How to measure equipment strength

Throughout racing, certain drivers have been criticized for relying on strong cars to overcome their deficiencies as drivers, with Dale Earnhardt, Jr. and Danica Patrick probably leading the list. While fans can intuitively sense the underachievement of these obvious cases in recent seasons by observing Earnhardt and Patrick’s results versus their teammates, it is far harder to draw comparisons between drivers on different teams, especially if the driver on the superior team has had more success. How much of that is due to the difference in drivers and how much is due to the difference in teams?  While results are measurable, equipment is generally not, so fans traditionally have to speculate on equipment differences based on their interpretation of what they see on the track.

For some time, I’ve been trying to personally come up with a measurement for which drivers and teams have had the strongest equipment, and I think I have had considerable success. While you definitely cannot separate the driver from his/her team as top drivers will tend to have stronger equipment (in part to their ability to relay information about setups to the crew), you can make comparisons by comparing teams’ results on horsepower tracks to more difficult tracks that require driver input. I initially started this analysis for NASCAR because the track distinctions are more obvious: I judged restrictor plate tracks and 1.5-2 mile intermediate ovals as “horsepower tracks” where car differences are more significant than driver differences, which I believe to be a reasonable assumption. I included road courses, short tracks, Phoenix, Dover, Loudon, Darlington, Pocono, and Indianapolis as driver’s tracks which tend to require driver input. Interestingly, the current Sprint Cup schedule is split up exactly 50/50 between driver’s tracks and horsepower tracks with eighteen of each on the schedule since 2001 (although that balance will be broken next year with Kentucky on the schedule and one Loudon date removed).

I have two main measures of performance that I believe I invented and are used frequently on this site: percent beat and average percent led. Both statistics are fairly self-explanatory. The percent beat measures the percentage of cars on the track that a driver beat. For instance, Juan Pablo Montoya, who won the 2000 Indy 500 in his only IndyCar start, he has 100%, while Willy T. Ribbs, who finished last in his only IRL IndyCar start, is at 0%. The driver closest to the middle is Justin Wilson at 50.16%. Percent beat is a measurement of consistency that ignores differences in starting grid sizes, allowing better comparisons between drivers who competed in different eras (for instance, a CART driver who competed against 30-car fields in 1994 and a Champ Car driver who competed against 18-car fields in 2007) or different series (one could theoretically use this to compare an F1 driver from the ’70s, a CART driver from the ’80s, and a Winston Cup driver from the ’90s, but this is not advisable). Another thing I like about the percent beat statistic is that since it is on a 0-100 scale, unlike average finish, you can usually discern whether drivers have had above-average or below-average careers in terms of consistency based on whether they are above or below 50%. Average percent led is an even more self-explanatory method of measuring dominance. The laps led statistic that is frequently used is highly biased towards drivers who lead frequently in races with more laps (thus benefiting for instance superspeedway specialists like Dan Wheldon and Sam Hornish over road course specialists like Will Power and Justin Wilson, simply because superspeedway races have more laps, or in NASCAR, strongly benefiting short track specialists like Rusty Wallace or Denny Hamlin rather than road course specialists like Juan Pablo Montoya or Robby Gordon, simply because short track races have more laps). Average percent led removes that bias towards specific types of tracks; as a result, it is one of the best measurements of career strength. So in order to measure equipment strength in NASCAR, I completely ignored driver’s tracks and simply calculated the percentage beat for each team on horsepower tracks only, which should give the best indication of a car’s abilities independent of the driver (although obviously the driver will still play something of a role). Given that, the teams that had the strongest to weakest equipment were as follows (updated through this weekend’s Kansas race):

