Retirements, Replacements and Rookies


He didn’t use the word ‘retirement’, but the sad news last week was that Matt Kenseth won’t have a Cup ride in 2018. It’s partly his choice. He could have done a Kasey Kahne and taken a ride with a lower-level team, but Matt said he wouldn’t race if he couldn’t be competitive. Kenseth will be replaced in the 20 car by 21-year-old Erik Jones.

Dave Moody summed up the situation on his blog:

Many observers assumed that the 38-time Cup Series winner would have no trouble finding a new ride for 2018 and beyond, but a changing economic climate and an influx of young driving talent appears set to relegate Kenseth to the sidelines, at least for now.

Kenseth said he realized the likely outcome when Kahne’s old ride was filled with 19-year old William Byron.

On Your Own Terms

One of Kenseth’s frustrations was not being to go out on his own terms. He didn’t get to do the farewell tour like Rusty Wallace, Jeff Gordon or Mark Martin.

So what does a career where you go out on your own terms look like?

I’m looking here at three measures of success: Wins, Top 5s and Top 10s. I used to regard Top 5s and Top 10s with a grain of salt, but the introduction of stages has given renewed meaning to those metrics.

When people discuss performance, they usually use last year’s statistics. If you’re a car owner, you’re trying to project future performance. You need more than just what happened last year. So I’m plotting not wins or placements, but the cumulative stats: The driver’s lifetime performance to date.

Let’s look at Dale Jarrett as an example. Wins are triangles, Top 5s are diamonds and Top 10s are squares.

Dale Jarrett Cumulative Stats

This is an example of what I’ll call an “S-shaped Curve”. Just so you don’t think I picked one example that fit my point, let’s look at four classic drivers and I’m showing only wins here to keep the graph clean.

The Model

You see some similarities, right?  As a first shot at a model, let’s divide the S-shaped curve into three parts.

The three parts of the model are:

  • Gearing Up: When the driver is just getting started: The first win, maybe awhile before the second and third.
  • Full Throttle: The driver wins consistently
  • Easing Off: The wins come less frequently. When the driver stops winning, the curve becomes completely flat

The Next-Generation Drivers

There are some subtleties, especially when you look at the careers of more current drivers. For example, here’s Jeff Gordon’s wins:

Jeff Gordon's Cumulative Wins vs. Age

You see the same general trend with Gordon, but some of the details are a little different. You’ll find a lot of similarities between Gordon’s graph and many of the drivers of his era.

The Gearing Up Period Has Gotten Shorter. We talk a lot today about drivers starting much younger, but the shorter gear-up period started way before the current youth movement. What is now the XFINITY Series started in 1982. Drivers do their gearing up there (and in other series) because they won’t have a couple years to prove themselves once they get in Cup.

Drivers Have Multiple Full-Throttle Periods. Back in the day, the driver was the main variable. As race cars became more complicated, the people and technology surrounding the driver have become just as important. You can see this in Gordon’s win graph.

Just in case you can’t, let me show you a little better. See what happens if we divide up the graph at the points where Gordon changed crew chiefs?

Jeff Gordon Wins vs. Age by Crew Chief

The Slope

The lines I’ve drawn tell you the average number of wins per year. You learned this is school. Remember y = mx+b where m was the slope and b was the intercept?

The slope is nothing more than how steep the line is. It’s the rise over the run.

Rise and Run Related to Slope

For this analysis, the steeper the slope the better. For Gordon, the slope of the line from age 22 to 27 (the Evernham years) is 9.82 wins per year. You can calculate them out for each crew chief. (Excel actually has a button you can press to get a best-fit line.)

Crew Chief Slope (wins per year)
Evernham 9.82
Loomis 4.00
Letarte 1.54
Gustafson 2.10

Gordon’s progress is typical of most (but not all) drivers. They start off strong, but as time goes on, they don’t win as frequently. In the old days, drivers didn’t make the kind of money they do today. Retirement wasn’t an option the way it was today.

Slope gives you a really good idea how the driver is likely to perform (barring injury or personal issues). Let’s look at another driver whose current owner opted not to renew his contract.

Kasey Kahne Wins

Like Gordon, you can break Kahne’s record into discrete segments. They mostly track with owner, but remember that there was a mess with Evernham becoming GEM, then merging with Richard Petty Motorsports and then Kahne running with two different teams one year before Red Bull…

In the first six years of is career, Kahne’s slope was 2.26. In the last six years, it’s less than one. (You go two years without winning and that’ll happen.) This is the curve of a driver closing in on the plateau.  

But Rookies Have No Slope!

Is it a risk to replace a known quantity — even if it’s not a big quantity — with a total unknown?

Yes. Definitely.

But thanks to XFINITY and trucks, there is no such thing as an unknown anymore. Erik Jones doesn’t have a long enough record in Cup to make comparing his numbers meaningful, BUT the slope for his career wins in both Trucks and XFINITY is 3. Kahne never won three races in an XFINITY season.

What About Matt?

Kenseth is often cited for his consistency and you can see it here.

