Strokes Gained, OWGR and the Craft of Catching Out Flawed Golf Data

I remember a summer evening in Hai Phong, sitting in front of a screen with the Strokes Gained table of a golfer who had just won a weekend event. The Approach column blazed green, the Putting column glowed red. According to that table, he won because of superb approach play and below-average putting. But when I rewatched all 18 holes, I saw almost the opposite: he putted with a firm hand through the front nine and won because of three decisive putts on holes 16, 17 and 18. The table was not mathematically wrong. It simply told a different story from the one I heard with my heartbeat. Every set of numbers can lie; my job is to catch it in the act.

I did not come to golf from a golfing family. I came from the track, from a fall at the 350-meter mark at the Hoi Khoe Phu Dong city games in 2026, and from the months that followed, when a hamstring injury forced me off sprinting and toward a different sport to observe. Golf chose me in a very counterintuitive way: it speaks the same language as track and field — the language of distance, rhythm, and managing a body under changing conditions. And it is also the sport where data, as I am about to tell it, can be both beautiful and empty at once.

What I want to do here is an investigation into how golf data is produced, how it gets distorted, and why a technically perfect analytical table can fail to contain a single usable line of information. In other words: I want to catch the data-production engine itself in the act.

Where the golf data era began

Golf has a structural advantage for becoming a data paradise. Every shot is a discrete event with a starting point, an endpoint, and a measurable result. Eighteen holes, a par of 72, each stroke a sample. There is no running clock, and for most of the round there is no opponent physically blocking you — only the golfer, the ball, and the ground.

The real turning point came from the work of Mark Broadie, a professor at Columbia Business School. He developed the Strokes Gained concept around 2026, built on a deceptively simple question: how many strokes does a shot from a specific distance and lie take, on average, to finish the hole? When you compare that average against actual outcomes, you get the net contribution of each shot. Broadie's book Every Shot Counts, published in 2026, turned that idea into the foundation of virtually all modern professional golf analytics.

The technical infrastructure came with it: ShotLink, the PGA Tour's shot-tracking system, running since the early 2000s. ShotLink records the ball's position after each stroke, distance to the hole, and lie type, producing the raw data from which Strokes Gained is calculated. By the mid-2010s, Strokes Gained had become the standard language of analysts and television commentators alike.

But here is where I want to stop. Strokes Gained is not an objective measure of the quality of a shot; it is a comparative measure against the average of a specific population under a specific set of conditions. If the baseline database changes — the course, the weather, the comparison group — the meaning of the same number changes too.

The four main categories are SG: Off the Tee (driving), SG: Approach the Green (approach shots, widely rated by analysts as the metric most correlated with scoring), SG: Around the Green, and SG: Putting. When someone says a golfer won thanks to "approach", they are usually citing the Strokes Gained table. But my question is always: at what stage of the event did this player win, under what conditions, and is the sample big enough to support a conclusion?

How Strokes Gained works in practice

To understand Strokes Gained, picture a giant baseline table. For example, if the average PGA Tour player standing 150 yards from the hole on the fairway needs 2.9 strokes to finish the hole, and your shot leaves the ball 8 feet from the hole, where the average is 1.2 strokes, then your shot has created an advantage of roughly 1.5 strokes over the average. Sum that across all 72 holes and you have a single number: if positive, you played better than the tournament average; if negative, worse.

The advantage of the method is obvious. It separates the shot from emotion and from reputation. It allows a shot on a hard hole to be compared with a shot on an easy hole using the same yardstick. It turns golf from a crude statistics game — greens in regulation, fairways found, putts per round — into a sport analyzable stroke by stroke.

The drawbacks are just as clear. First, Strokes Gained depends on the database. If ShotLink misreads a ball position, or if a tournament lacks full tracking, the data loses value. Second, small samples. Eighteen holes is a weak sample; four rounds is better but still easily distorted by a few lucky putts or one disastrous hole. Third, and this is the point I keep emphasizing, Strokes Gained measures outcomes, not intentions. A technically sound shot that lands half a yard off the green can score negative, while a careless shot that stops next to the hole scores positive.

I once wrote a long piece praising how a golfer survived by scrambling around the green in strong wind at a coastal event, only for a colleague to point out that his Strokes Gained table was heavily negative — because his scrambles left the ball in a worse position than the average for that lie, even though they looked spectacular. The lesson: the human eye is easily fooled by beauty, and data knows no sympathy. Both need to be cross-checked.

From Strokes Gained to OWGR: data does not live in a vacuum

Strokes Gained is competitive data. But golf has a second layer of data that governs a golfer's career: the ranking system. And that layer is both more complex and far more political.

OWGR — the Official World Golf Ranking — is the ranking system for male professional golfers, born in the mid-1980s. It is the gateway to the majors and to almost every elite event. How it works: a tournament with a strong field awards more points; points decay over time, usually on a two-year cycle; a player must meet a minimum number of events in that cycle. It sounds dry, but this formula decides who gets into majors, who gets into signature events, and who keeps their tour card.

