A statistical breakdown on the Brooks-Rory Player of the Year debate

Even Rory McIlroy was surprised by his 2019 PGA Tour Player of the Year award, an honor determined by the vote of PGA Tour players. The number of votes received by each nominee is not disclosed, only the winner, though Rory publicly stated that he had voted for Brooks Koepka. Was this outcome predictable based on player records? The question boils down to weighing Koepka’s three wins (including the PGA Championship), top major finishes and nine top 10s against McIlroy’s wins at the Players, Tour Championship, RBC Canadian Open and 14 top-10 finishes.

To investigate, I looked at Strokes Gained: Total data from 1996 to 2018. As a predictive model, this stat would have correctly placed only 14 of the past 23 winners. That’s not a great record, and it’s not surprising, either. In strokes gained, a one-stroke difference between first and second counts the same as a one-stroke difference between places 49 and 50.

This forced me to pivot and look for a measurement that placed an appropriate premium on winning, a larger reward for winning a major (or the Players Championship), a reward for top nonwinning finishes, plus extra credit for wins against stronger fields. The Official World Golf Ranking points gained in a year fit the bill. For all of the deficiencies of the OWGR, it does a pretty good job at the top, and points gained in a year takes into account all of the factors just mentioned. Case in point: OWGR points gained would have correctly predicted 17 of the past 23 POY award winners.

To improve the model’s predictive power, I added the number of major wins, a Players win indicator and the number of other wins as factors on top of OWGR points gained. These factors combined would have correctly predicted 22 of 23 winners. (The only miss was in 2008, when the model predicted Tiger Woods but the award went to Padraig Harrington.) This means that in POY voting, those casting ballots place even more emphasis on major victories and other wins than the OWGR computations.

The Winner is…

When applied to data from the 2019 season, this model, by a razor-thin margin, gives Brooks Koepka the edge over Rory McIlroy for Player of the Year. While the award going to McIlroy was a surprise, the close margin of my model suggests that the current voting process is far from broken or that Rory won on the basis of popularity as some suggest. I’m not sure why the PGA Tour shrouds vote totals in such secrecy. It’d be interesting to know how close was the final tally.