Skip navigation
Sign up to follow your favorites on all your devices.
Sign up

Dr. Diandra: What loop data can — and can’t — tell us about passing

Following the conclusion of the NASCAR Cup Series at Bristol Motor Speedway, drivers recap the night race and look ahead to the Round of 12.

Drivers expressed some strong opinions about the Next Gen car’s passing ability after Bristol. Kevin Harvick and Denny Hamlin, among others, felt the car made it too hard to pass.

Brad Keselowski agreed that it was hard to pass, but opined that it’s supposed to be hard.

The passing complaints surprised me. NASCAR’s loop data reported 2,690 green-flag passes for the 2022 fall Bristol race. That’s 980 passes more than the 1,710 green-flag passes recorded for last year’s fall Bristol race.

So are the drivers wrong? Perhaps their comments reflect the accumulated frustration of a long night plagued by so many equipment problems?

Numbers don’t lie. But they also don’t give up their truths easily.

Loop data

Each car carries a transponder that emits a signal unique to that car. Wire loops embedded in the track (and on pit road) record each of these signals. The loops capture a car’s precise position on track — and its position relative to other cars.

The graph below shows green-flag passes by race for the 2022 season. Because races are different lengths (and tracks different sizes), it’s hard to compare data.

A vertical bar chart showing the numbers of green-flag passes as determined by loop data

But superspeedway races stand out for having thousands more green-flag passes than other types of races.

I’ve always been skeptical of passing metrics at superspeedways. Those extraordinarily large numbers just tell us that two or three lanes of cars traded positions a lot. That doesn’t measure passing in a way that illuminates the racing.

What I hadn’t appreciated until I dove into these numbers is that they’re not exactly what you think they are at other types of tracks, either.

According to loop data, Chase Elliott made more green-flag passes than any other driver at Bristol. But did it really take him 154 passes to go from 23rd to second?

Although Elliott’s transponder switched positions with other cars’ transponders 154 times, not all of those events are what I think of as “passing.”

I view passing as capturing a position and holding it for more than a straightaway. But that’s not what loop data is designed to measure.

Old-school data

NASCAR doesn’t make detailed loop data publicly available. What I can access is each driver’s running position each lap. Using that data, I developed a different kind of passing metric.

I’ll use Kyle Larson as an example. Bristol is a good race for this type of analysis because there were no green-flag pit cycles. Counting accurately is confusing enough as it is.

The next graph shows Larson’s running position as a function of lap number. Caution laps are shaded yellow, although you can probably infer cautions from the position changes.

A scatter plot showing Kyle Larson's position as a function of lap number for the fall 2022 Bristol race

I examined each green-flag segment, noting Larson’s positions at the start and end of each segment, and how many times he changed position in-between. The table below shows the results.

A table summarizing the positions gained and lost for each green-flag segment at the 2022 Bristol fall race

Larson started fifth at Bristol and finished fifth. Over the course of 420 green-flag laps, he made 31 passes and was passed 15 times. That produces a pass differential of +16, meaning that he gained 16 more positions than he lost.

If he gained so many positions, how did he end up fifth? He lost 16 positions during pit-stop cycles. That’s not the number of positions he lost on pit road. That’s positions lost including factors like other drivers staying out.

Loop data attributes 109 green-flag passes to Larson. It’s not that one number is wrong and one is right: They’re measuring different things.

More passing or not?

At this stage of the metric, I only feel confident in my results for drivers who finished on the lead lap. Five drivers accomplished that feat at both the 2021 and 2022 fall Bristol races. I compare their passing data in the table below.

A table comparing Dr. Diandra's metric for passing with the loop data for the 2021 and 2022 fall Bristol races

My metric shows passing up by 11.9%, while passing loop data on the same set of drivers shows it up by 55%. If you think of my metric as defining successful passes and loop data as measuring attempts, it shows that drivers had to make more attempts to pass for each successful pass this year than they did last year.

By that measure, the drivers are right that it is harder to pass.

But they are passing.

At this point, it’s impossible to tell whether the limitation is the Next Gen car itself or a level of competition that’s produced 19 different winners this season already.