Strokes Gained T2G Regression: Long or Short
You may have noticed a new tool under the tools tab, PGA Expected T2G Regression. In a similar vein of the SG Putting Regression tool, the T2G Regression tool takes Strokes Gained Tee to Green over the last 2 years, and uses linear regression to predict a player’s SG Tee to Green this year. Then, we compare it to their actual SG, to see who might be over-performing or under- performing. You can find it here: https://dailyfantasywinners.com/pga-expected-tee-green-regression.
As a side note: while making this model, I learned one thing that surprised me. I initially used 2014 SG data as a feature, but I found that it actually did not help the linear regression at all. Essentially, SG T2G two years removed is not predictive of today’s performance (that being said, a more advanced machine learning model might find it useful).
Which also brings me to the flaws in this model and how to interpret it. One thing that this model is blind to is improvements in a player’s game. If player A makes a big swing change that drastically improves his swing, this model is going to say, “Hey, this guy is over-performing!” When in reality, he’s just become a better golfer. That’s why in my analysis below, I’m going to reference age quite a bit. For golfers over 30, or 40, I don’t really buy sudden drastic improvements in their long game. But a player in their 20s? It makes a lot more sense.
Another issue is simply a lack of data, for players who play mostly on the Euro tour, Sergio Garcia and Jon Rahm for example, we don’t have enough previous year data. When we don’t have enough data, I automatically give a player the average strokes gained on the year. So, both these players have huge DIFF’s, but I don’t doubt their skill.
Now, onto the analysis. Borrowing a couple of terms of day-trading (this seems to be popular in the DFS tout industry), here are a few players I’m “Long” on (who I think will perform better in the future) and a few players I’m “Short” on (who I think will perform worse in the future).
Kevin Kisner (SG Actual: 1.392, SG Predict: -.03) – So let me get this straight, a 33 year old with mediocre driving distance and no track record of good performance is suddenly the 10th best Tee to Green player on tour? I don’t buy Kisner at all, and he is one of the players I’m most confident is due for some serious regression.
Keegan Bradley (SG Actual: .557, SG Predict: .89) – While Kisner’s age works against him when looking at sudden improvements in performance, Bradley’s (31) is working for him. He’s consistently been a top 20 hitter on tour but has been a bit mediocre this year. I’m a week late to the party here with Bradley finishing 8th at the Travelers, but he’s still a little too cheap on FanDuel at $7,600.
Marc Leishman (SG Actual: 1.071, SG Predict: .22) – Leishman is also in his 30s and his recent performance has come out of nowhere. He also has been running hot putting according to our Putting regression tool. He’s now priced as an elite player, and he certainly is not.
Jim Furyk (SG Actual: -.041, SG Predict: .72) – Furyk is two years removed from a year where he had the second best long game on tour. Now, he’s below average. He’s likely somewhere in the middle, although it’s possible there’s something up with his body at age 47. His recent two top 30s are encouraging.
Bubba Waston (SG Actual: .365, SG Predict: 1.40) – Bubba legitimately had one of the best long games on tour the past two years. This year? Not so much. I think I get into an argument every other week with my brother Aaron over whether Bubba will return to form or not. There doesn’t seem to be anything wrong with his personal life, and it doesn’t seem like he’s hiding an injury. I’m not about to jump all over him in DFS, but I think he’s always good to have in a few lineups.
Questions or Comments? Follow me @maxjsteinberg on twitter or post on our forums!View all posts by Max J Steinberg