Here are my final driver rankings for today’s race at Kansas. I compiled them by crunching the practice and qualifying data from Kansas on Friday and Saturday, and then melding it with the historical data:
- Matt Kenseth
- Kyle Busch
- Kasey Kahne
- Martin Truex Jr.
- Greg Biffle
- Carl Edwards
- Mark Martin
- Clint Bowyer
- Jimmie Johnson
- Brad Keselowski
- Aric Almirola
- Ryan Newman
- Dale Earnhardt Jr.
- Kevin Harvick
- Brian Vickers
- Sam Hornish Jr.
- Tony Stewart
- Paul Menard
- Joey Logano
- Marcos Ambrose
- Kurt Busch
- Jamie McMurray
- Ricky Stenhouse Jr.
- Juan Pablo Montoya
- Jeff Gordon
- Jeff Burton
- Elliott Sadler
- Regan Smith
- Casey Mears
- David Ragan
- Travis Kvapil
- David Gilliland
- Bobby Labonte
- Danica Patrick
- David Reutimann
- Landon Cassill
- David Stremme
- J.J. Yeley
- Dave Blaney
- Michael McDowell
- Timmy Hill
- Mike Bliss
- Josh Wise
Really nice job Jed! Your system seems to do extremely well at capturing the top of the field especially!
How is model performing YTD, excluding Daytona?
Thanks Erik! I admit it’s nice to have three winners in the bag already. But clearly there’s a lotta luck in that. My hot streak could turn cold this week.
And, I think it’s easier to pick the top 5, say, than the fourth 5 (15th-20th). The top cars tend to run up front where there’s less chance they’ll suffer damage. They get better pit stalls. And we tend to get more info/coverage of them during practice/qualifying. Just a theory, though. Maybe later this year I’ll dig into the numbers and check on whether my top 5 picks are actually doing better than my fourth 5.
Bynum, excluding Daytona I from both years, I’m running slightly worse this year so far: 9.79 SD this year vs. 9.36 SD last year.
This year I crushed Phoenix, Vegas and Bristol compared to last year, and was a little better at Martinsville. I started thinking I was the man. But then I got pummeled at Fontana, Texas and Kansas.
Given the fact they’re running a completely new car, I feel OK with the performance so far. I’ll keep tabs on it and post updates periodically.
Not sure how easy it would be to do, but if you try to exclude accidents how are the models performing? Maybe exclude a driver if laps completed is < (total laps * 0.9) or status is Accident?
For example for Fontana you had Hamlin ranked 4th. He very well should have finished 1st or 2nd but with the last lap shenanigans he finished 25th. Same would apply to Mark Martin for that race, ranked 7th finished 37th. I don't think anyone expects a model to account for human behavior or mechanical breakdown.
Bynum, I’ll think on this because it would certainly be nice to know how I’m doing minus crashes. All catastrophes, actually: crashes, mechanical failures, pit road problems, penalties, etc. The trick is figuring out something that’s consistent and comprehensive.