Jim Furyk replaced the US team's analytics operation last week. Forbes reported on October 14 that Scouts Consulting Group, the statistical firm that supported the US in team competitions since 2016, has been let go ahead of the 2027 Ryder Cup at Adare Manor. Justin Ray and Hunter Stewart will lead the new group. "I love and I have a lot of trust in our analytics right now," Furyk said. Scouts had been in the room for four of the six matches since they arrived. The Americans won two of them.
The timing is not subtle. The US lost back-to-back Ryder Cups at Marco Simone and Bethpage, and the Solheim Cup fell to Europe in September. Angela Stanford, the US Solheim captain, told reporters afterward that in Spain the year before, the team had "people giving strokes gained in each match, even in practice." In the Netherlands they did not. Two events, two different captains, one recurring gap between what the data covers and what the competition demands.
Golf Digest made the gap concrete at Bethpage. Keegan Bradley sent Harris English and Collin Morikawa out together in foursomes, twice, both times against Rory McIlroy and Tommy Fleetwood. A pairings model from the analytics site DataGolf had ranked the English-Morikawa combination 132nd out of 132 possible pairings from Bradley's twelve-man roster. The Americans never made it past the sixteenth hole. "A golf analytics website is not the be-all and end-all," Golf Digest allowed. But when the math says a combination is dead last and the captain sends it out a second time, the distance between data and decision is visible from the gallery.
The interesting question is not why the captain ignored the model. It is what the model measured that the standard analytics did not. Stroke-play strokes-gained rankings, the kind published weekly on tour, tell you who played best over the last twenty rounds. They are built for individual performance across a full field over seventy-two holes. Match play asks a different thing entirely: who is better together, on this hole, against these two opponents. Individual rankings cannot answer that, because the question contains a variable they do not track.
DataGolf's pairings model adds what it calls a compatibility term. The strongest foursomes partner is not the player with the highest ranking. It is the player whose strengths do not overlap with yours. If one player dominates long par fours and the other handles par threes, the pair covers more of the course than two players who are both excellent from 180 yards. The estimated compatibility effect is small in isolation — a few hundredths of a stroke per hole — and in alternate shot, where a single mistake on any swing can end the hole, small edges compound fast.
Europe recognized this earlier. Twenty First Group, the analytics firm that worked with Thomas Bjorn's 2018 team, built qualifier projections months ahead of the event so that Bjorn could test pairings before the roster was set. The US operation, by contrast, has historically waited for the team to form and then optimized lineups from existing data — which means optimizing from stroke-play numbers that were never designed for match play.
Furyk has also shortened the qualification window. Points for the 2027 team start at the American Express in January and stop when the roster locks. The old rolling two-year cycle rewarded sustained performance. The new system selects twelve players in peak form and gives the analytics team months to study how they fit together before they arrive in Ireland.
Whether this solves the problem is open. Ray and Stewart have eleven months and a roster that might not take shape until summer. What they have that prior operations did not is a captain willing to restructure the team around the question the data is supposed to answer, rather than the other way around.
The lesson is not limited to team golf. Every golfer with a launch monitor tracks carry distance and spin rate. Those numbers are real. But the question they answer — how far does the ball go when I strike it well — is not the question that determines the score. That question is closer to how far the ball finishes from the hole, on average, from 140 yards. Proximity, dispersion, the gap between a good swing and a missed one. The data that matters most is often the data nobody collects, because the obvious numbers feel sufficient.
Numbers tell you what happened. The right question tells you what to change.