bears no relation to its contents, what do we measure with — and where have we put the actual tennis?