This is where a problem emerges that I consider the earliest notable signal of the past few years: the conflict between LIV Golf and the traditional ranking system. LIV Golf, backed by Saudi Arabia's Public Investment Fund (PIF), operates on a 54-hole format with a team element. LIV applied for OWGR points and was rejected in October 2026. The stated reasons involved the 54-hole format, the team nature, and event classification.

What is worth thinking about is that both sides are speaking the language of data while defending two different definitions of what creates value in golf. One side believes ranking must rest on verified fields and an open competitive standard. The other believes capital and star power are the real measure of value. When two frames of reference clash, what gets caught in the middle is the golfer's career — and the viewer's trust.

In June 2026, the PGA Tour and PIF announced a framework agreement, shocking the golf world, when only a short time earlier the two sides had been in open conflict. That agreement showed that market data — broadcast rights, capital flows, and the ability to control top stars — carries no less weight than a ranking table.

Ball Rollback: when the rules become data too

Another battle over the yardstick is the story of the golf ball flying farther, known as the ball rollback. The USGA and The R&A announced a change to the ball specification in December 2026. Under it, from 2028 the rule applies to elite competition, and from 2030 to the rest of the golf movement.

This topic is often oversimplified into the question "does the ball fly too far?". But the real issue runs deeper: regulators believe driving distance has grown to the point of changing the meaning of many classic golf holes, threatening the collective memory of the sport. Holes designed to test a long second shot are now "flown over" by a drive. The course becomes a shorter equation, but also a poorer one in narrative terms.

I once argued about this with a friend who works in equipment engineering. He said simply: if you want to reduce distance, change the specification. I countered: if you change the specification, you may be fixing a symptom rather than a cause, and you are creating two different definitions of the sport — one for professionals and one for amateurs. That argument has no winner. But it is exactly the kind of argument I enjoy: no absolute right answer, only consequences to be weighed.

At the data level, the decision matters because it starts a technology-transition cycle: manufacturers must redesign balls, golfers must adjust spin and club choices, and entire distance-data models become obsolete. When the yardstick changes, every old comparison table becomes a historical document.

The counter-intuitive angle: the window of empty data tables

Here I want to return to the opening story and expand it into a larger finding. What I realized after years of reading analytical tables is not that data often lies, but that data is often read at exactly the moment when it has nothing to say.

Picture a perfectly presented golf analysis: it has structure, a table of contents, a taxonomy, a slot for every metric, a risk framework, a conclusion section. It looks so professional that the reader easily forgets the first question: what is the input data? If the list of information points is empty — no golfer, no event, no hole, no metric — then a formally perfect analysis will contain nothing but cells reading "insufficient information to assess".

This is the point I find most counter-intuitive and most worth discussing. A golf analysis can be beautiful in every structural respect yet empty in every conclusion, and that condition is not the fault of the table — it is a failure at the ingestion stage. When ingestion fails, every downstream analytical layer collapses like dominoes: no technical subject can be identified, no form curve can be built, no event can be placed on the tournament ladder, and no analysis of governance, rules, risk, narrative, or transmission flow can proceed.

I once saw a technical golf analysis with full sections for SG: Off the Tee, SG: Approach, SG: Putting, course fit, and key metrics — but every cell read "insufficient data". That sounds useless, but it is actually a mirror. It forces me to check my own reflex: am I reading data, or am I reading the feeling the data was designed to evoke? I increasingly believe the most important skill of a sports commentator is not analysis, but checking the input before analyzing.

There is a test I apply to any golf data table before trusting it: count the verifiable information points. If that number is smaller than the number of conclusions the table delivers, then the table is borrowing against its credibility, and one day it will have to pay. This is the principle I drew from my own failure on the track: a runner who champions improvisation can produce a breakthrough, but that same improvisation kills performance when there is no discipline behind it.

What really stands behind a good golf data table

If the algorithm matters, what matters more is the input data. In golf, input quality comes from three sources: the course, the weather, and how the shot-tracking system is operated.

The course creates context. The same shot from 150 yards has different value on a firm, fast-running fairway at a coastal links course than on a damp, soft fairway inland. If the baseline database does not separate course types, your Strokes Gained mixes apples with oranges. This is why shots at The Open — the traditional championship on links courses — often require their own interpretation.

Weather creates variance. Wind affects every shot, and wind at a coastal course can blow in different directions on each hole. A data table that does not record wind conditions is a table that cannot be reproduced.

How the system is operated creates systematic error. If ShotLink is only installed at major events, then any conclusion about golfers on smaller tours is limited. If a tournament has few fully tracked holes, an already small sample grows smaller. And if the data operators carry bias, then even raw data can be shaped.

I once looked at a comparison of two golfers based on the previous season and wondered whether anyone had checked whether they played under the same course conditions. The answer, I suspected, was no. Golf analysis usually runs faster than verification, and this is where beautiful data becomes risky data.

Is OWGR fair?

