The Lesson of the Blank Scorebook: When Cricket's Data Pipeline Quietly Lies

The Lesson of the Blank Scorebook: When Cricket's Data Pipeline Quietly Lies Last week a file landed on my desk. 'Stage-1' was written at the top, 'Domain Label: cricket_world' at the bottom. That was all. The title was empty, the source was empty, the list of information points was empty. And somewhere in that emptiness was a line that took me back years — 'Entities Involved: identify from the information points above.' That is an instruction, not an answer. The system had mistaken its own directive for the value of a field. I read the file twice, then pulled out my old scorebook. The one from 2026-97, thirty-four days with Bengal. It too has a page that is blank. Rain washed out a day's play in Guwahati, the official sheet has those boxes empty, and my own book has them empty too, because I did not write a single ball that day either. Emptiness filled honestly and emptiness filled quietly are two different things, worlds apart. In cricket we have always treated the scorebook as truth. Runs, wickets, overs, averages — these numbers stand in the court like witnesses. But a scorebook never fills itself. Somebody fills it. Some scorer picked up the pen; some did not. Someone wrote it wrong and corrected it later; someone wrote it wrong and never felt the need to correct it. In 2026, when I was on the Ranji circuit, the official sheets reached the desk a week late, and half of them were wrong. I began keeping a separate book — 1,140 overs, ball by ball. That habit has never left me. Every number is checked against two independent records. It is slow work, it cost me two scoops, and it is the reason editors hand me the corrections page. Now the same thing is happening in the world of software, and cricket journalism is not immune to it. A Stage-1 file whose 'Information Points' are empty should have said: 'I found nothing, do not proceed.' Instead the system proceeded. It wrote 'Article Type: Unclassified,' meaning that when nothing at all was known, it saved itself by calling it unclassified. 'One-sentence Summary' was left empty along with fields like 'Author Stance: N/A.' And on top of it all, the Entities Involved field was filled with the instruction itself. This is more than a technical glitch — it is a habit, the habit of covering emptiness with some label rather than admitting it. In my profession we call it 'writing half a truth into the file.' This does not feel to me like something outside cricket. During the six weeks I spent in Delhi covering the 2026 match-fixing investigations, I built a 412-entry chronology — every deposition, every bank transfer, every board minute. That work taught me one thing: the most dangerous thing in a record is not a lie, it is a blank that nobody noticed. When a date is left empty in a file, anyone can later put their own story into that gap. And when a system starts filling its own gaps, the question is no longer whether the data is right or wrong — it becomes whose data it is. The clearest example in cricket is what we call 'official statistics.' Pick up any franchise valuation. Pick up any broadcast-rights figure. What percentage of those numbers carries a source? I sat through the IPL's first auction in 2026 with a printed purse sheet and no working definition of 'retention.' Two months later Brendon McCullum made 158 not out off 73 balls at Chinnaswamy and Kolkata Knight Riders beat Bangalore by 140 runs. I travelled with KKR the whole season; they finished sixth. I filed sixty-one roster updates that spring. The lesson was that the contract comes before the cover drive. I still maintain a roster spreadsheet — contracts, ages, overseas slots, injury status — refreshed before every press conference. This is why when an empty Stage-1 file reaches me, I do not feel like crying; I remember that blank page from 2026. The only difference is that the page had rain written on it — the cause was recorded. This new file has no cause recorded. There is just a label, 'cricket_world,' which tells you this is about the world of cricket but not what the matter is. In cricket journalism this habit has a direct effect. When a data pipeline starts producing structured output even from empty input, the journalist has two roads: either treat that output as truth, or go back and find the original source. The first road is fast; the second is slow. The first builds threads; the second builds trust. I have been on the second road for forty-two years, and I know it is the only one on which a byline survives. A byline is a promise, and the archive is where you keep it. But here is a counter-intuitive point I would apply to myself. It is easy to attack an empty file. It is clean moral ground — the technology is wrong, the journalist is right. The problem is that we journalists often do exactly