Golf and the Data Paradox: Distance Rises, Scores Stay Flat

ShotLink logs roughly 40,000 shots per week at a PGA Tour event. In the 2026 season, the tour's average driving distance reached 297 yards, up nearly 25 yards from 2026. Over those same two decades, greens-in-regulation held almost flat, and the average score per round fell by only about a stroke and a half. One data line surged; two stayed level. Golf got longer without getting easier.

When I opened that week's analysis file — eight sections, correct format, labeled "golf" — every content field was empty. No number, no name, no date. Such a report cannot be called a conclusion; it is a gap presented neatly.

Reading golf through data requires understanding the collection system behind it. ShotLink is the PGA Tour's shot-tracking system, running since the early 2000s on lasers and on-site volunteers. Every shot records its start point, end point and distance remaining to the hole. In 2026, the PGA Tour made Strokes Gained — a metric measuring a shot's stroke advantage against the field average — part of its standard dataset, after Mark Broadie, a Columbia University professor, published the method in his book Every Shot Counts. Since then, how a golfer is evaluated changed at the root: people stopped counting putts and began measuring the value of each shot against expectation.

Based on my experience tracking rounds, most golf arguments skip this data layer. In Nha Trang, where I live and work, amateur and regional professional events rarely have shot-tracking systems. People remember the deciding putt, not the approach that set it up. Data is never in a hurry; it waits for someone who knows how to read it.

Governance moves to its own rhythm. In June 2026, LIV Golf launched with backing from Saudi Arabia's Public Investment Fund (PIF), playing 54 holes, shotgun starts, no cut, with team scoring. In June 2026, the PGA Tour, DP World Tour and PIF announced a framework agreement. In October 2026, OWGR — the world ranking founded in 2026, calculated on a rolling two-year window — refused LIV points. Those three dates reshaped the board every golf data report must stand on.

The first piece of evidence lies in the Strokes Gained breakdown. When Broadie splits a professional's scoring into four groups — off the tee, approach, around the green and putting — most of the gap between players comes from approach shots, not putting. Approach shots decide standings more than putts do, even though fans only remember the putt. This is golf's oldest blind spot: the most visible thing is not the most important one.

The second piece of evidence is the distance paradox. From 2026 to 2026, the PGA Tour's average drive rose nearly 25 yards, per ShotLink. If distance were the variable that decides scoring, average scores should have fallen accordingly. They did not. The reason lies outside the swing: courses were lengthened, rough grown thicker, greens made flatter and faster. Organizers offset the advantage technology delivered. Rising distance and rising course difficulty ran in parallel, canceling each other out.

The third piece of evidence belongs to governance. OWGR refused LIV because the 54-hole, no-cut, team format does not fit the current points method. That decision pushed a group of elite golfers out of the ranking system, weakening the very sample used to compare generations. When Jon Rahm left the PGA Tour for LIV in late 2026, a top-tier observation vanished from the standard data series. The Tiger Woods era, with 15 majors, gave way to a generation led by Scottie Scheffler, but the baseline for comparing those two generations has thinned.

The fourth piece of evidence sits at the equipment layer. The USGA and R&A announced a change to the Overall Distance Standard (ODS), applying to elite play from 2028 and to recreational players from 2030. The rules bodies chose to intervene on the ball rather than the club, because the ball is the easier variable to control and disrupts the club market less. Behind a technical decision is a commercial calculation: changing the ball is cheaper than changing an entire club industry.

This chain of evidence transmits downstream along a clear path. If the ball flies shorter from 2028, the value of approach play and distance control rises; ShotLink data will have to recalibrate its baseline. Equipment makers adjust product ranges. Broadcasters and data providers reprice rights packages. A change at the rules layer pulls a chain of repricing through the market layer.

The counterintuitive angle sits here: we tend to overprice the variable we measure best. ShotLink measures driving distance almost perfectly, accurate to the yard. It measures a course's difficulty far less well. So every argument that golf is being ruined by distance rests on the best-measured variable, not necessarily the most important one.

Correlation is not causation. Distance rose at the same time golfers began training with launch monitors, building strength and optimizing launch angles. Equipment is only part of the equation. Blaming the entire cause on the club is misreading the data: we are measuring one variable and inferring about a whole system.

The same error appears at the layer of judging people. A golf data model easily overvalues the long drive — the easy-to-measure variable — and undervalues course management, which is nearly impossible to measure. That is why golfers who win majors through course management are often valued below their true level by models. Being pushed out of the game is the fastest way to see the whole board.

The signal for the next cycle lies at two marks. The 2028 ODS timeline will be the first test of whether golf truly revalues approach play; ShotLink data after that year will answer. And if OWGR and LIV find a way to reconcile, the baseline of field strength will shift, forcing every Strokes Gained figure to be recalculated. I write the report, close the file, and the market reopens on its own.

A report sitting in a drawer is not a conclusion but a chart waiting for its time axis. When golf measures almost everything except difficulty, is it optimizing for what it can measure, or for what it actually needs?