This question sounds political, but it is really a question about data. A ranking system does not arise naturally; it is the product of a series of decisions about weights, time windows, and minimum rules. Change any component and you change the order on the board.

For example, if you increase the weight for events with strong fields, you are assuming competition is the only measure of quality. If you increase the weight for number of entrants, you are assuming that scale equals greater difficulty. Every choice is reasonable within a certain logic, and every choice creates a loss or gain for a specific group of golfers.

Young golfers, golfers returning from injury, and golfers playing on less-covered tours are the groups most affected by how these weights are designed. This explains why debates about OWGR tend to run hotter than debates about Strokes Gained: Strokes Gained measures shots, while OWGR decides opportunity.

And here is the point I always raise: a ranking is not the truth, it is a hypothesis signed by people with power. When that hypothesis stops being treated as a hypothesis and starts being treated as truth, the sport loses part of its capacity for self-criticism.

The things I see that the data table does not

If I had to name one skill golf data cannot replace, it is the ability to observe a body and its rhythm. Golf, though it looks static, is a sport of physicality and nerves. A golfer tiring at the 15th hole will reveal it in clubhead height, in the moment of impact, and in how he reads a putt line. No metric in a Strokes Gained table records that.

I once watched a golfer play superbly over the first two rounds and collapse over the final two at a major, and to this day I remember how his shoulders began to drop from the 12th hole of the third round. The data would say he lost Strokes Gained on Approach over the final two rounds. The data would not say he lost rhythm after four straight days under pressure and after a hooked shot on the 13th that he could not forget.

This is not a call to turn away from data. It is a call to read data alongside experience. The empty summer stands of 2026 taught me to hear a match with my heartbeat, not with the sound — and golf, more than any sport, is best watched with both at once.

The putts that never appear in the table

Strokes Gained: Putting is one of the most misunderstood metrics. It compares a putt's outcome with the average from the same distance, but it cannot measure the weight of the moment. A three-meter putt on the 8th hole of the first round has the same statistical value as a three-meter putt on the 18th hole of the final round — but its psychological value is not the same.

When analysts say "putting is an unsustainable metric", they are right in principle. But that leads to a wrong conclusion if applied mechanically: that success in majors depends less on putting. In reality, the majors — hard courses, strong wind, fast greens — are exactly where short putts decide the winner. This is the kind of contradiction between aggregate data and situational data that I love to mine.

I have followed many final rounds and seen a recurring pattern: the winner gains from strong approach play across the week, but the person who decides the title is usually the one who putts well in the final two hours. The data table is correct but easily misleads when not combined with tournament structure.

Golf in Vietnam and the question of domestic data

I live in Hai Phong and work with golf in Vietnam, so I take a particular interest in the question: where is Vietnamese golf data right now? The honest answer is that it is in a formative stage. We have quality courses, amateur events and some professional events, and rising young golfers. But we do not yet have a shot-tracking system on a scale comparable to ShotLink, nor a database thick enough to compare Vietnamese golfers against themselves over time.

That is not just a technology issue. It is an issue of who creates the data and for whom. If data serves only short-term media, it struggles to hold long-term value. If data is built for golfer development, it can change the training curve. I believe this is one of the biggest gaps in Vietnamese golf today — and it is an opportunity, not a dead end.

As the Vietnamese golf movement expands toward younger groups, the need for basic data infrastructure — at minimum ball position, shot type, hole outcome — will soon become urgent. Without it, we will keep judging golfers by emotion and by a few crude metrics such as stroke count, and we will repeat the very mistakes international golf went through more than a decade ago.

The most worrying thing is not data, but faith in it

There is a paradox I have come to see more clearly. When data is scarce, people distrust it. When data is abundant, people trust it entirely. Both attitudes are equally dangerous.

Strokes Gained, OWGR, ShotLink data, course-fit models — all are tools that reduce uncertainty, but none eliminates it. Golf is a sport where a sudden gust, a shot into a bunker, or a moment of shaky hands can reverse any forecast. That is not a weakness of the sport; it is its nature.

To me, a good commentator is not someone who predicts the result correctly, but someone who honestly describes the state of uncertainty. When the data says one thing and my eyes say another, I have learned to hold both and wait. That is the discipline I drew from the track: trust neither a single feeling nor a single equation.

I used to think golf was a sport of numbers. Now I think golf is a sport of numbers placed in the right positions, and the real skill is knowing where to place them.

An open ending

If there is one thing worth carrying away from this story, it is a principle for reading golf data: always ask where the data comes from, what it measures, and when it expires. From the failed starting line to the commentary booth: every scar is a map. And the map I trust most does not chart the winner; it charts the places where I once read things wrong.

Golf is entering a period where data is no longer a supporting tool but foundational infrastructure. The next question is not whether to use data, but: who owns it, who verifies it, and which story gets pushed aside when a single number is chosen to stand in for the entire truth. When you watch a final round this weekend, try to remember this: the data table on the screen may be right in every column, but the real story is happening on a hole that no metric can measure.