The Score of Zero: How an Empty Feed Becomes a Full Story

In June 2026 at Edgbaston, India versus Pakistan in the Champions Trophy was reduced to 48 overs. The ticket cost £28. India made 319/3 — Rohit Sharma 91, Virat Kohli 81 not out — and bowled Pakistan out for 164 in 33.4 overs, a 124-run win. I charted all 202 balls of that chase by hand: field position, bowler type, shot. Back home I wrote 3,000 words with scanned field maps. Forty thousand people read it in nine days. The thing that bothered me most about that exercise was not a wicket. It was an empty column in my own spreadsheet. On some deliveries the seam position simply was not visible on camera, and I left those cells blank. When I later turned those columns into a graph, the spreadsheet treated the blanks as zero. The graph rendered. It was handsome, it was clean, and it was lying precisely in the region where I had no information at all. I counted the same sequence thirty times before the stand went quiet, and every time the same lesson came back: the dangerous number is not the wrong one; the dangerous thing is a space that looks filled while being empty. Eight years later that problem returned to my desk at far larger scale — not inside a spreadsheet this time, but inside an entire analysis pipeline. Cricket's data supply chain runs like a relay. Ball-by-ball feed, then the tracking system in the stadium, then the broadcaster's graphics desk, then the newsroom, then the person writing the headline off the scorecard. Every hand-off carries one assumption: what the previous hand gave us is complete. No station checks whether something actually arrived or whether the envelope was empty. The Stage-1 result delivered to me was exactly that empty envelope. Here is what was in it, stated plainly, because this is the real match of this report. Article title: N/A. Source: N/A. Article type: unclassified. One-sentence summary: blank. Author stance or purpose: N/A. Information points: an empty list. Entities involved: could not be extracted, because they derive from information points. Time sensitivity and source quality: not assessed. It is worth stopping here. Stage-2's entire job — identifying entities, fact-checking, inferring stance, judging how fresh a story is — is born from one thing: the information points. When that list is empty, every floor above it stands without a foundation. The difficulty is that it still looks like it is standing. The output I received carried no error message, no warning, no red banner. Only a very tidy table in which every cell said N/A or said nothing. This failure has three properties, and all three are easily understood from a cricket field. First, it is silent. An empty list is a valid data structure. No exception is raised, nothing crashes. Cricket is the same: figures of 0-0-0-0 on a scorecard are perfectly legal, though they cannot tell you whether the bowler was good or simply never bowled. Zero and absence look identical, and that is the deepest deception of all. Second, it takes a formatted shape. The most dangerous property of any dashboard is its design. When a table has eleven rows, eight filled and three reading N/A, the reader's eye does not stop on the three empty cells. It reads the eight and assumes the table is true. That is exactly what happened to my 2026 spreadsheet. Third, it compounds. Each successive stage adds a layer on top of the previous output — a conclusion, a belief, a draft headline. The first layer was empty, yet by the fifth layer you can be handed a complete, fluent, credible paragraph none of whose sentences traces back to a primary fact. I have watched this repeatedly in my own work. In any long story, the most dangerous sentence is not the one you can check; it is the one you never feel the need to check. The remedy proposed in this report is technically simple and professionally necessary. Put a hard assertion at the Stage-1 exit that stops the output before it travels further if it is incomplete. Four checks: a title exists and is not N/A; a source exists and is not N/A; there are at least three information points; and the one-sentence summary is filled. If any check fails, raise an error rather than proceeding. The threshold of three is not arbitrary. One information point is an anecdote. Two might be coincidence. Three form a sequence for the first time — and a sequence is the first thing you can actually verify. I learned that rule by applying it to myself. I do not declare a pattern until its threshold is fixed in advance and the sample size is printed in the copy. Three deliveries are not a pattern. And when the sample is small, the bravest sentence available is: this is not yet a pattern. There is something cricket learned long ago that data pipelines have not. Cricket names absence. A scorecard says did not bat, retired hurt, absent hurt; a result says no result. Every absence has a category, a name, and therefore an interpretation. That pipeline had no such vocabulary. It had only N/A. And N/A is not a name — it is a shrug, a word that claims the information has been verified when in truth the information never arrived. I have seen the inverse too, and it matters for balance. 19 December 2026, Adelaide, India bowled out for 36. Josh Hazlewood 5/8, Pat Cummins 4/21. There was no shortage of data that day. Behind every wicket was a delivery, behind every delivery a length, behind every length the lacquer of a pink ball behaving differently in that light. I wrote that night frame by frame in Thirty-Six, six thousand words. No empty cells there. It was a scorecard full of data missing exactly one thing: panic. And panic cannot be measured, so it cannot be stated — only shown. The difference is that on that night the absent thing was in the game, not on the scoreboard. In this pipeline the absent thing was on the scoreboard and nobody counted it. Now the part where I disagree with the common view — and I will state that view in its strongest form first, because an inversion deserves a fairly stated consensus. The common view is: more data, better analysis. Two dozen cameras in the stadium, tracking on every ball, GPS on every fielder, an angle on every shot. That view is not wrong. It is the reason so much of what gets asserted in the commentary box has already been disproved by the data desk. The data desk won the tactical argument; there is no doubt about that. But the danger no longer comes from a shortage of data. It comes from better formatting of the absence of data. Bad numbers stay in the commentary box — they are hard to turn into three publishable paragraphs. An empty cell becomes a paragraph very easily, because nothing contradicts it. The biggest failure here was not technical but editorial. There was no one who opened the envelope and said: this is empty, so I am not forwarding it. An analyst's job is not only to count, not only to interpret, not only to write — an analyst's job is also to refuse. That is the most important property of a model, and I learned it from empty seats rather than from any press conference in a crowded room. A model is only an argument that consents to being checked, and the first condition of being checked is that it declares where it is incomplete. There is a second error to avoid here, because I have a habit of noticing that every good fix eventually fills up too. A gate that lets everything through becomes a ritual. Assuming a process works because it has never caught anything wrong is the easiest delusion available. Any gate has to be tested — deliberately feed it an empty output and see whether it blocks. A door that never closes is not a door; it is a picture of a door on a wall. And if the gate is catching nothing, the problem is its placement, not its number: the check belongs where absence enters, not at the end of the pipeline. My question for the next cricket calendar comes out of this, and it is about habit rather than technology. When a dashboard hands you a list of twenty stories and two of them have an empty space where the source should be, will you count those two, or read twenty and move on? Cricket gave did not bat a place in its scorecard vocabulary because the game knows that absence is information too. In that chart I re-ran thirty times, one column is still blank. I left it blank. Turning it into zero would have made my story faster and my work worthless.