Predicting the 2023 Formula One Season

Caitlin Ree and Aashna Sibal

Predicting the 2023 Formula One Season

Source: formula1.com

With the help of the Netflix docuseries Drive to Survive, Formula One is quickly rising in popularity and gaining a new audience outside of its typically European viewership. Formula One is the highest-class international open-wheel single-seater car racing league; for Americans, it looks a bit like IndyCar racing. Drivers race at different circuits around the world to compete for both individual Grand Prix wins and the championship title. Formula One teams, called constructors, also compete for a championship title, with two drivers to each constructor.

Drivers (and their respective constructors) get points for scoring in the top ten out of 20 in each race: first place gets 25 points, second gets 18, third gets 15, and so on, until tenth place gets 1 point. An additional point is given to the driver with the fastest lap of the race. At the end of the season, which typically includes 18-23 races, the points are tallied, and the driver with the most points wins the championship; in the case of a tie, the driver with the most wins gets the championship title. Similarly, the constructor whose two drivers scored the most points combined wins the constructor’s championship.

We looked at past data from all the drivers and constructors who are competing in the 2023 season to predict what the standings will be at the end of the season.

Methodology

In order to predict the final 2023 season standings, we used each driver’s past seasons’ results to graph their trajectory. For each driver, we filtered out their standings only at the end of each past season. Then, based on what the graph looked like, we fit either a linear regression or exponential model to the data. After plotting the current data from past seasons as well as the model, we were able to make predictions about the 2023 season’s final results.

The methodology was slightly different for rookies, drivers with only one season under their belts, and drivers who switched teams between the 2022 and 2023 seasons. For drivers who started in Formula One in 2022, we took their standings from 2022 and their team’s projected standing for 2023 (times 2 to account for 20 drivers) and averaged it out (e.g. Zhou Guanyu finished 18th in 2022 and Alfa Romeo is projected to be 6th in the standings, so Zhou’s prediction is now 14th). For rookies, we just took their constructor’s projected standing for 2023 and multiplied it by 2. Finally, for the two drivers who switched teams, we took their predicted place and averaged it out with their new team’s projected standings.

Analysis

By using the trajectories of each driver, we were able to visualize their careers. For example, Fernando Alonso, who has been racing for 19 total seasons, shows a downward trajectory since his prime years. Alonso’s racing prowess is difficult to analyze, though, because he has driven with multiple different constructors with vastly different budgets and quality of cars. Based on our analysis of past results, Alonso is predicted to finish in 15th place overall. After factoring in his switch to Aston Martin, his projection remained the same.

Bottas' trajectory also seems to be going downwards, although that was not always the case. Compared to other drivers, his standings have one of the highest fluctuation rates, as his standings were not consistently rising or declining for longer than three seasons at a time. Since 2019, Bottas’ standing has either stayed the same or declined, with a sharp decline in 2022. Based on our model analysis, we predict Bottas to finish 10th.

Like Bottas, Hulkenberg also had an upward and then downward trend in his standings. 2023 is Hulkenberg’s first season racing since 2020, but because he is racing for the same team as before (Haas), that did not factor into his projection. Our analysis predicts that he will finish in 16th place this year.

As opposed to the previous drivers with a downward trend, Alex Albon is a relatively new driver to already have a sharp decline. However, Albon was dropped from powerhouse Red Bull after 2021 and moved to lower-tier Williams for the 2022 season, which very much impacted his placing. He’s predicted to place last in the 2023 season.

Carlos Sainz, on the other hand, has a steady upward trajectory. From placing 15th in his first Formula One season to 5th in the most recent, he shows promise of remaining a top competitor. Based on our analysis, Sainz is projected to place 4th in the 2023 season.

Right behind Sainz is Sergio Perez, who placed 16th in his first Formula One season in 2011. Since then, he’s improved tremendously, finishing in the top four since 2019 and rising to 3rd in 2022. Our analysis predicts that, due to other drivers improving more rapidly, Perez will finish in 5th place.

Similar to Perez, another competitor that shows a strong upward trajectory is the powerhouse Max Verstappen. Although he started out placing 12th in his first Formula One season, his final season standing improved or stayed the same in every season with the exception of 2017. For 2023, our analysis predicted that he will retain his two-year streak of finishing in 1st.

Charles Leclerc has a similar trajectory: he placed 13th in his first Formula One season and since that has improved each season from the last with the exception of 2020. Based on our analysis, we predict that he will earn 2nd place standing again in 2023.

George Russell also has an upward trajectory, and it’s quite a sharp one due to his moving from bottom-tier Williams to top-three Mercedes in 2022. Our analysis predicts that he will improve one place in 2023, finishing in 3rd place.

Esteban Ocon dropped places from his maiden season to his second, but he has improved every season since then. We predict he will finish in 7th place this year; this would be his best finish ever.

