The Empty Screen in Boston: Reading Golf When Data Has Nothing Left to Hide

January in Boston. On the fourth floor of a small edit suite, snow smeared across the window, I opened a file my research assistant had sent overnight.

The filename: SG_master.xlsx. Four column headers — SG: Off the Tee, SG: Approach, SG: Around the Green, SG: Putting. The formatting was suspiciously clean: bold headers, hairline borders, right-aligned numbers, conditional formatting shading negative values red.

The body was empty.

No player names. No rounds. No tournaments. Only a scaffold, carefully built for something that had never existed.

I sat looking at it for about three minutes. What unsettled me was not the emptiness. It was the first three seconds. In those three seconds my eye read the frame before it read the content. It looked like a data table. It had the shape of a conclusion.

An empty screen forces me to read a match the way I read an unedited manuscript.

In nearly twenty years in this trade, I have learned that most errors in sports journalism do not come from wrong data. They come from correct frames filled with nothing. A chart with axes, labels and colour will be believed. A blank space will not. And my profession, at this moment, is producing a great many charts.

Twenty years of golf learning to count

Golf entered the data age roughly a decade behind football and basketball, for one very specific reason: nobody passes the ball. No passes means no network, and no network means no positional data. Every shot is an isolated event across sixty hectares, and for nearly a century the only way to measure it was to count final outcomes: total strokes, fairways hit, greens in regulation, putts per round.

Those three metrics — GIR, driving accuracy, putts per round — share a common defect. They measure a shot by the shot that follows it, not by the shot itself. A 300-yard tee shot into rough loses nothing if the next swing finds the green. A perfectly struck approach that finishes forty centimetres off line still counts as a missed green. Traditional statistics cannot separate the skilful from the fortunate.

In the early 2000s the PGA Tour deployed ShotLink — a per-hole data collection system recording coordinates for nearly every shot by nearly every player in nearly every round. In parallel, Mark Broadie, a Columbia professor, developed Strokes Gained: instead of counting outcomes, it measures the expected value of a ball's position. A good shot moves the ball from a spot with an average expectation of X strokes to a spot with an expectation of X minus something. That difference is strokes gained.

It was a genuine leap. For the first time golf had a unit that could be added and subtracted across skills, across holes, across tournaments. Four standard categories emerged, and the PGA Tour began publishing them publicly in the 2010s. Within a decade they had replaced almost the entire vocabulary of player evaluation in press rooms.

This season places those four columns inside an ecosystem stretched in four directions at once. Sport: four majors running from Augusta in April through the PGA Championship, the U.S. Open and The Open, closing with the FedExCup playoffs. Money: the PGA Tour–LIV Golf confrontation, backed by Saudi Arabia's Public Investment Fund, which has turned prize structure into an argument about the nature of sporting labour. Institutions: the Official World Golf Ranking and the unanswered question of who is awarded points. Equipment: the ball rollback finalised by the USGA and The R&A in late 2026, applying to elite competition from January 2028 and recreational play from January 2030.

All four axes run on one data layer. That is where the trouble starts.

Four columns, and what they leave out

Strokes Gained is not a physical measurement. It is a comparison against a baseline. Every time you read that a player gained 2.1 strokes on approach for the week, you are reading the result of a subtraction between him and the field average, adjusted for lie and terrain.

Three consequences follow, and almost none of them appear in the summaries.

First, putting is the most volatile of the four categories, and therefore the least safe to extrapolate. A PGA Tour player takes roughly 28 to 30 putts per round. That sounds like a lot, but most are inside five feet — effectively automatic. The number of putts that genuinely distinguish skill is closer to six to eight per round. With six to eight observations, variance swallows the signal in most cases.

Second, approach is both the category most correlated with scoring and the most stable week to week. If you were allowed one column to forecast a season, take approach. Public data has said so for over a decade, and it is the thing tournament coverage forgets whenever it builds a story around a putt on the 18th.

