When the Model Runs Out: A Data Audit, Not a Match Report

I have a spreadsheet I carried from Delhi to Liverpool. In 2026, while sitting my Class 12 exams, I built a 736-match Elo-plus-Poisson sheet and filed previews for a new-media site at 1,500 rupees apiece. My model gave Croatia a 14% chance of reaching the final; the market said 7%. They got there. I'm telling you this because last week I had another spreadsheet open on my laptop. It was completely empty. There was a title, just a title. And in every cell below, a single phrase: 'N/A — insufficient information.' What I had in front of me was the output of a Stage-1 deconstruction, and it was a text-extraction pipeline failure. No golfer's name, no tournament, no strokes-gained data, no course, no scorecard. There was only a domain label — 'golf' — and the rest was blank. For me, as a sports writer, this situation is not new. In 2026, during lockdown, when the Bundesliga restarted behind closed doors on 16 May, I set 218 closed-door matches against 1,142 fixtures with crowds. The home win rate fell from 43.1% to 33.5%; home goals dropped 0.31 per game. I wrote 4,000 words arguing the effect was real but small and badly confounded by fixture congestion. This is my process: sample first, its limits second, conclusion third. And if the sample is zero, the conclusion must be zero. In professional golf, data scarcity is a feature, not a bug. On the women's tour, ShotLink or strokes-gained data is often unavailable. At Tokyo 2026, when Aditi Ashok led through 54 holes at Kasumigaseki, I had hole-by-hole scorecards for 60 players, and that was it. I built a proxy and flagged her putting inside ten feet as an edge the market had at 200/1. She finished fourth at 15-under, one shot off bronze. But my point here is that if I hadn't had Aditi's scorecards — if I'd just had an empty file saying 'insufficient information' — I would never have been able to write that story. And if I had written it, it would have been fiction. Talking about data quality in golf is difficult because the sport suffers from information asymmetry. The PGA Tour has ShotLink, which tracks every shot. But the European Tour, the LPGA, and especially the Indian and Asian tours have much more uneven data. One player might have a full set of strokes-gained data; another might have just a scorecard. In my experience, this asymmetry is a fairness issue. If you do data-driven journalism, you have to acknowledge that your model will work better for players with better data. It's a bias I include in my work every time, but one I cannot eliminate. When a Stage-1 pipeline fails, it's a reminder to me that the data I have is the result of someone else's effort. Someone fills out a scorecard, someone runs ShotLink, someone builds a stats API. If that chain breaks, my entire analysis collapses. So when I look at this empty spreadsheet, I have nothing to say about golf. But I can say this: as a data analyst, my greatest strength is knowing when my model isn't working. And today, it isn't. I have nothing to look forward to except one question: if we don't take data quality in golf seriously, how many 'Aditi Ashoks' stay off our radar?