Darlington Raceway hosts the first of the three races (Sunday at 6 p.m. ET on USA Network) that will eliminate four drivers from the Cup Series playoffs.
But which drivers?
That’s a harder question to answer this year than in previous years. Only nine points separate the fifth seed from the 16th seed. In addition, the changes the Next Gen car caused make historical statistics of questionable utility.
So I developed an algorithm to predict the first four playoff eliminations — based only on 2022 numbers.
How to build an algorithm
Predicting the first drivers eliminated from the playoffs is a significantly different task than predicting the championship four. I used the following metrics in my algorithm.
Top-10 finishes: By definition, none of the first four playoff-eliminated drivers will have won a playoff race. Therefore, the algorithm must value running well and scoring points over winning. Playoff drivers have between four (Chase Briscoe) and 17 (Chase Elliott) top-10 finishes.
Average finishing position: A driver with a low average finishing position has few bad finishes or enough good finishes to offset the bad ones. Elliott has the best average finishing position by far, at 10.5. Briscoe and Denny Hamlin are tied for the worst average finishing position at 19.0.
Average running position: Although Hamlin and Briscoe have the same average finishing position, Hamlin has twice as many DNFs. Average running position provides a better measure of drivers’ strength in races they didn’t finish. Elliott has the best average running position (10.2) and Austin Cindric the worst (17.9).
Stage points: The points spread between drivers seeded fifth and up is small, so stage points will make a difference. The caveat is that some crew chiefs choose track position over good stage finishes.
Here’s the ratings from my first iteration of the algorithm.
As always, though, there’s room for improvement.
The Mishap Multiplier
DNFs are high this year, as are accidents and spins. Some drivers have a lot of penalties. I folded these into what I call the mishap multiplier. I didn’t want it to be too big an effect, so I set it to one for the driver with the fewest mishaps. That’s Logano, with three penalties and 10 caution-causing incidents.
I set it to 0.8 for the driver with the most mishaps. Hamlin holds that unfortunate distinction with 17 penalties and 14 accidents/spins.
After scaling the numbers for everyone else, I subtracted 0.01 for each DNF.
Chastain drops from second to fourth, while Hamlin drops from 10th to 14th. But Hamlin’s avoided penalties recently. Chastain’s performance in the second half of the regular season wasn’t as good as in the first. Plus, all the teams know a lot more about the Next Gen car than they did at the season’s start.
I compared metrics from the first half of the regular season to the second half. The results confirmed some things I already suspected, but also held a few surprises.
Bowman’s average finish is 7.7 positions worse in the second half of the regular season relative to the first. His average running position is worse by 5.7 positions. Eight of Bowman’s 10 top-10 finishes came in the first half of the regular season.
Kyle Busch’s average finishing position worsened by 7.3 spots, although his average running position dropped by only 3.1.
Their drops are more than double the next-biggest decrease in average finishing position. Austin Dillon has a 3.3-position drop in average finishing position and a 1.2-position drop in average running position.
On the surprise side, Daniel Suárez tops the list of improved drivers. His average finishing position is better by 4.1 positions and his average running position by 6.1. Suárez had six fewer penalties and five fewer accidents/spins in the second half of the regular season.
Hamlin and Cindric tie for second-most-improved average finishing position at 3.92 positions higher. But Cindric’s average running position is 2.0 positions worse, while Hamlin’s is 3.7 positions better. Hamlin’s team reduced the number of penalties committed in the second half of the season, too.
I expected Kevin Harvick to be on the list of big improvers, but he isn’t. His average finishing position got better by only 0.31 positions in the second half.
Chastain has mixed results. His average running position is better by 1.2 spots, but his average finishing position is worse by 2.2. I bundled all of this data into an “improvement bonus,” which I worked into the algorithm.
Here are the final rankings:
The top eight drivers have small improvement bonuses. Elliott’s is closest to zero because he’s been so consistently good this season.
Again: This algorithm is designed to predict the first four drivers eliminated, not the final order for the season.
And because it’s my first stab at developing an algorithm, there’s lots of room for me to be wrong. For example, I may not have weighted the different elements properly.
There’s a significant gap between the final rankings for the last four drivers and Cindric. It’s possible Cindric and perhaps even Hamlin could exit in round one if any of the drivers below them perform unexpectedly well.
As far as Darlington, Hamlin had a 9.9 average running position in the spring despite a 29th-place finish, He led 42 laps. Bowman, Briscoe, Cindric and Dillon all had average running positions of 18.5 or worse, and led no laps.