The Empty Cells on Golf's Data Sheet

Last August a four-page spreadsheet landed in my inbox at two in the morning, Boston time. The sender was an editor I have known for eight years. The sheet had every standard column: SG: Off the Tee, SG: Approach, SG: Putting, average driving distance, greens-in-regulation rate, scrambling rate, OWGR position, top-10 major finishes across three seasons. Six cells were empty. No footnote, no question mark, no unverified tag. The email ended with one line: fill in the rest for me, deadline is nine.

I sat with those six empty cells longer than the task required. After twenty-one years watching this industry, I believe the most dangerous moment in sports writing comes when someone is obliged to fill a blank with whatever sounds most plausible. A blank screen forces me to read a tournament the way I read an unedited manuscript. A blank is not neutral. It has gravity, and that gravity always pulls toward the easiest story to tell.

Golf is the most densely quantified individual sport in the world. Mark Broadie, a Columbia University professor, built the Strokes Gained framework and published it in full in Every Shot Counts in 2026. The PGA Tour operates ShotLink, which logs every shot to the foot. Independent platforms such as Data Golf build forecasting models from that same feed. Major Championship entry is allocated through the Official World Golf Ranking. A touring professional today walks into a tournament week carrying more numbers than a footballer accumulates in an entire career.

That density produces a dangerous illusion: that any data sheet must be full. When a sheet has forty columns, people assume column forty-one only needs filling in. Three of golf's largest disruptions in four years point the other way — LIV Golf launching in June 2026 with backing from Saudi Arabia's Public Investment Fund (PIF), OWGR refusing to award points to the 54-hole format, and the USGA and R&A ball-distance limit taking effect in January 2028 for elite competitions and January 2030 for recreational play. More numbers mean more places to hide a gap.

Based on my experience covering tournaments and press conferences, I read golf through eight layers. Each layer has its own kind of blank, and each kind of blank has its own way of being filled wrongly.

Start with technical data. Strokes Gained divides a round into four areas: off the tee, approach, around the green and putting. Broadie's research shows approach play correlates most tightly with scoring at tour level, while putting is the most volatile of the four. This is where most analysis goes wrong. A hot putting week rarely repeats, and extrapolating it into form is the elementary error of the discipline. It is also the easiest cell to fill, because it produces a tidy story: the putter is hot. A season is one sentence in a book a decade thick, and a tournament week is a comma inside that sentence.

Behind the metrics sits form and career trajectory. A decent observation window for form is five to ten events, not two. The age curve for a professional is not flat: peak driving distance usually arrives early and declines before scoring skill does. Injury travels along the kinetic chain, from ankle through hip to shoulder, and an ankle problem can completely change the attack angle of an iron. No model builds that from a scorecard. It requires someone standing at the practice area before the gates open.

Then comes tournament structure — the part treated as dull and therefore filled carelessly. The cut at most PGA Tour events sits at top 65 and ties; survive it and you earn prize money and OWGR points, miss it and you earn neither. A Tour Card decides the following season. The Korn Ferry Tour is the main feeder stage, and Q-School is the narrow route back. The FedExCup finale applies starting strokes based on season standing. All of this is verifiable, citable, checkable. Precisely because of that, it is treated as filler, when it is what determines who is still playing in September.

Governance is the loudest layer. The tension between the PGA Tour, LIV Golf and PIF is not an algorithm problem. OWGR withholding points from LIV is an institutional decision, and it reshaped the major pathway for an entire generation. The same applies to prize structures and media rights. In the summer of 2026, aged twenty-nine, I sat in a World Cup press conference in Russia and was cut off by an older male reporter while asking about a tactical shape. He said women should ask about a player's family instead. I did not argue. I spent three weeks re-analysing the data, and the resulting piece was republished by forty-seven outlets. The lesson was not to be braver. It was this: they doubt the voice before they hear the argument. I learned to gather the evidence first and the expectation second. Every piece of mine has carried a verification block ever since.

Rules and equipment is the quiet layer. The ball-distance limit published by the USGA and R&A lands differently on professionals and recreational players, with two separate timelines. In golf a single penalty stroke can decide a tournament. Slow-play enforcement exists on paper and is applied inconsistently. And the yips — involuntary movement on short putts — are almost purely psychological, measured by no index at all. This is where the data stops and people begin.

The risk surface runs across competitive, psychological, injury, commercial, governance and system risk. In my trade, the largest risk sits outside all six. It sits in a data sheet that looks complete being passed to a reader while nobody checks how many cells are genuinely empty.

The public narrative layer runs on templates. Golf has a handful of near-automatic story shapes: the new king crowned, the dynasty consolidated, the redemption arc, the defector's price, the Career Grand Slam chase. Only five players in history have completed the career Grand Slam: Gene Sarazen, Ben Hogan, Gary Player, Jack Nicklaus and Tiger Woods. Those five names are real and verifiable. But when a young player wins twice in six weeks, the new-king template arrives before any data sample is large enough to say anything.

The longest chain is industry transmission. A win travels through the course, the equipment brands, sponsorship and broadcast rights, data and betting, and finally down into junior scouting. That last link interests me most and carries the least data. Scouting networks in developing countries find genuine talent and also produce lottery tickets and broken families. No ShotLink records that part. The ball rolls on the grass, but I am reading the money moving behind it.

The irony is that public debate about golf analytics almost always aims at the wrong target. People argue over which metric is right — Strokes Gained or scoring average, a Data Golf model or official stats. But the credibility of a metric is not in its formula. It is in the blanks left around it. A model that does not state what it ignores is not a model; it is a claim decorated with numbers.

I would argue more data has made golf coverage less honest, not more. With three metrics, a writer is forced to admit they do not know much. With two hundred, a feeling of completeness appears, and that feeling is the best soil for unsupported assertions. Coldness is a long-term strategy, not a personality defect — and in this trade, coldness means refusing to write a sentence I cannot verify, even when that sentence would travel further.

What I want from the next generation of golf analysis is not a more accurate model. It is a small change in publishing format: two blocks at the foot of every piece. A verification block listing the source of each figure. And an unknown block stating plainly what the writer could not verify and will not guess. One article willing to print that we have no injury data on this player is more honest than ten articles speculating about form.

Those six empty cells, I left them empty. I sent the sheet back with a short note explaining why no source could confirm those six metrics, and suggested either cutting them or publishing with a data-limitation note. The editor chose the second option. The piece carried fewer numbers than planned, and it was more accurate.

If a golf data sheet can be empty in six places without anyone noticing, the problem was never a shortage of data. The problem is an industry so practised at filling gaps that it has forgotten a gap is also information.