Don't Mistake an Empty Input for a Clean Article: The Pipeline Failure That Quietly Shapes Cricketers' Futures

I start with a small technical incident that will sound far from cricket — but it changed the way I work. Some years ago, at a digital sports desk, I built an automated scraping tool that pulled hundreds of match reports and team announcements a day into a database. One morning I noticed records with 'N/A' titles, 'N/A' sources, and empty summaries. My first reaction was: 'Fine, these must be small stories, nothing in them.' Three weeks later, when I filtered those records, I found that a large share were reports about domestic players who had shone in the Ranji knockouts — exactly the data that was the spine of my work, quietly converted by the system into 'no information.' I raise this in a cricket context because over the past few years the Indian selection apparatus has become rapidly data-driven. Franchises have analytics teams, boards have performance tracking, and media have automated feeds. When a system cannot read an article and treats it as empty, it is not being neutral — it is quietly deciding that there was nothing in that article. And when hundreds of such articles are treated as empty, a player who should have surfaced vanishes from the data map. I have a concrete example. In 2026, coaching a Delhi under-16 side while freelancing in analytics, I built a 1,200-cell batter-versus-bowler grid from 60 IPL matches. Before the final I wrote that the Mumbai Indians vs Rising Pune Supergiant match at Hyderabad on 21 May would turn on whether Pune needed more than 40 off the last four overs against Jasprit Bumrah and Mitchell Johnson. Pune needed 11 off Johnson's last over, and Mumbai won by one run. After that day I stopped writing about momentum and started writing about matchups. But the point is not my grid. The point is that part of the data needed to build it was coming through the same automated pipeline that sometimes returned empty output. My model was only as reliable as the data fed into it. That is the lesson I applied everywhere after — in cricket commentary and in data system design. Look at the technical diagnosis behind this kind of failure. When both title and source show 'N/A', the first possibility is not that the article had no information. The first possibility is that the system could not read the article at all. Title and source normally come straight from page metadata — the HTML title tag, Open Graph tags, the canonical link. Both going empty at once almost always signals that what reached the system was either empty or not the body text. The second signal is when classification succeeds — the system knows this is a cricket-Asia domain article — but the body extractor fails. That means the classifier was running on something else: an article ID, a URL, or domain metadata. The body extraction step failed separately. The third, and cleverest, signal is that the summary is empty while author stance is 'N/A'. Empty and 'N/A' are different failure modes. That means different stages of the pipeline received different states of input — some stage saw an empty string, some stage saw a default value. This is not the story of 'the article was too short to yield a stance.' This is the story of 'some stages received no content at all.' The most likely causes: a JavaScript-rendered page, a paywall or login wall, an article selector mismatch, an encoding problem. And above all of these, the possibility that frightens me most — a wrong record reaching Stage-1 through task routing. You may ask what this has to do with cricket. It has a direct link. Today a large share of domestic performance in Indian cricket is recorded through digital feeds. Media coverage of Ranji Trophy knockouts is thin — I have experienced this myself. When coverage is thin and the system reading that thin coverage also sometimes returns empty output, the name of a player who scored a knockout century never gets recorded at all. And the selection committee, the franchise scout, and the analytics team all decide from that same incomplete data. My claim is not that every selection is data-driven. My claim is that when data goes missing, that absence works in someone else's favour. A player already in the headlines has data from many sources — TV, newspapers, social media, the franchise's own team. A player outside the headlines has data from only one or two thin feeds. If that feed fails, his existence is erased. That is not bias — it is infrastructural bias. Here I want to step back and write the strongest counter-argument to my own thesis. Someone will say: 'This is overblown. Real selection still happens through eye-test performance, coach opinion, and team balance. Data is only an aid.' And honestly, that argument is largely right. In Indian domestic cricket, regional scouts, old coaches, and people present at the ground still hold decisive sway. I come from that world myself. But that argument holds on one condition — that the people deciding have access to that player. When a selection committee receives a list of 40 players and 8 names are missing only because their data came through empty in the pipeline, then even the eye test never reaches those 8. Let me state the condition under which I would abandon my position. If any board or franchise had an independent, manual audit process that checked every season how many performances went unrecorded in the digital feed, and that audit were public — I would accept the problem is under control. I have not found such a public audit yet. Now the part that troubles me most. This kind of failure makes no noise. The system does not crash, no error popup appears. An empty string is simply converted to a default value 'N/A', and the record moves on. Whatever analysis happens downstream treats that record as a 'no information' article and discards it. No one learns that a player's name was never in the system at all. This is exactly like the empty stadiums of 2026 changing the sound of captaincy. That season I watched the IPL across three UAE venues with no crowds and wrote that the silence changed the nature of captaincy — bowlers could hear field changes, and Sharjah's short square boundaries inflated six-hitting by design. Then 19 December arrived: India 36 all out in 21.2 overs at Adelaide, Hazlewood 5/8, Cummins 4/21. My editor wanted 800 words on the players' state of mind. I wanted a taxonomy. I took three weeks, rewatched every ball, and built a four-category model — seam movement, indecision, scoreboard pressure, technical fault — and decided that no collapse piece after that would be written without assigning every wicket a category. The inverse of that discipline is also true. When data comes through empty, my entire framework for categorising a collapse stops working — because the evidence needed to assign a category has been quietly removed by a system. My taxonomy is only as strong as its input. Here is another counter-intuitive point that has troubled me for a long time. We generally assume more data yields better selection. My experience says more data is better only when its failures are visible. If data failure stays hidden — as 'N/A' — then more data means more hidden failure, and models built on it are less reliable than before. This is the same mistake as explaining Spain's 2026 World Cup exit by saying 'the bowling wasn't good enough.' On 1 July at Luzhniki, Spain completed 1,029 passes and lost to Russia on penalties after a 1-1 draw. I wrote 2,400 words arguing the failure was geometric, not stylistic — Spain's possession traced a U around Russia's five-man low block, and the half-spaces stayed empty for 120 minutes. The real lesson of that piece was not that Spain played wrong. It was that a statistic that does not show an empty space will tell you a winning story. The same applies in cricket. If your data does not show which performance went unrecorded, your data will tell you the story of the most reliable selection — when in reality it may be telling you the story of the most convenient one. Now to the solution. It is not complex. Every stage of the pipeline needs a hard check: if the title is empty or the list of information points is empty, label that record 'extraction_failed' and do not pass it downstream. Do not call it 'no information.' Call it 'information not retrieved.' The distinction looks small but it changes the whole system. Saying 'no information' lets the system move on and the player disappears. Saying 'information not retrieved' makes the system stop, send an alert, and have a human check what was missed. The second need is downstream. At the end of every season, someone should measure how many of the dozens of domestic tournament reports that appeared in the media were recorded in the digital feed. If the gap between coverage and data is above 20 percent, the selection process has a structural blind spot, whether anyone admits it or not. On the first major knockout series of the next domestic season, my eye will be on whether any team, when announcing its squad, said which players were picked on data and which on eye-test performance. The day a franchise publishes how many empty records appeared in its scouting database during the season, we will know this was not just a technical flaw. If that does not happen, a year from now another article will be written under this same title — and in it, too, no one will know which century was buried under an empty string. I regret that my own first system did something similar, and it took me three weeks to catch it — when a single warning message would have sufficed. That lesson is bigger than cricket.