The Monster Energy NASCAR Cup Series heads to its first road course race of the year at Sonoma Raceway. The 1.99-mile road course is highly technical, and while it is one of two road course tracks that we have data for (the third road course is the new roval at Charlotte), the other road course (Watkins Glen) has faster sweeping corners. Beware of borrowing too much road-course history from Watkins Glen when making your driver evaluations for Sonoma this weekend.
My machine learning model finds that track history at Sonoma is far more valuable than overall road-course prowess. Additionally, my model finds that year-to-date performance and both short- and long-run practice speeds from final practice are statistically significant. As always, I use the RotoViz Driver Sim Scores I've developed to gauge race-winning upside.
One other note: Since 2011, there has been an incident rate of about 20% at Sonoma, so we should expect on average around 7-8 drivers to have major incidents. These incidents create plenty of cautions and also (along with stage cautions) create opportunities for strategy to come into play. As a result, I'll be looking at some longer odds as well as a group bet.
As always, make sure you shop around for the best lines.
Kurt Busch: +1350
Busch simply missed it in qualifying, which is why he'll start 23rd on Sunday. However, starting position isn't a statistically significant factor in my model because of the incidents, cautions and strategy we see at Sonoma. The older Busch brother has an average finish of 5.6 at Sonoma since 2011, tops among all drivers in the field.
The model puts him in the top tier of drivers along with co-favorites Kevin Harvick and Martin Truex Jr. as well as Denny Hamlin and Clint Bowyer. However, I prefer Busch to Hamlin with practice times as the tiebreaker. Busch posted a faster lap in both practice sessions as well as a better 10-lap average.
Brad Keselowski: +2000
Keselowski is one of those drivers who ordinarily would be a trap at Sonoma. He has a stellar history at Watkins Glen, which certainly doesn't hurt, but his performance at Sonoma has been bland. Prior to last year, the highest Keselowski had even finished at Sonoma was 10th (in 2011).
However, last year Keselowski led 15.5% of the race en route to a third-place finish. This reminds me of Hamlin, who struggled for years at the track until a breakthrough second-place finish in 2016 followed by a fourth in 2017.
Like Hamlin, Keselowski could follow up his breakout Sonoma showing with another top-end performance. Per the model, Keselowski projects to finish just behind the top tier of drivers, and in the Sim Scores two of his fifteen most comparable drivers were Sonoma race-winners.
Jimmie Johnson: Top Finisher in Group C +315
This bet isn't to win, but I don't hate Johnson at +2000 at 5dimes if you want to go that route. Instead, I'm picking Johnson to finish tops among four drivers: Keselowski (+260), Jamie McMurray (+260), Ryan Blaney (+285), and Johnson at +315.
Johnson will probably fly under the radar, but his Sonoma résumé is that of a top-10 driver with race-winning upside. Johnson has the best average finish among all drivers at Sonoma since 2009, including a win in 2010. His practice times weren't blistering, but they held up against the field, ranking 13th on an averaged basis. Johnson should hang around and contend — he has completed every Sonoma race since 2005.
My machine learning model puts Johnson in a second tier of drivers that includes Keselowski (as well as Busch, Kyle Larson and A.J. Allmendinger), and a tier above McMurray and Blaney. Johnson also has more upside than any driver in Group C according to the Sim Scores.