Third, and most ignored: Strokes Gained is calculated against the baseline of that week's course setup. If pins are soft, rough is short and wind is calm, the best player in the field posts modest numbers, because everyone scores well. If the course is a brutal U.S. Open test, the same performance produces an enormous figure. Same player, same shot, two tournaments, opposite readings. The number is not wrong. It needs context, and context does not live in the spreadsheet.

The twenty-four putt problem

I have watched this scene play out in press rooms many times: a player finishes a round on 24 putts and is instantly crowned the best putter of the week.

Unpack it. On 24 putts, he saved roughly four to five strokes against average. Most of that came from hitting approach shots close — credit belongs to the preceding swings. Another portion came from long putts that happened to drop, which the player himself, asked privately, usually admits was luck.

The problem is not that media reports 24 putts. The problem is that the figure is used to define a player when the sample is one round. When the stands are empty, the match exposes what tactics conceal. Here what gets exposed is a basic statistical rule: extremes in small samples regress toward the mean, not because a player got worse, but because the extremity was never durable.

The reverse holds too. A player with a bad putting week gets labelled out of form when he may be in perfectly stable condition and simply encountered seven outlier situations. I have spent weeks matching round-by-round putting figures for the same player across a season. The spread between his best and worst rounds is usually wider than the gap between the 30th-ranked and 150th-ranked player on season averages.

In other words: the confidence interval of a single round is wider than the distance between two tiers of ability. Any conclusion drawn from one round — about form, about future, about class — is describing noise, not signal.

Heat maps: this century's astrology

The most popular visualisation in televised golf in recent seasons is the heat map. A green rendered in colour gradients showing where balls landed, where they stopped, where players scored. A dispersion map of tee shots. A zone chart of fairway positions.

These images are beautiful. They carry one property that worries me every time: they create a feeling of understanding without transmitting any information about mechanism.

A green heat map tells you player X's balls cluster on the left half. It does not tell you why. Did he aim there deliberately because the pin was there? Is he fighting a swing-path issue that sends the ball left automatically? Is it the by-product of a swing change mid-transition? Three entirely different causes, three entirely different implications, and the heat map cannot separate them.

Based on my experience tracking rounds across many tournaments, I have watched analysts reach three contradictory conclusions from the same heat map, in the same week, on the same broadcast. This does not happen because anyone is incompetent. It happens because the image does not carry enough information to rule out hypotheses.

That is why I call the heat map this century's astrology. Not because it is false, but because it is true to the exact degree that any interpretation becomes plausible. A tool that can justify every hypothesis tests none.

What makes this worse is market structure. Most golf analysis is produced to fill gaps between rounds, not to answer a specific question. When airtime is fixed and data is formally unlimited but substantively finite, pressure tilts toward producing attractive images rather than firm conclusions.

Course fit: the story told backwards

One of the most common genres before a major is course fit — which players suit which course.

In principle it is a sound problem. Narrow courses with thick rough reward accuracy. Long courses reward distance. Fast, sloped greens reward green-reading. Coastal links reward low ball flight into wind. There is real technical logic underneath.

In practice, most course-fit pieces I read are written in reverse. First identify the likeliest winner, based on recent form and ranking. Then find a course attribute that matches his technical profile. Finally, present the causal chain in the opposite order.

This method leaves a very recognisable trail. It explains the chosen names well and ignores others who also fit but lack recent results. It always finds a reason for the winner and always finds a reason for the loser.

I call this the prophet's paradox: a system that explains everything predicts nothing. And it has a real cost. Every course-fit story built backwards is a missed opportunity to ask a genuinely predictive question — for instance, whether a player with elite approach numbers but a weak short game can survive thick rough and small greens. That question is answerable. It is simply less entertaining than the story.

The ball rolls; the money moves behind it

The ball rolls across the turf, but I am reading the money moving behind it.

No part of modern golf depends on money for its data quite like this one.