Lando Norris, returning to drive for Mclaren, seemed to be steadily improving since debuting in 2019. He dropped one place from 2021 to 2022, and we predict that he will drop another place to 8th in 2023.

Others have remained somewhat stagnant in their time in Formula One, like current Aston Martin driver Lance Stroll. While Stroll switched teams after his first two seasons, he did not improve or worsen much. From our analysis, Stroll is predicted to place 11th, very similarly to his past results.

Magnussen also has a very similar trajectory, with his place typically fluctuating between 9th, 16th, and in between. 2020 was an exception to this trend, when he finished in 20th place. Our analysis predicts that he will finish in 13th place this season.

Pierre Gasly would be projected to finish second-to-last, but he made the switch from lower-tier Alpha Tauri to Alpine, who are projected to finish 3rd in the constructor’s championship. After averaging out his projection with a 6th place finish (for his constructor times 2), we predict that Gasly will be 12th.

Hamilton’s trajectory is unique, as it seemed to usually fluctuate from 1st to 5th from 2007 to 2020, but since then has been consistently declining. His place also stayed stagnant from 2017 to 2020, with Hamilton finishing in 1st every season. Given his decline in 2022, we predict from our analysis that he will finish in 6th place this season.

The two drivers who are returning for their second full seasons, Yuki Tsunoda and Zhou Guanyu, were projected using their first season’s standings and averaging it out with their constructor’s projection. This resulted in Tsunoda being predicted in 17th and Zhou in 14th. For the three rookies, their predictions based on their constructor’s projections are 9th for Oscar Piastri, 18th for Nyck de Vries, and 19th for Logan Sargeant. Obviously, predicting that a rookie will place 9th overall sounds like a stretch, but it is in line with Mclaren’s projected placement of 5th. Additionally, it may make more sense for Logan Sargeant to be 20th because of his constructor’s projection and rookie status, but Alexander Albon has such a downward trajectory that his prediction is technically lower.

Final driver standings predictions:

1. Max Verstappen

2. Charles Leclerc

3. George Russell

4. Carlos Sainz, Jr.

5. Sergio Perez

6. Lewis Hamilton

7. Esteban Ocon

8. Lando Norris

9. Oscar Piastri

10. Valtteri Bottas

11. Lance Stroll

12. Pierre Gasly

13. Kevin Magnussen

14. Zhou Guanyu

15. Fernando Alonso

16. Nico Hulkenberg

17. Yuki Tsunoda

18. Nyck de Vries

19. Logan Sargeant

20. Alexander Albon

Final constructor standings predictions:

1. Red Bull

2. Ferrari

3. Alpine

4. Mercedes

5. Mclaren

6. Alfa Romeo

7. Aston Martin

8. Haas

9. Alpha Tauri

10. Williams

Model Limitations

As with any model, the models we used to predict the final standings for the 2023 Formula One season come with limitations. For example, we used each racer’s career since their first Formula One season, meaning that the timelines we analyzed differed across racers. We did so to increase accuracy in predicting each racer’s final placement, but a possible change to address this is to compare a subset of the same years for each driver. Of course, unique limitations would come with this model as well.

In addition, the model does not take into account other factors that could influence drivers’ performances during the seasons. Possible examples include injuries that affect driving/practicing, the amount of DNFs (Did Not Finish) that a driver acquires that may impact their final standings, and switching teams from season to season. The quality of a driver’s car, which is based entirely on their constructor, can heavily impact the potential a driver can reach. The prediction is based entirely on trajectories, and some drivers have less obvious or observable trajectories than others.

Another limitation is the applicability of each model (e.g. each model’s p-value). Hulkenberg, Verstappen, Perez, and Sainz had stellar p-values that were very close to zero, so their models fit their data very well. Others, particularly those with a straight line as their trajectory like Magnussen and Stroll, had very high p-values; their data was very hard to fit to any model and, as a result, their trajectories look mostly like the average of their career results. Hamilton’s model, too, although it looks like it would be a good fit, had a very high p-value. Overall, though, most drivers had a p-value less than 0.1, which is decently significant considering the amount of data available.

Conclusion

Because these predictions were made solely with data from past seasons, they do not take into account anything from what we already know about the 2023 season. With pre-season testing done and two races already completed in 2023, we have somewhat of an idea of which teams and drivers have drastically improved or declined. Notably, Aston Martin is off to a shockingly impressive start, with one of the fastest cars on the grid and veteran Fernando Alonso placing 3rd in both races. On the other end of the spectrum, Mclaren, a consistent middle-ground constructor, is having issue after issue with their cars and seems to be trending towards a disastrous season. Despite these surprises, though, we look forward to seeing how our predictions compare to the final 2023 standings.

Source:

http://ergast.com/mrd/