The Eight Dimensions of Golf Analysis: Data Methodology, Gaps, and the Limits of Models

There was an evening when I opened my report file and found it empty. Not a formatting error, not a font error. Eight analytical frameworks sat there, fully titled, fully table-lined, fully provisioned with empty cells — and inside every cell was a dash. I sat still for about ten minutes. For someone who reads numbers for a living, a file like that is not a failure of data. It is a reminder that data never arrives on its own; someone has to be responsible for bringing it home, checking it, and saying plainly when there is nothing there.

I write this for golf readers in Japan and Vietnam — people used to opening a leaderboard on a Sunday morning and finding a number. But I want to tell a different story: the story of the frame behind the number, how it is built, where it collapses, and why a good analyst is not the person with the most numbers but the person who knows when to say "I have nothing yet."

Context: a sport both rich and poor in numbers

Golf is a strange sport on the data map. On the surface, it is an analyst's paradise. Every shot is a discrete event with coordinates, distance, and outcome. No teammates blocking the view, no defence adding noise, no referee deciding most of the result. If you want to measure an individual under pressure, golf gives you near-ideal conditions.

But at a deeper layer, golf is data-poor in a way football or basketball is not. A PGA Tour season has dozens of events, each four rounds, each eighteen holes. That sounds like a lot. Yet when you slice it to answer a specific question — say, how good is this player with irons into the pin when the wind is against at fifteen miles per hour or more — the sample shrinks frighteningly fast. That is the foundational paradox of golf analytics: a vast total of data, but very thin usable data for any single question.

The arrival of Strokes Gained changed the game. Mark Broadie, a professor at Columbia Business School, systematised the metric and popularised it widely through his 2026 book. The core idea is methodologically elegant: instead of counting strokes, you measure the value of each shot by the difference between the expected strokes remaining before and after that shot, calculated against a common baseline across the entire field. A three-metre putt on the seventeenth and a seven-metre putt on the second suddenly speak the same language.

But I learned one thing doing this work: every new metric carries a new trap. Strokes Gained is not a verdict. It is a way of slicing the truth, and every slicing method has its blind spot. Strokes Gained measures the value of a shot, but it does not measure the cost of having to make that shot inside a human head. And as always, a poor analyst takes the pretty number and forgets what was just said.

I came to golf after a decade in football data. I built models, I was wrong, I sat through footage frame by frame to understand where I went wrong. That experience taught me that every field shares one disease: people want a conclusion before they have enough evidence. Golf is no exception. It is just that in golf the disease is harder to spot, because a leaderboard looks so convincing.

Core analysis: eight dimensions and a load-bearing frame

Whenever I have to build an analytical frame for golf, I always start with eight dimensions. Not because eight is a magic number. Because eight is the minimum for a sports story not to collapse under interrogation. Each dimension answers a question, and each question can be answered with "not enough data." That is the most important feature of this frame.

Dimension one: technique and shot data

This is the layer readers assume is settled. Strokes Gained: Off the Tee, Strokes Gained: Approach, Strokes Gained: Putting, Strokes Gained: Around the Green. Distance, accuracy, fairways hit, greens in regulation. These four main metric groups, plus course fit, form a technical profile.

But I want you to notice the biggest gap at this layer: advanced metrics are only trustworthy when the evaluation data is dense enough; with a thin sample, a pretty number is just a lucky wind packaged as a chart. A player hot with the putter for three weeks can lead Strokes Gained: Putting, but three weeks is not a career. The analyst has a duty to state the sample size before offering praise.

Tracking numbers across courses in Japan and on the major systems, one thing stands out: course fit is chronically underrated. A player with a high-ball iron profile thrives on soft turf but looks lost on a fast, firm course where the ball must run and bounce differently. Ignore the course variable and every comparison is hollow.

Dimension two: player and form

Here we have amateur and professional rankings, recent events, major records, and age. Each piece is familiar. The problem is in fitting them together.

The age curve in golf differs greatly from basketball. Golf tends to peak late, around the early thirties, and lasts long if the body holds. That makes reading form far subtler. A twenty-two-year-old leading a regular event does not mean he owns the future; he has merely hit a good point on the arc of a human life.