The PGA Tour operates a merit-based prize architecture: money by finishing position, FedExCup points by finishing position, membership by finishing position. LIV Golf operates a structurally different one: guaranteed contracts, team formats, 54 holes, shotgun starts, no cut. These are not two versions of the same product. They are two theories of how athletic labour should be paid — one rewarding outcome, the other rewarding participation.

The resulting wave of movement produced a chain of events the industry usually tells as a money story. Jon Rahm signed with LIV in December 2026, already a major champion and at the peak of his career. Cameron Smith moved in 2026, immediately after winning The Open. Brooks Koepka moved to LIV and then won the 2026 PGA Championship. Phil Mickelson won the 2026 PGA Championship at fifty before making the switch.

What matters about those facts is not the facts. It is that they break an assumption on which the entire ranking system was built: that the only path to proving class is accumulating points on OWGR-sanctioned events.

LIV withdrew its OWGR application in 2026. Since then, its players must rely on major results to hold ranking positions — and to earn major invitations, they need ranking positions. The loop closes to their disadvantage, and it produces a data effect few discuss: a significant share of this sport's elite is competing outside the public data layer.

For analysts, that means the database we use to determine the best player in the world has a structural hole. Any ranking built from ShotLink data automatically excludes a group of genuinely strong players. Any conclusion about generational dominance depends on where you draw the boundary of a generation.

The negotiations between the PGA Tour, the DP World Tour and PIF — opened by the June 2026 framework agreement and still unresolved — are not merely governance. They are about who controls the sport's foundational data layer for the coming decade.

The transfer market is a mirror reflecting the fears of whoever signs the contract. In golf, the specific fear is being erased from the measurement system.

The twelve-year-old lottery ticket

At the deepest layer of every sports data pipeline is a person who was never entered into the spreadsheet.

I have spent much of my career watching recruitment systems. Golf's cost structure is harsher than most. A young tennis player can train against a wall with one coach. A young golfer needs a course, a membership or green fees, a set of clubs refitted as the body grows, a swing coach, a short-game coach, a strength coach, a psychologist, and travel and lodging for junior events.

In the United States that system runs through national junior tours, where a middle-class family can spend the equivalent of a university tuition every year, for years, on a twelve-year-old with unproven potential. In Korea and Japan the pattern is often relocation: the family splits, the mother moves with the child to the US or to a training base, the father stays behind to earn.

The scouting network does not merely find talent. It manufactures lottery tickets, and sometimes, broken families.

I want to state this with data rather than sentiment, because sentiment in this area is systematically undervalued. Look at the structure of the career path. A young player must pass through junior and amateur ranks, then Q-School, then the Korn Ferry Tour — the PGA Tour's feeder circuit — before touching a full card. At each level the pass rate narrows, and the operating cost does not narrow at the same rate.

On the Korn Ferry Tour, most players do not earn enough in prize money to cover a season's costs. Which means the family investment does not end at eighteen. It extends to twenty-two, twenty-five, twenty-eight. And for every player who breaks through, dozens paid the same invoice.

The success story of an Asian golfer in America is usually told as a story of cultural discipline. That framing both flatters and erases. It flatters because it converts an economic consequence into a spiritual quality. It erases because it does not mention what the family sold, which sibling went unfunded, or how many years the household lived in a state of waiting for an outcome nobody guaranteed.

On the other side of the Pacific there is a symmetrical blind spot: Americans read the discipline of Asian academies as an explanation, when much of it is a function of having no safety net. A child in a system with a safety net can quit golf at fifteen and go to university. A child in a system without one often cannot, so she continues.

That continuation looks like discipline from outside. From inside, it is a sunk-cost calculation.

Ball rollback: the argument isn't about distance

In late 2026 the USGA and The R&A finalised changes to the Overall Distance Standard, limiting ball flight after a comment period that ran through much of that year. The rollout was tiered: elite competition from January 2028, recreational play from January 2030.

