One interesting topic I’ve wanted to tackle for some time is forecasting the IndyCar championship. Who is most likely to win the championship at any point in the season? Further, what is the probability of a driver finishing in a specific position in the championship?
In my opinion, there isn’t currently a good way to evaluate team strategy in IndyCar based off of the traditional statistics of starting and finishing position. If a car improves throughout a race, was it because of good strategy, great passes, or just being fast? It’s not particularly easy to tell and even if you watch the race, there’s too many cars and different things going on to get a good measure of what happened by the eye test.
New for 2019, I will be keeping track and updating Expected Points (xPts) for every race and for the season as a whole. xPts is the number of points we would expect to see a driver earn in a race given how he ran as judged by their average track position (ATP) and ATP25. The last 25% of the race is given extra weight as it is when the race is finally coming down to the wire and performance is more crucial. If two drivers both had an ATP of 5 but one had an ATP25 of 3 and then other an ATP25 of 18, while they both had good days from their general ATP, we would still expect the former to score more points than the latter.
Single Seater will be “forecasting” each race of the 2016 IndyCar season, with win probabilities posted to the site after qualifying is over. These will be up either late Saturday or early Sunday morning — if a driver scratches from a race, an additional update will be posted.
- It’s easier to update and keep track of. There aren’t 20 different variables I have to gather together to make a prediction, so I can easily post them for each race.
- There is so much randomness in racing. Crashes, mechanical failures, caution flags — these can all have a drastic affect on the race, and they’re almost impossible to predict. Instead of trying to with a ton of variables, I just skip over it altogether. Also, it helps to distinguish the signal from the noise — data that isn’t very predictive. I’ve chosen variables that are historically very predictive of race success, and that’s it.