Strokes Gained and the Trap of Beautiful Data: What Golf Is Measuring Wrong

In swimming, one hundredth of a second is one hundredth of a second. The scoreboard flashes and the argument ends. On the track, Usain Bolt's 9.58 seconds in Berlin in 2026 needs no further interpretation. Golf breaks that rule. A round of 68 can be an act of total control from tee to green, or it can be a day rescued by the putter. The same number, two opposite stories. That makes golf the hardest sport I have ever covered to read at a glance. No scoreboard tells you who is actually playing better. There is only the scorecard, and the scorecard is the final output of hundreds of small decisions it never records. Which is why I spend most of my working hours reading Strokes Gained tables instead of leaderboards. And why I have grown increasingly suspicious of the very data I once trusted without question. Strokes Gained came out of research by Mark Broadie, a professor at Columbia Business School, published in his 2026 book Every Shot Counts. Instead of counting strokes, he measures the value of each shot against the tour average from the same distance and the same lie. The PGA Tour collects the data through ShotLink, with tracking equipment on every hole recording ball flight, distance and where the ball came to rest. The output is four main columns: Off the Tee, Approach, Around the Green, Putting. Add them up and you get Tee-to-Green and Total. It sounds scientific. And it is scientific, until people start using it wrong. On American broadcasts, Strokes Gained appears as a graphics layer over the screen: a horizontal bar, a few coloured dots, a number with a plus or minus sign. Viewers are shown who is better than whom over that stretch of holes. They are not told that the number only means something against a full season of history, that a player sitting at minus 1.2 strokes on approach can still be the tenth-best approach player on tour. Broadie's research produced a conclusion many found uncomfortable: roughly two thirds of the scoring difference between professional golfers comes from the long game, meaning driving and approach play. Putting accounts for only about 15 percent. Around the Green adds roughly 10 percent. The rest lives in things that are hard to measure, such as hole management and shot selection. In other words, the sport the media calls a battle on the greens is actually decided from 150 to 200 yards out. I checked this against Scottie Scheffler's 2026 season. He won nine times that year, including the Masters, the Players Championship and Olympic gold in Paris. Scheffler's putting was average, with stretches that were plainly poor. Yet he led the tour in SG: Approach and SG: Tee-to-Green for most of the season. The iron, not the putter, is what carried him to the top. Compare that with Tiger Woods at the 2026 Masters, his fifteenth major title at the age of 43. Woods did not hit the ball far that week. He hit more greens than anyone in the final group and barely made a positional error. It was a victory of distance control, not of power. So why does the public still believe golf is a putting story? Because the putter creates moments. A six-metre putt that breaks an opponent is what makes the highlight reel. A seven-iron approach from 175 yards to three feet is just a line on a table. Media rewards emotion, and emotion lives on the green. That is the biggest blind spot in golf analysis today: we measure very well, but we tell it very badly. There is another problem: sample size. A player's SG: Putting swings violently from week to week. One 18-hole round supplies about 30 putts, far too few to conclude anything. Serious analysts wait for 40 to 50 rounds before they will speak of a trend. But news desks need content every night. So three rounds of data get packaged as form. In another sport, I have seen the same pattern with distance-covered figures in football. A midfielder who runs 12 kilometres is praised as a workhorse. But if four of those kilometres are aimless lateral running, the number merely reflects waste measured with precision. Golf has its own version. A player who hits 14 greens but leaves the ball 15 metres from the pin every time will post a beautiful greens-in-regulation figure and a terrible score. Proximity to hole is the metric that tells the true story. Where people think there is only passion, I find the mathematics of the ball. But mathematics is only right when the sample is large enough and the question is specific enough. Another common confusion is equating Strokes Gained with scoring average. The two metrics answer two different questions. Scoring average tells you how many strokes you took. Strokes Gained tells you how many strokes you took relative to expectation from the positions you put the ball in. A player who keeps finding the rough but saves par with the putter will post a very high SG: Putting and a negative Total. He is covering mistakes on one hole with skill on the next. Over time, that way of playing does not hold. The distance arms race is the clearest example of both the power and the limits of data. Average driving distance on the PGA Tour has climbed steadily for three decades, from around 270 yards in the 1990s to above 300 yards at the top end today. Three factors drive it: player conditioning, club and ball technology, and the way data rewards distance. Strokes Gained confirms that hitting it far helps. It does not confirm that hitting it far always helps. On narrow holes, that benefit is cancelled by the probability of finding rough, and the season-long summary does not separate the two contexts. Course context makes the picture harder still. At TPC Sawgrass, where Pete Dye designed narrow corridors and lurking water, the choice off the tee matters more than raw power. At Quail Hollow or Torrey Pines, driving distance translates directly into scoring advantage. The same driver, the same swing speed, two entirely different values. Strokes Gained records outcomes, not fit. That is why the concept of course fit exists, and also why it is so often abused to justify poor decisions. Wind and weather are the second variable. ShotLink records shots in real conditions but cannot separate how much of the outcome came from skill and how much from luck. An approach to eight feet in a 30 km/h gust at Augusta carries far more technical value than an approach to eight feet in still air. The stat table does not distinguish the two. The institutional backdrop muddies things further. Since 2026, LIV Golf has operated on capital from the Saudi Public Investment Fund, running 54-hole, no-cut events with large purses. The OWGR system, a two-year rolling ranking weighted by field strength, refused to award points to LIV because the format does not meet its criteria. The result: players such as Jon Rahm, who moved to LIV in December 2026 on a contract reported in the hundreds of millions of dollars, slid down the world ranking and lost major championship places. On 6 June 2026, the PGA Tour, the DP World Tour and the Saudi Public Investment Fund announced a framework agreement to merge their commercial interests. That process remains unfinished. This affects the data directly. When part of the elite field plays in a system without full ShotLink coverage, we lose the ability to compare. The PGA Tour's stat tables become a picture with holes in it, and every conclusion drawn from them needs a warning label. When the curtain falls, the truth begins. The most interesting part of this job is not the number that gets published, but the question of what does not. The counter-intuitive view I hold to be correct: golf does not lack data. Golf has too much data and too few questions. An entire industry is racing to collect more metrics: clubhead speed, attack angle, spin rate, ball speed, putt roll. Most of those new numbers only describe the shot that already happened; they do not explain why it happened. We know a player lost 0.4 strokes on approach on Saturday. We do not know whether he chose the wrong club because the wind shifted or because he misread the terrain. The biggest lesson of the past season is not in any column. It is this: when someone presents a very confident conclusion without citing specific data, it is most likely an assumption dressed up in terminology. I have received plenty of submissions to the newsroom, 1,500 words long, full of Strokes Gained and course fit, with not a single verifiable number. The structure was beautiful. The content was empty. I am not against data. I am against the habit of using data to end an argument rather than extend it. A good stat table should make the reader ask more questions, not nod in agreement. If an analysis leaves you feeling that everything is now clear, it has probably left out the most important thing. The world of sport is not fair, but it always hands you a microphone to tell the truth. The problem is that the same microphone is handed to people who have nothing to tell. The right direction for golf analysis in the coming years is not more sensors, but translating data into a language viewers can understand. Knowing Scheffler leads SG: Approach is a fact. Explaining why his irons work in a headwind at Augusta is knowledge. The purpose of analysis is not to replace intuition but to train it. I hold that as an editorial principle. Data is a map, not the territory, and a map is only useful to someone who knows what they are looking for. For me, that is also a reminder about the craft of writing. If an article about golf can be produced without watching a single round, it does not deserve to be read. And next season will answer an old question: when everyone has the same data, the difference will once again come down to the eye that knows how to look.