Public debate has centred on one question — whether distance should be limited at all. A second axis gets less attention: how the rule operates across two groups with entirely different swing-speed profiles.

A professional swings the driver above 180 km/h. An average male amateur in middle age sits around 130 to 145 km/h. The impact of a limit designed around high-speed test conditions is not distributed evenly across those groups. Some industry equipment research suggests slower swingers lose less distance, but other data contradicts it, and most of the detailed evidence sits with manufacturers rather than the public.

Which returns us to this article's central problem. A regulation shaping the sport for decades is being assessed on a dataset whose most important part is not fully published. Fans are asked to trust the conclusion without being given the tools to check it.

There is another dimension technical analysts usually skip. Equipment rules hit the supply chain. Club and ball makers must reallocate R&D budgets, redesign product lines, and build inventory for a multi-year transition. And at the end of that chain sits a golf course deciding whether to buy land to lengthen a hole — a decision whose value is tied directly to how far the ball travels.

The distance debate, followed to its end, is a debate about land.

The media heat cycle

Every golf season runs on a predictable narrative cycle. Before a major, media looks for an heir. After a major, media looks for a decline. Between those points sit roughly eight weeks of stories built on thin data.

In a generational transition, the cycle accelerates. Each time a young player wins a regular event, a run of pieces announces a new reign. Each time an older player wins a major, a run of pieces announces that age doesn't matter. Both are written without anyone checking whether they cancel each other out.

A season is only a sentence in a book that runs decades long.

A long-run data observation worth recording: peak competitive ages in men's professional golf, measured by major win frequency, cluster between 28 and 38, with a long tail past 40 for a small group. That tail exists because golf is a sport where course management and emotional regulation partly offset physical decline — something no pure speed sport can claim.

But the data does not say age doesn't matter. It says age is a variable with a wide distribution. Very different.

The contrarian angle

Here I want to invert an assumption most of sports analytics runs on: that our problem is a shortage of data.

Golf has more data than anyone can process. ShotLink has logged every shot on every hole in every round for over two decades. Third-party platforms layer simulation on top. Equipment makers collect swing-level data. Nobody is short of data.

The problem lies elsewhere, and it fits in one sentence: data has shifted from a tool for opening arguments to a tool for closing them.

When a number is cited in a press room, it stops being a hypothesis to test. It becomes a verdict. People once argued with observation and personal credibility — both easy to challenge. Now they argue with charts, and charts win because they look objective, not because they are right.

This is especially dangerous in golf for a structural reason: golf has a naturally low signal-to-noise ratio. Only about seventy players per week survive the cut, and of those only about ten have a realistic chance to win. From ten, you can build a story about anyone. With four strokes-gained categories split by round, you can build a story about any player at any position on the leaderboard.

Confidence in golf analysis today is not proportional to the certainty of the conclusion. It is proportional to the number of charts that can be built.

Which returns us to the empty spreadsheet in Boston. What I learned from it was not that empty data is bad data. What I learned was that the data frame is built before the data exists, and it shapes the conclusion before any observation occurs. Build the frame first and you have already chosen the kind of story you will tell.

Coldness in analysis, in the end, is a long-term strategy: say less than you know, and know exactly how much you have.

What remains

It would be easy to close with a call for more caution. I don't want to. Golf does not become more interesting when people write less about it. It becomes more interesting when people write about it more accurately, which is considerably harder.

If there is one thing I want to carry out of this season, it is a small change in reading habit. When you see a strokes-gained chart this week, read it as a question, not a conclusion. When you see a player called the best putter of the week, ask how many putts the sample contains.

And when you see a beautiful spreadsheet, check whether there are any numbers inside it.

A season always ends. Fans move to the next. Studios close their projects. And in the quiet weeks between seasons, when the courses are empty and nobody is logging data, that is when I reread old statistical tables and discover that much of what I believed for a year was noise formatted attractively.

That does not cost me faith in the sport. It costs me faith in how I used to read it — and that is a necessary loss.