It’s time for the third race of the 2021 F1 season, the Formula 1 Heineken Grande Prémio De Portugal! This racetrack, the Autodromo Internacional do Algarve, is usually referred to as Portimão, which is how I’ll be referring to it in the rest of this post. It’s exciting to be back into the swing of the F1 season after Bahrain, Imola, and the extended break between those (plus the two week break since Imola). I enjoy Portimão on the schedule because the time zone math gives me an extra hour to grab a quick jog and a coffee on Sunday morning.
As my day job has mostly focused on designing and delivering events this year, I haven’t had many opportunities to code. F1 is a data-driven sport, which is part of why I enjoy it, and can be a great resource for data to practice data visualizations of creating machine learning models. Earlier this week, I thought about modeling this race and making predictions about winners. Ultimately, I decided not to for several reasons and I decided it’s important to unpack the reasoning behind why you might not want to model/predict a certain event or outcome. I’ve broken my reasoning down into a few categories:
Lack of Races at Portimão: F1 has raced at Portimão once (2020), so there’s a lack of historical data.
2021 Driver Swaps/New Lineups: Let’s start with the low-hanging fruit: Haas. Two brand new drivers in their lineup this year, both of whom lack previous F1 experience, making this their third F1 race ever. The only real prediction we can make with some certainty is that Mazepin does a Mazespin sometime over the weekend. In terms of other new drivers, Tsunoda at Alpha Tauri has promise to crack the top 10, but he needs to stay consistent.
This week there have been a plethora of videos about how the drivers who have swapped teams this year aren’t adjusting quickly. Vettel is looking dreadful in the Aston Martin, Checo had a pretty horrible race last outing with Red Bull, Ricciardo couldn’t hold his own against Norris and got team ordered in McLaren, Alonso is back and driving for Alpine but hasn’t impressed, and Sainz is just okay in the Ferrari.
Only 12/20 of this year’s drivers raced last year with their same team: Hamilton and Bottas (Mercedes), Norris (McLaren), Leclerc (Ferrari), Giovinazzi and Raikkonen (Alfa Romeo), Latifi and Russell (Williams), Ocon (Alpine (but Renault)), Stroll (Aston Martin (but Racing Point)), Gasly (Alpha Tauri), and Verstappen (Red Bull).
Season/Track Conditions: Last year, Portimão was Round 12 and took place in October. This year, Portimão Is Round 3 and takes place April/May. This may make a huge difference for getting the tires warmed up. As mentioned above, we’ve a plethora of new drivers, returning drivers, and drivers who have swapped teams. 2021 has seen a shortened testing period, a shift in the first race from Australia to Bahrain, and a dry (Bahrain) and wet (Imola) race. Drivers are still adjusting to their cars this year and we haven’t seen two races under the same conditions yet.
Portimão was a late addition to lineup last year and track conditions were questionable. We saw extensive track repairs needed last year, especially after Vettel’s incident with a drain cover in the third practice. Hopefully this year we see improved track quality/conditions.
Low- vs. High-Rake: I won’t subject you to Otmar Szafnauer-levels of complaining about how FIA’s regulation changes this year may be affecting low-rake cars more than high-rake cars, but we have seen Mercedes and Aston Martin struggle a bit at the beginning of this season. Hamilton is still working his hardest to get that Merc where it needs to be, but I think we’ll continue to see this low/high-rake drama continue to play out. With the regulation change and only two races under it, modeling again becomes uncertain for current conditions.
The Bottom Line: A lack of historical data in general, lack of driver data for eight of the drivers in their current teams, drastically different seasons between the only two races (2020/2021), and changes in FIA regulations in the last year all gave me pause in creating a model for this race and doing data-driven predictions based on historical data.
That said, do I think we’ll see many of last year’s top 10 in this year’s top 10? Absolutely. I think ( & hope for the sake of my Fantasy F1) that Norris moves up into the points here, which is possibly based on his last two races (P4, P3). I don’t think Vettel will crack the top 10. Predicted Winner: Verstappen.
TL; DR: Too many caveats, not enough data, no confidence in predictions. Looking forward to an interesting weekend!