Italian GP 2023

Italian GP 2023

If the weather conditions in Zandvoort cannot stop Max Verstappen from winning, then what can? Going to Monza, Max is set to break another record in case he wins; he will be the first driver who has taken 10 consecutive race wins! Vettel won 9 races in a row in 2013, again with Red Bull. Moreover, Williams performed great in a race they were supposed to be bad at. Monza is definitely a circuit that suits their car, so let’s wait and see how they perform here – and whether Logan Sargeant can get closer to Albon’s performance.

Of course, the model gives Verstappen the highest probability to get the pole and the race win with Perez coming 2nd. Albon is predicted for a rather low position but I think he will do much better than that. Lawson replaces Ricciardo for a 2nd race and it will be interesting to how he handles the whole race week, it might well be his chance for a future full-time race position.

Driver Qualifying Prediction Actual Qualifying Result Race Prediction Actual Race Result
Verstappen 1 2 1 1
Perez 2 5 4 2
Sainz 6 1 2 3
Leclerc 5 3 3 4
Russell 7 4 6 5
Hamilton 3 8 5 6
Albon 14 6 10 7
Norris 4 9 9 8
Alonso 8 10 8 9
Bottas 11 14 11 10
Lawson 20 12 18 11
Piastri 10 7 7 12
Sargeant 18 15 19 13
Zhou 9 16 12 14
Gasly 13 17 13 15
Stroll 12 20 15 16
Hulkenberg 15 13 17 17
Magnussen 17 19 20 18
Ocon 16 18 14 R
Tsunoda 19 11 16 R

5 thoughts on “Italian GP 2023

  1. Hi! I’m curious to understand how the model makes it’s predictions, or what it’s based on?

    Reviewing the predictions, it’s hard to project that Bottas, or even Zhou will come anywhere in the top 10 at Monza?!

    For some reason I feel that Piastri will do better than the model predicts? He’s been very impressive this season, and at another low downforce track (similar to Spa) he could come in the top 5 perhaps?

    I’m interested to see how the Alpine’s fair out this weekend and excited to see Lawson back at the wheel!

    Thanks!

    1. Hi! You’re right. It’s not clear how the model predicts Bottas or Zhou in the top 10. I do agree that Piastri will probably do better than predicted.

      The model uses several features. Most of them are documented in the feature engineering blog posts esp. in part 2. The model makes pairwise comparisons (i.e. for each pair of drivers, it calculates the probability that one will finish ahead of the other – separately for quali and race) and then this is converted to a final ranking. You can actually view these pairwise probabilities via the API.

  2. Will it be possible to add the following factors?
    – Weather condition
    – Team pit strategy
    – Pit stop timing
    – Previous years result
    – Percentage of each driver being at top2/podium?

    Thank you.

    1. Previous year result is already taken into account by the model. The ‘percentage of each driver being at top2/podium?’ is not directly a factor/feature by I am using some statistics (e.g. min/max/average) of each driver’s and each car’s position in the last few races. This takes into account the current form of the driver or car.

      The weather would be a super useful feature but I have not found any source that contains reliable data for each past race. Furthermore, I would need accurate weather prediction for the upcoming race if I were to use it as a factor in the model.

      The other two (team pit strategy and pit stop timing) would not be possible to get into the model because I cannot know beforehand what each team will do in the upcoming race. I can find it for previous races but I cannot use it to predict the future.

Leave a Reply

Your email address will not be published. Required fields are marked *