When the Golf Data Sheet Comes Back Empty: Strokes Gained, ShotLink and the Trap of Evidence-Free Analysis

On a morning in Hai Phong, I reopened a golf analytics file that had been set to run overnight. The first page listed exactly one field: domain — golf. The second page: article title, none. Source, none. Player, none. Venue, none. Information points, none. The file ran dozens of pages, and nearly all of its content was instruction text telling the reader how to fill in the content, lifted verbatim from a template shell. A sports analytics system had returned an empty result while preserving the polished appearance of a finished piece of work.

Nine years in commentary work taught me to hunt for errors in numbers. One unit off, a wrong sample, a lopsided comparison between two events of different tiers — those are familiar kinds of lying and I usually catch them. Every statistic is capable of lying; my job is to catch it in the act. But the lie that surfaced that morning belonged to a far harder category to detect: a gap presented as if it were content.

The story looks like it is about a corrupted file. It lands precisely on the sorest point in golf analytics today, in the West and in Vietnam alike: we have memorized the vocabulary of data before we have enough data to speak.

How Strokes Gained Actually Works

Strokes Gained emerged from the work of Mark Broadie, a professor at Columbia Business School, who in the early 2010s began recalculating every shot on the PGA Tour around a single question: from this position, how many strokes does an average tour player need to hole out? The PGA Tour made Strokes Gained an official statistic from the 2026 season, built on ShotLink, the shot-tracking system deployed from the early 2000s.

The mechanics are simple enough to be routinely misread. A shot from roughly 150 metres in the fairway carries an expectation of about 2.98 strokes to finish the hole. If the player puts the ball on the green five metres from the pin, the remaining expectation is about 1.8 strokes. That shot is scored as 2.98 minus 1.8 minus 1 — a gain of 0.18. If the ball finds a bunker and the remaining expectation is 2.6, the shot is worth negative 0.62.

The key word is expectation. Strokes Gained does not measure the shot. It measures the gap between what was expected and what happened, against the tour-average baseline. A negative shot does not mean a bad shot in absolute terms; it means the shot left a worse situation than average. That distinction is skipped in a great many Vietnamese write-ups that quote the statistic.

The four main categories are: shots off the tee on par 4 and par 5, approach shots into the green, shots around the green, and putting. Adding the first two gives long game; adding all four gives total. Broadie's best-known finding, after analysing millions of shots, is that at professional level long game explains roughly two thirds of the scoring difference between players, while putting explains roughly one sixth.

Put another way, the sport the public still imagines as a battle on the greens is largely decided on the ground in front of them. That truth sounds jarring to mainstream coverage, and it is also the most consistently ignored point in domestic golf reporting.

The Trap Lies in Stability, Not in Accuracy

This is where the story turns absurd. Strokes Gained is meticulous in measurement, but the stability of each category varies enormously season to season. Putting is the most volatile group and carries the lowest year-over-year correlation. Approach-the-green is far more stable, and therefore far more predictive.

The consequences are concrete. A player wins an event on a scorching putting week worth more than two strokes gained per round. Coverage immediately crowns him the new putting king. The following season he ranks ninetieth in putting and is declared to be in decline. Both labels are wrong. That putting week was a small sample sitting in the tail of a distribution, and regression to the mean simply reclaimed what it had lent.

Scottie Scheffler is the instructive counter-example. His foundation sits in approach play and chance creation, not in putting. When his putting dipped during one stretch, most of the media conversation circled the putter, while the thing actually holding him at the top was the most stable category of all. Rory McIlroy's career follows similar logic: driving and approach are the bedrock, and putting is the variable that decides whether he wins or finishes second. Jon Rahm, during his peak PGA Tour years, was also best described through approach-the-green figures and the ability to control ball flight in wind.

No ShotLink, No Strokes Gained

The VGA Tour took shape as Vietnamese golf exploded in course numbers and player numbers. Across roughly a decade, the national course count grew from a few dozen to more than seventy, pulling behind it a domestic professional schedule, amateur events, and a young generation of players such as Nguyen Anh Minh, for years Vietnam's top-ranked amateur and a representative of the country at continental level.

What has not kept pace is the data infrastructure. ShotLink requires a crew charting every shot on every hole, measuring equipment, and a standardized system running across the entire event. At the world's leading professional events, that cost is treated as non-negotiable. At most regional and domestic events, it does not exist.

The consequence is simple and rarely stated: any Strokes Gained figure attached to a domestic tournament must come from a charting system operated by the organizer. Without that system, the statistic does not exist. A piece claiming player X led the field in approach play, with no charting source cited, is technically a story assembled out of thin air.

I am not saying this to nitpick. I am saying it because I once did exactly that.

Three Kinds of Error, and the Most Dangerous One

There are three ways to get golf data wrong. The first is fabrication: assigning a statistic to a player with no verifying source. This kind is easy to expose and usually collapses on its own. The second is silence: no analysis at all, just a retelling of events. Safe, useless, and readers leave.

The third is the frightening one. It is the silent null — a system, an article, a report that returns an empty result while still formatted as a finished product. The full skeleton, the serious headline, the sections in the correct order, and not one piece of information inside. A skimming reader takes it for analysis. A busy editor approves it for publication.