I always recall a principle from youth development: young players who mature early are over-used, and golf is not immune to that. A group of muscles not yet fully grown is pushed into a dense schedule, constant travel, high-intensity practice — and the bill usually arrives years later, not immediately. The leaderboard does not know this. Only injury history does.

Form, then, is not a number. It is a time series with memory. The same score, produced by a body in recovery and by a body in full bloom, carries entirely different value. I do not believe in luck; I believe in cultivated probability.

Dimension three: the tournament system

Golf runs on tiers of events. Majors at the top, a regular annual circuit, special events, regional tours. Each tier has its own points scale, purse, and eligibility conditions. Ignore this structure and every analysis loses its footing.

A simple example: a win at a strong event is not the same class as a win at an event with a thin field, even if both are titles. The world ranking system, launched in 2026, tries to quantify that difference by weighing field quality and the time decay of results. A good finish in a strong field carries more weight, and old results fade with time.

Many fans read the rankings like a dictionary and forget it is a weighted graph. The same position in mid-May and mid-December are not the same. The system structure decides who plays, where, and for how many points. That is the invisible part with the heaviest weight.

Dimension four: context and governance

This is the dimension fans of the leaderboard often skip, and the one that has most shaken modern golf in the first half of the 2020s.

Professional golf once had a relatively clear order: the American circuit at the centre, the European circuit behind it, regional tours below. Then a new league with enormous financial backing from Saudi Arabia's Public Investment Fund punctured that order. From 2026, big money began moving players, and in June 2026 the parties announced a framework agreement to seek commercial consolidation. By August 2026, the negotiation had not closed.

For a data analyst like me, the central question is not who is right. The central question is: when the tournament structure changes, the value of every individual metric changes with it. A player moving to a circuit with fewer events and a thinner field collects numbers that, on the old baseline, look beautiful. But that beauty is a consequence of context, not of skill.

Elimination is the key to the transfer market. When you see a player surge in metrics after switching tours, the first move is not to praise but to ask: on what baseline were those numbers produced, in what field, under what course conditions. If you cannot answer, the number is only the echo of an empty corridor.

Dimension five: rules and equipment

Golf is a sport where playing rules and equipment regulation can change the fate of a whole generation. This is rarely mentioned, but it is a hard variable no long-term model can ignore.

The two main governing bodies for the rules are the royal golf association of Britain and the golf association of the United States, which jointly issue regulations with global reach. Over the past two decades they have tightened matters relating to club grooves, to the anchoring of the putter, and to a research project on hitting distance. Each tightening rewrites part of a generation's data.

For the analyst, this is a lesson in false invariance. When you compare driving distance across two decades, you are comparing two different legal environments. Without noting that, every growth chart is a technical lie. I always add a note on the legal environment before drawing any long-run trend. Gaps in a data table also speak, if we are willing to listen.

Dimension six: the risk surface

Risk in golf is not only injury. It includes competitive risk, psychological risk, career and commercial risk, governance risk, and systemic risk.

A young player who suddenly rises usually carries three risks at once: surging expectations, a denser schedule, and early commercial contracts. Each can damage the development curve. Outwardly he looks healthy and ascending. Structurally, he is walking on a thin shell.

This is why I always separate physical risk from performance risk. A player can win three events in a row and still sit in a high-risk zone, because real-time conditioning data — the thing I once lacked — shows the load has long exceeded the safe threshold. Error here is not a calculation mistake. It is a signal.

Dimension seven: public narrative

Every player, every event, has a story the media tells. A comeback, a dominance, a new generation, the gift of youth. Those stories exist because they have pull. But for an analyst they must be tested.

I often ask: how long would this story survive if all commentary were removed and only the data remained? If the answer is "a few weeks," it is a hot streak, not a trend. If the answer is "a few years," it belongs in a model.

What did NOT happen often tells the truth better than what did. A player praised for winning two majors in a year — the more telling fact is that he missed the cut five other times. The silence of unmentioned rounds is the fullest data on consistency. The media rarely counts the things that did not happen. The analyst must count on their behalf.

