What’s Happened to Long Green-Flag Runs?

A caller to Dave Moody‘s SiriusXM Speedway radio show argued that racing was worse because there aren’t any long green-flag runs anymore.

What do the data say?

Measuring Long Green-Flag Runs

I consider data from 1980-2021. Going back further in time is hard because caution data isn’t readily available.

I started by looking at numbers of green-flag laps, but unless you’re comparing tracks with the same length, you’re comparing apples to oranges. So I settled on studying miles of green-flag racing across seasons.

I counted up all the green-flag runs and calculated their lengths in miles. In order to compare seasons with different numbers of races (or of green-flag segments), I calculated the percentage. In other words, what percentage of all green-flag runs were less than 5 miles long? What percentage of green-flag runs was between 5.01 and 10.0 miles?

2021 as an Example

Let’s start by looking at this year, since it’s freshest in our minds. The first thing I did was look at a histogram of all the data, which is shown below, on the left.

A histogram showing the percentage of green-flag laps by length on the left; on the right, a boxplot showing the distribution of green-flag-run lengths for 2021.

Because the numbers drop off after about 100 miles, I used bigger bins for longer runs. The bins on the left of the histogram are each five miles in width, while those to the right are 10 miles in width.

The histogram for 2021 tells us that:

  • Green-flag laps of length five miles or less made up almost 20% of all green-flag runs.
  • Runs between 5.01 and 10.0 miles were about 14%
  • You can go down the line and see that there are many fewer of the longer green-flag runs.

One way of trying to sum all this data up in one or a few numbers would be taking an average — but that’s dangerous anytime you have a distribution that has a long tail. The small number of really long runs skews the average.

Average vs. Median

The right plot — a boxplot — is a bar whose height represents the middle 50% of all green-flag runs. In 2021, 50% of all green-flag runs were between 7 miles and 58 miles.

The lines show the minimum and maximum values of the distribution.

And then there are the outliers. I made them red diamonds so you’d notice them. These are values that are so far outside the ‘normal’ that they are flagged. In statistics, these values are often discarded, but we know they’re real and we can’t do that.

We had 184 green-flag runs in 2021 so far. Anything longer than about 140 miles is an anomaly. We have nine such anomalies this year, which is about 5% of the data. This is important because it tells you that really long green-flag runs (at last this year) are very rare.

The red bar running across the box is the median — the value for which half of all green-flag laps are longer and half are shorter.

  • The median for 2021 (through 24 races) is 24 miles. That means half of all green-flag runs were shorter than 24 miles and half were longer.
  • The fact that the median is not at the halfway point of the bar tells us that there are many more short runs than long runs. That’s confirmed by the histogram.

Sometimes, looking at the data from two different angles really helps.

In Search of Long Green-Flag Runs Over The Year

Let’s look at plots for earlier years. Here’s 2012, 2002, 1992 and 1982.

A histogram showing the percentage of green-flag laps by length on the left; on the right, a boxplot showing the distribution of green-flag-run lengths for 2012.
A histogram showing the percentage of green-flag laps by length on the left; on the right, a boxplot showing the distribution of green-flag-run lengths for 2002
A histogram showing the percentage of green-flag laps by length on the left; on the right, a boxplot showing the distribution of green-flag-run lengths for 1992.
A histogram showing the percentage of green-flag laps by length on the left; on the right, a boxplot showing the distribution of green-flag-run lengths for 1982

You can spot a few similarities right off:

  • All years have outliers — although some outliers lie much further out than others. Put another way, we have long green-flag runs in all four of the years shown.
  • With the exception of the 1982 data, the largest percentage of green-flag runs are five miles or less.
  • But even for 1982, the largest percentage of green-flag runs is 5-10 miles.

Pulling it all Together

What we really want is a way to show all this data on one graph. So here goes.

The graph below shows the median value for each year as a green dot. The grey lines bracket the green bar on the boxplots from each individual chart. In other words, 50% of all green-flag runs for each year were in between the lengths shown by the grey bars.

A scatter plot showing the median and the middle 50% of all green-flag runs from 2980-2021.