what that system was doing. We too fill blanks with labels. We write 'according to sources' and fill the blank. We write 'insiders in the team reveal' and use the same auto-fill the Stage-1 was using. After 2026 I struck 'according to sources' from my own copy, because a claim needs either a document number, a date, or a named person. That rule made me the slowest writer in the newsroom and the one they call at eleven at night when a name is spelled two different ways in two different filings. So when a system is taught to stop on empty input, that is not just a software lesson. It is our profession's lesson too. The technical suggestion made for Stage-1 — 'if not stage1.get(k): raise' — has a journalistic form: if a claim has no document, no date, no name behind it, do not give it the first paragraph. Spike it, or go find the document. Now the question arises: where in cricket are such empty files running? I can count three places, and I am watching all three. First place: coverage of domestic cricket. There are many Ranji Trophy matches that get no independent report. The official scorecard gets uploaded, but the story behind it — who played through injury, whom the selectors ignored, how the groundsman prepared the pitch — does not enter the record. That is a blank. And a blank later gets filled by someone's agenda. Much of what I saw in thirty-four days with Bengal in 2026 never made it into any newspaper. It is in my own scorebook. This is where the blank lines are most numerous, and where the most stories are. Second place: reporting on IPL rosters and retention. Every year before the auction, a flood of 'according to sources' stories arrives — who will be released, who will be retained, whose contract is expiring. How many of these have a document behind them? This is why my spreadsheet runs. When a story comes that a certain player has been released, I first check my sheet — age, overseas slot, injury, contract date. If the sheet says something else, the story needs a second source. It is slow. But the auction room keeps its own weather, and every bid leaves a shadow — to read the shadow you must first know the direction of the sun. Third place: coverage of women's cricket, especially after the 2026 home World Cup. Here there is a new problem. How many of the WPL's broadcast promises, match counts, audience figures, advertising rates come with a source? Comparing with 2026 is easy, but that comparison is only honest if someone audits whether the promises made after 2026 are actually being kept. Nobody asked how many cities the second WPL season was played in, or what tickets cost in those cities. This is the blank that nobody wants to fill, because filling it cracks the celebratory story. I have a rule I made between September 2026 and June 2026, when I covered forty-seven matches without once entering a dressing room — the IPL in the UAE, a pink-ball Test in Ahmedabad that finished inside two days, and the World Test Championship final at Southampton where India lost to New Zealand by eight wickets. Back then my method was proximity — the nets, the bus, the tray of tea — and proximity was banned. I filed four months of flat copy before admitting the problem was mine. I wrote a two-page presence protocol for the desk: who calls whom, at what hour, what gets logged. It is deeply unglamorous and it held. My copy since leans on process — selection logic, workload, travel — because I can verify process at a distance and I cannot verify a mood from a hotel room. The same applies to that Stage-1 file. That file is a failure of process, not of mood. It did not learn to stop when the data was missing. And when a process does not learn to stop, it turns its own instruction into an answer. 'Entities Involved: identify from the information points above' — that sentence sits in some system today, can reach some report tomorrow, and a fan's tweet the day after. That is the sequence. I saw that sequence in 2026 when a bank transfer receipt had a blank where the date should be, and over the next six months three different stories grew in that blank. So what next? I will not predict which system fails next year. I will only say that the next time you read a cricket report — whether about an auction, or Ranji, or WPL broadcast rights — ask one question: what date, what document, what name lies behind this claim? If you find no answer, understand that you are looking at a blank page that someone did not leave blank — someone tried to fill it and failed. And someone should have stopped that failure before it reached you. In my scorebook that 2026 page is still blank. I never filled it. Because it rained that day, and rain is not a score. The scorebook stays honest with its blank line. The question is whether our data pipelines can be half as honest, or whether they will always cover it with a label.