The Empty Column Is the Honest One: Auditing Indian Golf's Numbers

March 2026, DLF Golf and Country Club, Delhi. The first European Tour co-sanctioned edition of the Hero Indian Open — 156 players, four rounds, and one column on the spreadsheet open in front of me in the media centre. The column was headed: approach shots from 150 to 200 yards. For the first three hours, the rows for the 24-man Indian contingent had nothing written against them. I did not fill them in a hurry; I let them stay empty, because every shot's landing lie, pin distance and wind reading had to be collected first. When the column finally closed after the fourth round, the number was unforgiving: that group of twenty-four was losing 1.8 strokes per round to the European entrants from that one distance band alone. Anirban Lahiri won that week, and the model showed his gains were concentrated on and around the greens. The real lesson of that afternoon was not the number in the column. It was the three hours the column spent empty.

Indian golf's data ledger did not begin that day. In 2026, three years into a sports-science contract at a Bengaluru performance lab, I put a one-page proposal in front of the newly formed Tata Steel PGTI: let me record every shot. Two interns and I charted 13 events and 1,064 rounds that season on clipboards and a battered PDA — fairways hit, greens in regulation, putts per GIR, scrambling. The result was clean: the top ten on the money list scrambled at 62 per cent, everyone else at 48 per cent. Driving distance separated nobody at all. The first truth of that season is that the first PGTI season lives in pencil, not in a database. A circuit whose founding season was never properly databased cannot be analysed — only re-derived, line by line, with the gaps left visible.

The 2026 model was built on that same habit: 156 players, four rounds, every shot logged from the media centre. The model had a hand-built spine before it had a dashboard. I filed the finding as the 'approach-play tax' for a business daily, and from that week I wrote two versions of every piece — one for the sports desk, one for the business desk — never letting earnings language and national-first framing sit in the same article. That separation is not a technical rule. It is craft discipline.

By 2026 I had a weekly chart newsletter with roughly 900 readers. Then a Bengaluru second-screen app firm contracted me for the 2026 Russia World Cup: build an xG and PPDA model feeding 64 match previews for a broadcaster's digital desk. I logged 2,900 shot events. My pre-tournament table showed Germany's rolling xG differential falling for 14 straight months; Germany exited in the group stage. The same desk then declared a trend off a single 3.1 xG match. That night I added a standing footnote to every page I filed — n=1 is not a trend — and refused to publish single-match xG as analysis. It cost me two click-driven commissions and earned me a reputation for being slow and correct, which I preferred.

In March 2026 the PGTI suspended its season and the Hero Indian Open was cancelled. Locked down in Bengaluru, I did the only thing I trusted: re-charted five years of Indian Open broadcast footage, 2026 to 2026, 1,180 holes, to build a pre-COVID scoring baseline. When the tour resumed without galleries, field scoring average dropped 0.6 strokes across the first three events and was back at baseline by the fifth. That week I wrote the noise-versus-signal essay and set the rule I still keep — no trend claim before three events.

Now the point that sits above all of it, and the one most ignored in this data-forward era: an empty column is also a finding, and a filled column is often less honest than an empty one. I trust the number only after I can retrace the row — who recorded it, on what instrument, and what older figure it replaced. A field marked 'insufficient information' is not a weakness in the report; it is the report's boundary line. The danger is not in that empty field. The danger is in the report where eight section headings are fully populated and not one number underneath them has a source. Mistaking format richness for analytical depth is the most expensive error in this trade. I have seen documents where the technical assessment, the form read, the rules compliance and the industry transmission blocks all appear complete — and when you ask for the provenance of the first row, the answer is blank space. Before the dashboard, there was a column someone had to defend. Today there are plenty of dashboards and no column being defended.

This is where I distrust my own work fastest. For years I used N/A as a shield. The truth is that provenance-first habits quietly become footnote creep — the live figure buried under hedges and caveats until the copy reads like an audit memo. The fix is simple and I applied it to myself: cap the provenance at one clause in the body, park the full chain in a single closing note, and always lead with the number. The second suspicion is about hiding behind data. Saying 'insufficient information' is often an analyst's lazy refuge when a limited, labelled conclusion could have been drawn from the sample available. A boundary line and an excuse look alike; the difference is that a boundary line comes with a statement of what could have been extracted and why it was not.

Which is why the rush to turn correlation into causation bothers me most. One round of 64, or one top-ten, is an observation, not a pattern. I apply a three-event threshold to every 'Indian golf has arrived' claim and state the number at its true size. Aditi Ashok finished fourth at Tokyo — fourth, every time, with the 72-hole strokeplay format and the field attached; shaving a decimal off it to write a medal story is the easiest and most dishonest work available. Jeev Milkha Singh's No. 28 in the world, or Lahiri's T5, should be read against a three-event baseline, not against a single season's average. The ledger did not need polish; it needed proof.

The next signal is not in any player's score. It is at the door of the data pipeline. Any report lacking an article title, at least one information point, one named entity — player, tournament or governing body — and a genre is not publishable, however populated its sections look. I now watch the empty-field rate across every batch I touch; if it is above zero, the fault is in the pipeline, not in the content supply. The question is not who wins the next tournament. The question is whether, next time an empty column lands in front of you, you will rush to fill it or give it permission to stay empty.