There is obviously a bit of scatter, but it you look at just the green dots, it’s hard to argue that the median green-flag lap length exhibits any significant downward trend over the forty years of data shown. The lower limit is bounded because we can’t have negative miles, of course. But there aren’t any discernible trends for the upper ends of the grey bars, either.

Another Way of Looking at Long Green-Flag Runs

The last graph put all the boxplot data into one graph. We can put all the histograms into one graph, too. This graph has the advantage of having all the data in one place — and the disadvantage of being pretty darn busy. But the blog has been awfully monochrome to this point.

So here’s a stackgraph.

I know: It look like Charlie Brown’s sweater redesigned by Timothy Leary. But I promise there’s meaning here.

A stackplot showing the percentage of total green-flag runs by length for 1980-2021.

The bottom-most segment (in Christmas-tree green), shows the percentage of green-flag runs that were between 0 miles and 5.0 miles. There is definitely a very slight uptick over the years, but the percentage remains mostly between 15% and 20%. Some of the increase is due to the introduction of green-white-checker finishes. At all but the largest tracks, GWC runs are less than 5.0 miles long. I can’t argue that GWC finishes makes the racing less interesting, so I don’t see that uptick as a negative.

The red segment is the percentage of laps between 5.01 miles and 10.0 miles. Agains, it’s a little larger over the years — but we have to note that the rise in the last two years is probably because of the plethora of competition cautions instituted because of COVID

The blue segment immediately above the red represented green-flag runs of 10.01 miles to 15.0 miles, and so on. I grouped them in bigger segments as the numbers got higher because it was hard to see anything otherwise.

With the exception of a few years, this graph tells us that almost 50% of the green-flag runs every year in our sample were 30 miles or under. And 75% of the green-flag runs have been 75 miles and under.

So there haven’t been any meaningful changes in the vast majority of green-flag runs.

But It Sure Seems There Are Fewer Long Green-Flag Runs…

And whenever you think that, you want to think about the human tendency to remember outliers. There definitely has been a decrease in some green-flag runs, but you can’t really see it in the graphs above because we’re talking about a very small number of of runs.

I made a scatter plot, but instead of considering the median lap, I used the most outlying of outliers for each year.

A scatterplot of the longest green-flag runs each year from 1980 to 2021. This shows that the length of the longest runs has gone consistently down over time.

Aha! We finally found an obvious decrease over the years. The longest green-flag runs aren’t as long as they used to be.

The trend didn’t start with the introduction of stages, but you can see a distinct change in 2017. It is impossible for us to have a stage longer than 200 miles, so it’s impossible to have a long green-flag run longer than 200 miles anymore.

Let’s Look at a Really Long Green-Flag Run

Let’s examine the longest green-flag run in a little more detail. Back in 1981, a Dover race went green from Lap 30 to Lap 500. (Caveat: there are a number of races in the 1980s without detailed caution data. They were excluded for my analysis.)

The Dover race featured twelve lead changes (near the average for Dover races); however, most lead changes appear to be during green-flag pit stops. Neil Bonnet would lead 25+ laps, then someone else would lead one lap, and then Neil would be back for another long run. Bonnet led 264 consecutive laps at one point, from Lap 196 to Lap 459! He clearly had the car to beat.

Bonnet ended his lead at lap 265 when his engine blew. Cale Yarborough took the lead for 21 laps, and then his engine blew.

Only two cars (winner Jody Ridley, who led the last 20 laps and P2 Bobby Allison) finished on the lead lap. P3 was one lap down and P4 was nine laps down.

Nineteen of the 32 cars didn’t finish the race — that’s almost 60% of the field DNFed. Eight engines gave up during the race.

Many of the long-green-flag runs I examined have a similar pattern. It doesn’t sound like great racing to me.

To be fair, there are a few examples that might have been pretty good. In particular, there was a race in 1993 at Darlington with a 409-mile green-flag run that had a lot of what look like competitive green-flag racing. Still, only four cars ended up on the lead lap.

Conclusion

Yes, it’s now impossible to have a green-flag run longer than 200 miles because of stages. But we’ve never had a whole lot of 200 mile+ green flag runs in the forty years we’re considered here. When there is a long green-flag run with a lot of exciting racing, you’re much more likely to remember it than other green-flag runs.

Related Posts

Be the first to comment

Leave a Reply

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