StatCar Racing: New Hampshire

June 27, 2008 9:00am CDT

Review of Last Week

Road course chaos was in full force last week as many favored drivers had disappointing finishes and others surprised with much better than expected results.

Correlation between pre-race ranking and finish position was only 23 percent. In particular, the late-race spins that pushed back Tony Stewart, Jamie McMurray, Kevin Harvick, and Ron Fellows hurt our performance.

Matt Kenseth and David Gilliland were some of the biggest surprises, as both recorded top ten finishes despite less than spectacular records on road courses. Kenseth has been clawing his way back into contention over the past several weeks and now sits 12th in points. Expect him to continue his run of solid performance and make the Chase. The real question is whether he can move from recording top ten finishes to winning races. He needs to if he wants to win his second series championship.

Despite this relatively poor performance, my fantasy teams actually gained ground last week. Apparently, others had similar difficulty forecasting who would do well.

This Week

Three variables were good predictors of driver's finish place in last year's Lenox Tools 301: average finish in previous Sprint Cup races at Loudon, number of laps a driver carried the top speed at Martinsville earlier this season, and percent laps run in the top 15 at Phoenix earlier this year.

Like Loudon, Martinsville and Phoenix are short flat raceways, so including metrics from races at these venues is not surprising. Somewhat surprising, however, is that number of fastest laps was the variable with the best predictive power from Martinsville. Potentially, the number of fastest laps is the best variable from Martinsville that shows a driver's ability to get up to speed, navigate traffic, and capitalize on clean air, characteristics that are critical to success when racing at Loudon.

The variables and their impact coefficients are shown below.

Variable, Impact Coefficient
Constant, 23.1
Avg. Finish – Loudon, 0.19
Martinsville - Number of Fastest Laps, -0.08
Phoenix - % Laps in Top 15, -0.16


Computing Dale Earnhardt Jr.'s expected finish is done as follows:

Expected Finish = (1)(23.1) + (.19)(15) + (-.08)(70) + (-.16)(92.9) = 5.4

That expected finish is good enough to put Dale Jr atop this week's rankings. Expect to see his fellow Hendrick teammates Johnson and Gordon in close pursuit.

Rankings