# Early-Season Predictions: Don’t Believe Anything You Read — Yet

Don’t believe early-season predictions. There is nothing in a 36-race season that can be predicted on the basis of three races.

Statistics still says that you need a reasonable number of observations before an analysis means anything.

And that’s especially true when we’re talking three races that include a superspeedway, a road course, and the most unique of the mile-and-a-half racetracks.

Many writers don’t write their own headlines, so I’m loathe to call out specific articles. But ignore anyone who claims that NASCAR has (finally) achieved parity. These claims are way premature.

This has been going on forever. I wrote in 2012 that predictions shouldn’t be take seriously until we’ve had 10 races. Many of my examples here come from a 2018 post on the same topic.

## An Example: Cautions

Let’s look at the number of cautions per 100 miles after 1,3,5 and 6 races as an example. This plot was made after predictions that 2018 was set to have the lowest number of cautions ever — after six races.

2018 wasn’t anomalously low after 3 races or 5 races. In fact, 2018 only became the “lowest-caution season” after the sixth race. Let’s look now at the entire 2018 season as a function of race number.

It’s clear from this graph that the number of cautions bounced around until about race #10. It increased from there (from 1.7 per 100 miles to 1.8 per 100 miles), but that increase was gradual.

## Early Predictions are Bad Every Year

There wasn’t anything special about 2018. Here are the corresponding graphs for a few other years.

The cautions don’t always go up, either. Here’s the 2011 data

If made an early-season prediction after race 6, it wouldn’t have been close. Statistics change rapidly over the first 10 races, then settle in as the season goes on.

The same thing happens with driver standings. They fluctuate wildly at the start of the season. Austin Dillon went from 1st to 9th to 11th in three races.

Don’t be swayed just because someone is throwing numbers around. Make sure those numbers mean something. We haven’t achieved ‘parity’.

Not yet, at least.

#### Be the first to comment

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