Strokes Gained and the Verification Gap in Modern Golf Data

A 1.2-metre putt on the 17th hole on a Sunday is rarely a technical story. It is a story about breathing rhythm, about a green speed that changed since morning, about crowd noise behind your back, and about a wrist that shook for exactly one thousandth of a second. But on the big screen in the media centre, the only thing that appears is a line of text like "SG: Putting +1.8". The room nods, as if that line had finished the whole story.

I have sat in the media centres of PGA Tour events many times, holding a still-warm ShotLink printout, then looked up at the broadcast graphic. The two rarely match. Not because anyone is lying, but because they are produced by two different data pipelines, with two different baseline definitions, and almost nobody tells the audience that.

Where people assume there is only passion, I find the mathematics of the ball. And there, I learned that mathematics is only as trustworthy as its transparency.

Golf learned to count later than everyone else

Golf entered the data era later than baseball and basketball. Through the 1990s, a tournament's statistical book contained only a few crude figures: fairways hit, greens in regulation, putts per round, average driving distance. Those metrics counted outcomes; they did not measure quality.

The turning point came from two directions. The PGA Tour put ShotLink into operation in the early 2000s, a system of lasers, cameras and volunteers recording every shot at every event. At the same time, Mark Broadie, a professor at Columbia Business School, built the Strokes Gained method and published it systematically in Every Shot Counts in 2026. The core idea is simple: each shot is scored by the average number of strokes a tour-level player needs to finish the hole from that exact distance and lie. Better than average adds; worse subtracts.

From there, four categories became a common language: Off the Tee, Approach the Green, Around the Green, Putting. Television likes them because they are compact. Analysts like them because they separate skills. Young players like them because they say what to practise on Tuesday.

Everything comes with a price rarely mentioned: when a sport turns human behaviour into metrics, it also turns metrics into targets. Bryson DeChambeau won the 2026 US Open at Winged Foot by turning his body into a speed machine, then won again in 2026 at Pinehurst No. 2 with a more balanced version. Between those two markers, a generation threw itself into the distance race. In December 2026, the USGA and the R&A announced a reduced-distance ball rule for elite competition, due from 2028, a direct intervention in the very thing the data had rewarded.

The data pipeline does not cover the whole golf course

Strokes Gained is not a measurement. It is a comparison. There is no absolute "SG", only "SG against some baseline". That baseline is built from the data of a specific group of players, on a specific course system, in a specific period. Change the baseline and the number changes.

The PGA Tour publishes SG from its own ShotLink data. Independent platforms build their own models. Sensor systems fitted to club grips collect data from recreational players. Three sources, three baselines, three different results for the same round. None is technically wrong. But when an editor places a figure from one source next to a figure from another to tell a story, the audience is watching an invalid addition.

The deeper problem is coverage. ShotLink covers almost the entire PGA Tour, but the DP World Tour, the LPGA Tour, major amateur events and most of Asia's professional circuits record far more sparsely. When a player moves from Europe to the United States, his SG numbers do not simply improve. They are measured with a different ruler. Any cross-system comparison needs a caveat attached, and that caveat almost never makes it to air.

Thirty putts are not enough to say anything

This is where golf data collides with probability. A round has 18 holes. A player's putt count in a round usually falls between 25 and 32. In SG: Putting, round-to-round variance is large enough that one good putting week may not reflect skill at all; it may simply reflect the ball finding the hole a few more times than average. A bad putting week, likewise, is not proof of decline.

Long-run tour data shows a fairly stable trend: SG: Approach the Green is the most durable category across seasons, while SG: Putting is the most volatile. In other words, the metric television mentions most is the one that predicts least. The paradox is not in golf. It is in how broadcasters choose numbers: they prefer what is easy to narrate, not what is trustworthy.

There is also what the system cannot measure. A 40-centimetre putt is recorded identically to a nine-metre putt in the putts column. Putt distances in ShotLink are measured from an observation tower, and in real conditions some shots are still flagged as having an imprecise distance. The foundation of every SG: Putting calculation therefore carries a layer of systemic noise nobody publishes.

What actually predicts

Based on my experience watching tournaments live across many seasons, if I had to trust one metric over the long run, I would choose SG: Approach. The reason is structural: the approach shot sits under the greatest technical control; distance, trajectory, spin and angle of attack belong far more to the player than in putting, which depends on terrain and green speed, or driving, which depends on wind and hole design. That is why Scottie Scheffler won the Masters in 2026 and 2026 and Olympic gold in Paris in 2026 with a dense run of titles, while most of his edge lives in approach play rather than putting.

Rory McIlroy is the fascinating counter-example. Four majors, but his most recent major title dates to 2026. He has SG numbers good enough to win more. What stops him sits in the zone the tables cannot measure: the closing holes, the short putts, the tactical calls at the decisive moment.

A number never tells the whole story, but it always knows how to open one. Golf data describes very well what already happened. It predicts far less well who will win. And it is nearly blind to why one player can do it on a Sunday and another cannot. Tiger Woods, with 82 PGA Tour titles, level with Sam Snead's record, has never been fully explained by any statistical table.

The world ranking is also a data product

The OWGR is not a physical measurement. It is a formula: event strength, field quality, rounds played, points decay over time. Every change to the formula creates winners and losers. When LIV Golf emerged and clashed with the PGA Tour system, the ranking-points question became a genuine governance dispute. In March 2026, LIV withdrew its application for OWGR recognition. To fans it was a short news line. To a 28-year-old professional, it could be a major-championship door closing.

Nothing in that chain is natural. All of it is design choice. And every design choice can be examined, argued over, and should be published clearly.

The blind spot of the digital arms race

The biggest blind spot of modern golf is not a shortage of data. It is too much unverified data wearing the appearance of absolute precision.

In analytics there is a lesson worth remembering: when a data pipeline breaks, the most dangerous error is not the absence of a conclusion. The most dangerous error is a conclusion that sounds entirely reasonable, generated from an empty source, and impossible to falsify afterwards. A clean table is more persuasive than a messy truth. In golf, where every shot can become a line of metrics, that risk multiplies.

One small broadcast detail is worth noticing. The ball flight appears on screen with a perfect arc: not recorded footage, but a simulation graphic, built to look good rather than to be right. It is a fairly complete metaphor for the whole industry: presentation is prettier than measurement, and measurement is checked less often than presentation.

A verification standard is needed

I do not want golf to return to an era of fairways hit and putts per round. I want the opposite: more data, accompanied by a public verification standard, with published baselines, open methodology, stated coverage, and the places where the system does not know said out loud rather than blurred away.

The sports world is not fair, but it always hands you a microphone to tell the truth. When the curtain falls, the truth begins. Golf data will only be genuinely trustworthy on the day the people presenting it agree to talk about what they cannot measure.