Statistical predictions for the FedExCup Playoffs
August 23, 2017
By Matthew and Will Courchene, Special to PGATOUR.COM
- August 23, 2017
Inside the PGA TOUR
2016-17 PGA TOUR season recap
This month, we are providing predictions for the 2017 FedExCup Playoffs, the four-tournament run that begins this week at THE NORTHERN TRUST. The predictions are probabilistic; using 30,000 simulations from our statistical model, for each player in this week’s field, we calculate their probability of advancing to the Dell Technologies Championship, the BMW Championship, the TOUR Championship, and finally, of being crowned the 2017 FedExCup Champion.
Additionally, we provide predictions for this week’s event at Glen Oaks Club.
Let’s get right to the predictions. (For those interested, details on our model are provided after the tables).
We have taken into account those players who have publicly stated their intentions to not play in certain Playoff events. Here are the relevant probabilities for the 30 players with the highest probability of winning the 2017 FedExCup:
The results are pretty interesting. First, the top 3 players (Hideki Matsuyama, Dustin Johnson, and Jordan Spieth) capture the lion’s share of the win probability for the FedExCup (52.1 percent!). This is really high; for context, at a major championship, the top 3 win probabilities would typically add up to about 14-17 percent.
Second, while a player’s starting rank is clearly important, there are some players who are further down in the rankings that have fairly high win probabilities (Jason Day, Roryt McIlroy). This is mainly due to the fact that to win the FedExCup, you likely have to win an event along the way (and a player like McIlroy is more likely to do this than some of the players currently ranked above him in the FedExCup).
Third, notice that the top 10 players in the current FedExCup rankings are guaranteed a spot in the TOUR Championship.
Next, we highlight the players who are near the bubble for advancing to next week’s Dell Technologies Championship. Specifically, here are numbers 83 to 118 in the current FedExCup standings, ordered according to a player’s probability of advancing to next week:
Notice that all players ranked 83rd or better in the current standings are guaranteed to advance to next week. According to our model, there are 10-12 spots that are really up for grabs (depending how much of a sure thing you feel 90 percent is).
Finally, we have also predicted each player’s probability of winning, finishing in the top 5, finishing in the top 20, and making the cut at THE NORTHERN TRUST. Here are the 25 players with the highest win probabilities this week:
Now, a bit about our model and how we simulate a round of golf. If you understand how to simulate a single round, then it’s not a big step to understanding how we simulate the FedExCup. Keeping things as simple as possible, our statistical model can be thought of as proceeding in two steps.
First, we assign each player an expected score, based off of various characteristics of the player. For example, two-year scoring average is one input, as is the player’s performance at his previous event (although the former is much more important than the latter).
Second, to simulate a round of golf with our model, we add a mean-zero random term to each player’s expected score. So, while Rory McIlroy has a better expected score than most other players, in some simulations he will lose to players who have worse expected scores than him because he got a bad draw of his random term (or the other golfers got very good draws).
Additionally, players who are more consistent are given less variation in their random term. We perform many simulations, and from these we can calculate the desired probabilities (ex: McIlroy’s win probability is defined as the fraction of simulations where he was the winner).
The full details on our statistical model for predicting PGA TOUR events can be found here.
Brothers Matt and Will Courchene grew up in a Canadian household full of golf fanatics. In 2016, they launched a DataGolf blog in hopes of contributing fresh and unbiased insights to the sport. Matt, a PhD student at the Vancouver School of Economics, focuses on applied econometrics and causal inference, while Will, who has a Masters of Economics from the University of Toronto, focuses on statistical modeling and data visualization.