Richard Childress Racing – 73.02
Hendrick Motorsports – 66.94
Roush Fenway Racing – 66.76
Joe Gibbs Racing – 62.21
Stewart-Haas Racing – 61.72
Michael Waltrip Racing – 59.79
Earnhardt-Ganassi Racing – 57.33
Richard Petty Motorsports – 56.23
Penske Racing – 50.18
Red Bull Racing – 50.09
Furniture Row Motorsports – 45.42
JTG Daugherty Racing – 42.86
Wood Brothers Racing – 41.27
TRG Motorsports – 33.33
Front Row Motorsports – 32.54
Latitude 43 Motorsports – 26.84
Robby Gordon Motorsports – 26.02
Phoenix Racing – 23.02
Whitney Motorsports – 20.95
Tommy Baldwin Racing – 20.48
Germain Racing – 20.04
Braun Racing – 7.14
NEMCO Motorsports – 6.15
Prism Racing – 5.95
Gunselman Motorsports – 3.17

Just as with drivers, a team above 50% is above-average, while a team below 50% is below average. I conducted a similar analysis for the IndyCar teams on the six horsepower tracks on the schedule (Kansas, Texas, Chicagoland, Kentucky, Motegi, and Homestead) with the following results:

Chip Ganassi Racing – 87.74
Andretti Autosport – 76.01
Panther Racing – 68.51
Penske Racing – 68.39
A.J. Foyt Racing – 55.48
Newman-Haas Racing – 49.36
FAZZT Racing – 43.87
Dreyer & Reinbold Racing – 38.39
Luczo Dragon Racing – 32.79
Conquest Racing – 32.58
Dale Coyne Racing – 31.61
Sarah Fisher Racing – 25.85
KV Racing – 24.09
HVM Racing – 14.19

The Cup team ratings seem generally right although Roush, MWR, Red Bull, and Furniture Row are a bit higher than I would have guessed. The IndyCar ranking also seems decent except for the surprise of Penske in fourth; I guess Power’s inconsistency on these tracks doomed them behind Andretti Autosport (as even Marco and Danica beat Power in the oval points standing), but Panther in third is a bit of a surprise even with Wheldon and Carpenter’s great runs at Chicagoland and Kentucky. I then did this same analysis of Cup teams for years in the past dating to 1993. Some very interesting trends resulted. Anytime Roush had a stronger than usual season, its satellite teams of Yates and the Wood Brothers that it also provided engines to tended to be much higher as well, which does indicate that this is some measure of equipment strength. Several teams that collapsed pretty much declined according to this statistic from year to year with Yates declining from 68.70% (first among all teams) in 2004 in equipment strength to 58.93%, 50.20%, 44.25%, 39.02%, and 33.33% in its remaining five seasons. Bill Davis, Robby Gordon, and Chip Ganassi’s teams (the latter before the Earnhardt-Ganassi merger) had similar year-to-year declines as well that were easily identifiable. Additionally, in seasons when a generally-considered weak driver performed far better than usual, that driver’s team tended to consistently have among the best equipment on the track, the best example being Elliott Sadler in 2004 when driving for Yates (although most would have guessed Hendrick had stronger equipment this season, Yates actually had the strongest equipment of all teams and it is clear that Sadler and the then-aging Dale Jarrett were both underachieving). However, while it is clear that I am capturing something with this measurement of team strength, there were a few reasons I did not ultimately go with this particular measurement. Measuring an entire team and then comparing that team’s drivers to the overall team standards assumes that all teammates within the team are equal, and this has underrated certain teams that have some cars running better than others. I was very surprised when Hendrick Motorsports only came in first in the team rankings in 2007 (admittedly, its best season), but it is hard for me to argue that Childress was the best team in 2008, DEI was the best in 2006, Yates was in 2004, and Gibbs was in 2003 when none of those teams were fighting significantly for the title in those years; additionally some mediocre teams having better-than-usual seasons appeared higher than struggling multi-car powerhouses some seasons, such as Jasper Racing beating Childress in 2002, but I don’t think most would say Jasper was the stronger team that year. Teams like Hendrick and Roush are known to have a car or two lagging behind at all times, so comparing the #48 team to the Hendrick average does not necessarily make sense, since Johnson has definitely had better equipment than his teammates. Given that, when I ultimately conducted the analysis, I compared each driver’s career percent beat and average percent led to that driver’s career percent beat and average percent led on horsepower tracks, rather than comparing to the overall teams, but the analysis of teams nonetheless resulted in some interesting observations.
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: