Thursday, December 10, 2009

Correlation Matrix of Top 25 Drivers in 2009

Back in July, we took a look at mid-season correlations among drivers.

We found at the time that Kasey Kahne and Jimmie Johnson had the highest correlation among any pair of drivers.

Now that the season is over, we can finish that analysis and take a look at which drivers had the most similar finishes


You can click on the image for a larger version.

That correlation between Kahne and Johnson broke down by the finish, and you see they aren't even one of the highlighted "extreme" cases.

High positive correlation means the drivers often finished well together and poorly together, while high negative correlation means one driver finished well while the other finished poorly, and vice versa. Correlations close to 0 means the two drivers had performances unrelated to each other.

Among the 12 chase drivers, we find the most correlated pairs are
.37 Mark Martin and Ryan Newman
.32 Tony Stewart and Juan Montoya
.31 Jimmie Johnson and Carl Edwards
.30 Carl Edwards and Brian Vickers

The least correlated chase pairs were
-.36 Jeff Gordon and Greg Biffle
-.32 Kurt Busch and Kasey Kahne
-.31 Mark Martin and Brian Vickers

The overall highest correlation was .41 between teammates Clint Bowyer and Kevin Harvick.

Harvick also paired in the other two highest correlations, .40 with both Jeff Gordon and .40 with AJ Allmendinger.

The most negative correlation pairs were
-.48 Marcos Ambrose and Jeff Gordon
-.45 Ambrose and Harvick
-.45 Montoya and Jeff Burton

What's interesting here is to notice how well teammates do with each other. Positive correlations between teammates mean that they tend to perform well together and badly together, a sign that their teams have them dialed in about the same. We already saw that with Harvick and Bowyer, and we notice it as well between Kyle Busch and Denny Hamlin, who score a very high .38 between each other.

The correlations also suggest who has similar driving styles. Notice we see smart, conservative drivers like Martin and Newman together up there, along with aggressive pairs like Montoya/Stewart, Johnson/Edwards, and Edwards/Vickers up top.

Pairs of drivers with opposite driving styles will have negative correlations, so not surprising to see Montoya and Burton with a very negative number, as Montoya is overly aggressive and Burton is a smart and smooth racer. Maybe that's also why you see two very rough drivers in Kyle Busch and Denny Hamlin have a high correlation between each other.

As you check out the grid yourself, you will be able to see your favorite driver paired against all the others, and see how they stack up compared to each other.

email any comments to 36races@gmail.com