The third kind is dangerous because it wears the clothing of rigour. It does not lie with a wrong statistic; it lies with an absence presented as a presence. In data work this is called null propagation — a failure at the ingestion layer is never blocked, flows straight into the conclusion layer, and reaches the reader as an assertion.

In golf this failure mode is unusually likely, because the sport has a prestigious statistics architecture used by global media, and quoting it produces an instant impression of expertise. Nobody checks whether the number is real, because the outer shell is already convincing.

When the Data Goes Quiet, You Listen With Your Eyes

The empty stadiums of summer 2026 taught me to hear a match through a heartbeat rather than through sound. Every competition was suspended that year, and I sat in Hai Phong re-commentating old games, rebuilding the feel of a grandstand with nobody in it. That stretch made one thing clear: when there is nothing to lean on, you are forced to look harder.

For Vietnamese golf, that means returning to older tools. A notebook charting every shot by hole, by ball position, by club. The organizer's pin sheet to know where the flag sits that day. Video from a fixed camera angle. Together those three allow a hand-built Strokes Gained table with tolerable error, at least for approach play and putting.

I know this because I built a crude version of it. In 2026, still a maths-track school student in Hai Phong, I logged every sprint of a World Cup into a notebook and built my own statistics table. Nobody read it. But the act of writing by hand taught me to separate an observation sample from a feeling.

The Lesson From the 400 Metres

The 2026 fall did not stop me — it redirected the whole lane. I led my semifinal over 400 metres, cramped at the 350-metre mark, and finished last. What matters is that at 300 metres I held the fastest split in the group. Had someone stopped the data at 300 metres and written a report, that report would have concluded I was the fastest runner in the field. The data was not wrong. The reader of the data was wrong, because the race had been cut at the point most convenient for the conclusion.

Golf behaves identically. A player can lead after three rounds and collapse in the fourth. His putting figure through 54 holes may top the field; that figure predicts nothing about Sunday. Writing that he is the field's best putter on Saturday night is not incorrect. But the reader on Sunday night sees a completely different story, and loses faith in the writer rather than in the data.

From a failed starting line to a commentary booth: every scar is a map. My scar at 350 metres taught me that any data sample must be read alongside a question about how long the race actually is.

The Counter-View: Data Poverty Can Be an Advantage

This is where I argue against myself, because the claim that Vietnam lacks data slides too easily into the claim that Vietnam must buy an American system at any cost. I do not believe that.

Data density in the West has produced a generation of analysts who read statistical tables fluently but can no longer sit still for three hours to watch a round. They know a player's Strokes Gained. They do not know how he reads wind on a coastal hole, how he picks a club with the ball on a downhill lie, or how much his hand shakes over a short putt on Sunday.

None of that sits inside any charting system. A substantial share of golf's decisive value sits outside the measurable zone: reading terrain, managing bad situations, absorbing pressure. If golf analysis becomes only the reading of statistics, it shifts from observation into retelling.

The absence of ShotLink forces Vietnamese practitioners to retain the skill their counterparts in data-rich markets are gradually losing: looking by eye and recording by hand. The condition is a willingness to say I do not have that statistic, instead of constructing a plausible-sounding approximation.

Who Owns the Gap

Back to the empty file on my screen that morning. The fault sits in the data-collection layer, but the consequence lands on readers if someone decides to publish it. In sports analytics workflows this is the most overlooked failure point: there is no gate requiring at minimum one information point and one proper name before moving into interpretation.

In golf specifically, the missing gate produces three practical outcomes. Conclusions about form get published with no data behind them. Figures from events of different tiers get blended, so a strong number at a low-tier event reads as proof of elite quality. And players get described with sticky labels, when labels only hold for a very narrow window.

Where the data infrastructure is complete, all three outcomes still exist; they simply arrive later and in subtler form. Where the infrastructure is thin, they appear in the very first sentence.

A Concrete Method for Next Season

There is a low-tech fix any Vietnamese golf media team can try. Pick one hole. Chart every shot by the final group on that hole, every day, for the whole event. Record ball position, distance, club if visible, and hole outcome. After four days you have a small dataset that is real.

A small real dataset always beats a large constructed one. One hole, four days, three groups, twelve players — that is roughly a hundred and forty-four analysable shots, enough to compare average effectiveness from fairway against rough on that specific hole. Not enough to declare the best ball-striker in the field. That modesty is exactly what makes the conclusion credible.

What Remains After the Data Disappears

I tell the story of the empty file because it is a miniature of a habit spreading through the sports industry: we are teaching each other how to look as though the analysis is done.

Beneath that shell, Vietnamese golf sits at a fairly unusual moment. There are domestic tournaments, a young generation moving onto the continental stage, and a new audience discovering the sport through driving ranges and amateur events. What is missing is a culture of recording. A culture of recording needs no big budget and no expensive equipment, only one person willing to sit in one place long enough.

If someone on the VGA Tour next year spends four days behind the 18th green with a notebook, they will produce something no imported system can buy: the first dataset written by Vietnamese people about Vietnamese golf. And once that notebook is full, the next question will no longer be who hits it best, but whether we have the patience to read back everything we just wrote down.