From ShotLink to the PGTI: Where Golf's Numbers Come From, and Why That Question Comes Before the Verdict

At 6:40 on a Mumbai morning a spreadsheet opened on my laptop. Seventy-two holes of arithmetic, four columns, three decimal places. The final round of the Tokyo Olympic women's golf event, August 7, 2026. The sheet said this: Aditi Ashok gained 6.2 strokes on the field with the putter across 72 holes, lost 1.4 off the tee, shot 15-under 269, finished one stroke off bronze. The sheet said exactly that. The medal table said nothing. The report filed that morning did not carry the word “medal” in its first 200 words, and carried the phrase “one stroke” six times. Last week a different sheet landed in front of me. No headline, no source, no date, and blank space where the facts should have been. The first impulse that arrived is the oldest disease of this trade: fill it. A familiar name, a familiar figure, a familiar conclusion, and the page writes itself. I did not fill it. The day we start patching blank space with inference is the day our copy stops being about numbers and starts being about memory. Three lines sat at the top of that first sheet, and those three lines are the spine of the whole business: what is the source, how big is the sample, and when is it from. I learned to count shots only after fifteen years of watching them. Before that I wrote what everyone wrote: found the fairway, hot putter, the wind changed. Nothing in those sentences was false, which is exactly why they were dangerous. None of them could be checked. Before the model, we built the protocol; before the protocol, we watched. Golf's data sits on a ladder, and no comparison makes sense until you know which rung you are standing on. At the top is the PGA Tour, where ShotLink measures every shot from tee to green — lie, distance, pin coordinate. Below it the DP World Tour publishes a narrower statistical set, and the four strokes-gained columns are not available at every event. Below that LIV Golf releases a thinner public dataset, shaped by its 54-hole format and team presentation. And at the bottom are the tours where shot-level capture has to be built by hand. The PGTI lives on that bottom rung. A fifty-lakh-rupee purse does not buy ShotLink. Strokes gained itself needs one clean definition, because in India it is still received as a novelty. It splits four ways: off the tee, approach, around the green, putting. Approach carries the largest scoring effect, because the shot that decides distance and angle into the green decides everything downstream. Putting is the most volatile — a hot week goes cold by the next Thursday, which is why treating four rounds of putting as proof of form is the easiest and most misleading habit in the game. Without a sample size, strokes gained is not a number. It is a voice. The Hero Indian Open's dates hang on the same ladder. European co-sanctioning from 2026 is what turned the event into the product we recognise; the 2026 heritage is the product's address, not its engine. The shift to DLF in the late 2010s raised the course-fit question, and the cancellations of 2026 and 2026 showed how little slack this circuit carries. In 2026 I sold the PGTI the thing it had never owned: a standardised capture protocol. Fourteen events, three-person crews logging lie, distance to pin and pin coordinates for 2,300 competitive rounds, with local scorers trained at every stop, Bombay Presidency and Poona included. At the end of the season each player's scoring average was replaced by four strokes-gained columns. That changed the copy. “Found the fairway” gave way to “gained 1.8 strokes off the tee.” Every claim now carries a sample size or an explicit admission that the sample is thin, and no scoring average gets quoted without the field's strokes-gained baseline printed beside it. A number without its denominator is a rumour with better manners. In 2026 the work travelled to football. A broadcast data vendor hired me to model all 64 matches of the Russia World Cup: live xG, PPDA, pressing-height bands, delivered within 40 seconds of each event. I wrote the methodology page myself — eleven definitions, three decimal places, and a hard ceiling of one disputed call per match. The graphics team was made to display sample sizes on screen during the group stage. All sixty-four matches became a second screen that had to earn its trust, and the person whose trust it had to earn first was mine. From football I imported a discipline into golf copy: per-round arithmetic. A hot putter is no longer in my vocabulary. What is in it is strokes gained putting per round against the field. The 40-second clock taught a second habit — number first, meaning second, context last, or nothing at all. Then 2026 put all of it through examination. The Hero Indian Open was cancelled, nine