Dimension eight: industry transmission

Finally, golf does not exist in a vacuum. It is a chain from courses, equipment, and development, to events, broadcasting, sponsorship, and data.

When big capital flows into a new league, the effect does not stop at the player list. It spreads to sponsorship values, broadcast strategy, the betting market, and youth programmes in distant countries. In Vietnam and Japan, where I live and write, a small change at the centre can create a wave at the edge.

I always try to trace how far an on-course event travels. It goes from the leaderboard, through the news, through the contract, and finally to a twelve-year-old practising irons on a range on the outskirts. If I cannot trace it that far, my analysis is not finished.

Contrarian angle: when an empty data file is the best signal

Now I return to where I started. The empty file.

Across all eight dimensions just covered, the scariest thing is not a wrong number. The scariest thing is a frame that looks complete but holds nothing inside. It has titles, cells, formatting. A glance suggests the work is done. And that very completeness stops people from checking again.

I once lived through a period when matches were suspended entirely by a pandemic. Empty stadiums, a shattered calendar, broken match data. People assumed that with no new data, analysis must stop. Reality was the opposite. In that void, rebuilding the model mattered most. I had to find substitute data, look at precedents from historically disrupted seasons, use every indirect signal available. The result was not bad, because I did not pretend to have data I did not have.

That is the core lesson. Honesty about data gaps has higher analytical value than a number filled in by guesswork. When a model admits it lacks a variable, it becomes more trustworthy within the range it can still speak. When it claims to be complete while being hollow inside, it becomes a time bomb.

Correlation is not causation is a line everyone knows by heart and almost no one practises. In golf the temptation is overwhelming. A player changes clubs and wins the next week. A player changes coach and putts better. Those stories are attractive, factually true, and evidentially meaningless. To turn correlation into causation you need controlled variables, an adequate sample, and elimination. Without those three, every explanation is literature, not analysis.

I have asked the wrong question before. I once built a prediction model on raw data, ignored context, and got nearly six of ten late-round predictions wrong in a stretch. Looking back, the data did not betray me. I had omitted variables that were not in the tables. Data is never wrong; I simply asked the wrong question.

And another time, I ignored a real-time conditioning variable in a big match. Every pressing metric early on supported one conclusion. But the opponent's running intensity in the final twenty minutes overturned it all. Since then I have set a rule for myself: never conclude on pressing ability without conditioning data broken into time intervals. Every number is a confession not yet written into prose. My confession that time was impatience.

What fans should carry with them

I am not writing this to teach anyone how to read a leaderboard. Leaderboards are easy to read. The hard part is knowing what a leaderboard is not saying.

If you follow golf, I suggest a small habit. Every time you see a pretty number, ask three questions. First, over how many samples was this measured? Second, in what course and field context was it produced? Third, if you remove the ten best results, what remains? Those three questions will dismantle most of the illusions leaderboards sell us every week.

For young people who want to do analytical work like mine, a different note. The most important skill is not writing formulas. It is the ability to say "I don't know." A mature analyst is not the one who always has an answer. He is the one who knows precisely which answers he lacks the data to give. The empty file I opened that evening was not a failure. It was a reminder that methodology is the only asset that cannot be faked.

In the regular season, as each weekend brings a new champion and a new story, the pressure to conclude rises. I understand it. I want answers too. But I have learned that golf is a sport designed to punish haste. Four rounds, seventy-two holes, and a turf that forgives no one who misremembers the wind.

Perhaps the right way to read golf is as a multi-year sequence of experiments, where each event is only one re-run. Each re-run gives a signal, not a truth. Truth appears only after enough reverse verification, enough self-contradiction, enough endurance of the feeling of being wrong.

I will keep watching this season through that same eight-dimension frame. There will be moments when a dimension is empty, and I will say plainly that it is empty. There will be pretty numbers, and I will have to resist the urge to write about them at once. But that is the job. When data hides its face, error becomes the guide.

The question I want to leave you with — not to answer now but to carry through the season: if tomorrow every ranking disappeared, what signal would you use to know where a player truly stands on the arc of his career? I do not have a full answer. But knowing I lack one already feels like progress, compared with believing I had it long ago.