Matt’s first Full Throttle period was with Roush and a slope of 1.7. Those last five points are with JGR and the slope is 1.9. He won seven races his first year with JGR, but there wasn’t much overall change in his rate of winning after that. This is the advantage of looking at the record this way: you can see trends. Replacing Kahne with Kenseth would make sense, but Jones’ record in XFINITY and Trucks is better than either of those.


Winning Isn’t Everything…. Or Is It?

Here’s something interesting. If you look back at Dale Jarrett’s record, you’ll notice that the wins plateau before the Top 5s. The Top 5s plateau before the Top 10s.

Kenseth’s performance in terms of finishes (excluding wins) is pretty close now to what it was fifteen years ago. There’s been no falloff in those stats. But compare them to some other top drivers

Driver Top 5 Top10
Kenseth 10.6 18.7
Johnson 15.3 22.5
Keselowski 13.0 21.0


Moody put his finger on two issues in his blog:

Economics: Veteran drivers have larger salaries, larger performance bonuses, larger cuts from merchandise. There are fewer season-long sponsors and let’s also mention that there are fewer teams, including fewer lower-level teams that once offered younger drivers a chance to get that first bit of experience and prove themselves.

A glut of young, talented drivers: I’ll add to this by noting that holding onto a veteran driver doesn’t only mean you miss out on a younger driver. It means one of your competitors gets that driver. Let’s say JGR was willing to give Kenseth a one-year contract. There’s no guarantee one of those young, talented drivers will be available next year. There’s a reason Penske went to three cars: they wanted Ryan Blaney under their roof.

Let me offer a third consideration to add to the above two points.

Winning Really is More Important Now. This is what NASCAR wanted with all the changes to the Chase in the last few years and I think they’ve achieved it. Owners need drivers who can win three, four, five or more races a year. When your average rate of winning is two or under, I think you’re in danger. Wins (and stage points) put you in the final four.

I am sorry Matt Kenseth won’t be driving next year. I will miss him. But if I were a NASCAR owner, based on the analysis I’ve shown you, I think I would’ve made the same decisions they made.


Congratulations. You’re doing a little calculus here. The slope is the common name for the first derivative, which is the rate at which something changes. Speed is the first derivative of position: it’s how fast you change position. Acceleration is the second derivative of position. It’s the rate at which you change your speed.

If we really wanted to be clever with this, we would look at the second derivative of the cumulative wins. The first derivative gives you wins per year. The second derivative tells you how fast the first derivative is changing. When you aren’t winning anymore (when the graphs up there become flat lines), the second derivative is zero. You could theoretically extrapolate from there how long before a driver won’t win anymore.

Parting Shots

A few random interesting tidbits that came up during the analysis for the main article.

Two Exceptions to the Top 5/Top 10 Consistency Trend

Most drivers are pretty consistent throughout their careers (with the exceptions of those drivers with very long careers, like Petty and Waltrip). The graphs are mostly linear for the whole career.

In the very few drivers whose rates of Top 5s and Top 10s did change, it almost always gets worse. Here are the Top 5s for Kasey Kahne.

So if you’re analyzing data to figure out whose career is nearing its end, a falloff in rates of Top5s is another sign that the end is near. Kahne’s rate of Top 10 finishes, however, is pretty constant over his entire career.

Of the very few drivers whose rates of Top 5s changed, Joey Logano was the only one I found who got better as he got older and whose rate of top 10s changed, too. I was curious about Logano because he was so heavily hyped and then chalked up to be a disappointment.

The crossover happened when he was 23. Was everyone right that he was rushed into Cup before he was ready? Could be, but it’s worth mentioning that the crossover happened when he moved from Gibbs to Penske.

Jimmie Johnson & 3 More Years

I heard that Jimmie Johnson has said he plans on driving for three more years, so in light of this analysis, I thought it was worth looking to see whether he was showing any signs of slowing down. Where on the S-shaped curve is he?


That is pretty darn consistent. Sure, there’s a bump between ages 31 and 34 (he won the Championship each of those years), but within the limits of the eyeball, this is astoundingly consistent.

If he drives for three more years and keeps up this remarkable level of consistency, he’s in line to retire with 100 wins. This assumes, of course, that Chad Knaus stays with him for those three years.

Watch Out for Harvick

With Edwards, Stewart, Gordon, Kenseth, and Earnhardt, Jr. retiring (and Johnson talking about three more years and out), Kevin Harvick is now one of the older drivers in the garage. Is he starting to show signs of wear after all this time?

Umm…. No.

Almost all the graphs I looked had the highest slopes earliest in the career. Kevin Harvick bucks that trend.

Sure, he won a championship his first year with Stewart-Haas (Age 38), but the change in his driving record started five years earlier. So when people tell you that drivers peak when they’re young… here’s your counter example. He’s getting better.

Please help me publish my next book!

The Physics of NASCAR is 15 years old. One component in getting a book deal is a healthy subscriber list. I promise not to send more than two emails per month and will never sell your information to anyone.

1 Comment

Leave a Reply

Your email address will not be published.


This site uses Akismet to reduce spam. Learn how your comment data is processed.