scheduled PGTI events disappeared, and my retainers froze inside a fortnight. Eleven days is not a crisis; it is a schedule with no slack. The plan built inside those eleven days was a remote data room where 40 tour players self-logged practice sessions against a fixed template, plus a strokes-gained benchmark for the Indian Golf Union's Olympic group assembled from two years of archived scorecards. No crowds, no tournaments. We standardised the practice week instead. That season produced Indian golf's first season-in-numbers piece with zero tournaments played: 9 cancelled events, 40 tracked players, 11,000 logged practice shots. It became my standing template for covering absence, and a dormant-season file now stays open in every Olympic cycle whether or not anything is happening. Underneath all of it runs an accounting the golf industry prefers to whisper. India has roughly 200 courses and about a quarter of a million registered golfers — a small, expensive enclosure inside a cricket market. The best players still emerge from caddie programmes: Mukesh Kumar came out of the caddie ranks at Delhi Golf Club, S.S.P. Chawrasia out of Royal Calcutta. The first column is the culture; the rest is arithmetic. When a green fee collides with a household's annual budget and the caddie line remains the main door for talent, the question is never talent. It is access. Calendar arithmetic belongs in the same ledger: a cancellation is not an emergency, it is a schedule that was built without slack. Here the caution has to be loudest. Correlation is not causation. The line is repeated in every data report and broken in every data report. When a player's driving numbers dip in the same week his new driver makes the news, joining the dots feels natural. It is usually wrong. A results dip during a swing overhaul is normal, and injury — lumbar, wrist, elbow — travels through the swing's kinetic chain and surfaces several events late. Translating that as loss of form is the easiest and most dishonest translation available. The second failure is the missing date. Golf information decays fast: a preview is worthless once the event starts, a result goes cold within a week. Drawing a conclusion from a report with no publication date is an act of faith, not analysis. And when the source is missing too, a third hazard grows larger than both — narrative contamination. The PGA Tour, LIV and PIF story circulates so relentlessly that a blank page fills itself with it. Prior probability is not evidence. The fourth hazard is medal drift. Fourth is not a medal; it needs repeating in a fixed sentence, because the temptation rises after every big event. Tokyo was the largest audience moment Indian golf has had, and calling it a near-medal is, to my mind, an insult to the performance. 15-under 269 and one stroke off bronze is the whole sentence. There is no room to add anything to it. The fifth is the bench illusion. Two names are not a generation. Anirban Lahiri's DP World Tour titles and Shubhankar Sharma's titles are real achievements, but count the players: India has never had two genuinely international-level players at once, holding a top-20 standard week after week. Depth is a number, not a mood. Until the PGTI has its own shot-level data, that number cannot be produced — and a number that cannot be produced cannot be argued about. The sixth is the thin-data model. Small circuits invite extrapolation because the appetite for structure is sharpest there. My own rule is plain: publish the source, the sample size and the gap before publishing the conclusion. Some PGTI purses are partial, some undisclosed, some a season old. A model built on those is not golf analysis; it is estimation with a chart. Data is a team sport, and the consultant is never the only player. The three-person crews, the trained scorers at Poona and Bombay Presidency, the graphics team, the 40 players who logged their own practice shots in 2026 — not one number reached print without them. So when someone asks where a figure came from, the honest answer is not a name. It is a chain. Three signals will matter to me in the next cycle. First, whether any Indian event publishes its strokes-gained sheet in public, even if it is built by hand. Second, whether the Olympic group's benchmark survives the next four years or restarts from zero each cycle. Third, whether a number appears in copy with its source and date attached, or whether we fall back on the old habit. Numbers do not lie, but they tell incomplete truths, and an incomplete truth costs more than a clean lie. The next time a three-decimal figure appears beside a golfer's name, ask three questions: which source produced it, how many shots it rests on, and when it is from. The day those three answers arrive, the argument will move to the result — where it belonged all along.