by Jacob Whittle
Who is the greatest Formula 1 driver? The ultimate question of any sport. The endless problem of answering that question is that, like so many sports, Formula 1 has changed dramatically over the years. From when the World Championship began in 1950, racing cars bear very little resemblance to one another throughout each decade. Would Senna or Fangio have been as successful in the modern era? Could we have dropped Alonso or Hamilton into a 1960s F1 car and seen them win titles? Is it possible to compare drivers of completely different eras of racing? We shall investigate that possibility.
Who is the greatest Formula 1 driver? The ultimate question of any sport. The endless problem of answering that question is that, like so many sports, Formula 1 has changed dramatically over the years. From when the World Championship began in 1950, racing cars bear very little resemblance to one another throughout each decade. Would Senna or Fangio have been as successful in the modern era? Could we have dropped Alonso or Hamilton into a 1960s F1 car and seen them win titles? Is it possible to compare drivers of completely different eras of racing? We shall investigate that possibility.
One area of variation throughout the history of Formula 1 is the format of scoring points, used to decide the results of the F1 World Championship. For this investigation we will restrict ourselves to the three points scoring formats used from 1991 to the present day [footnote 1]. These formats are detailed in the table below:
Finishing position 
19912002 
20032009 
2010present 
1st 
10 
10 
25 
2nd 
6 
8 
18 
3rd 
4 
6 
15 
4th 
3 
5 
12 
5th 
2 
4 
10 
6th 
1 
3 
8 
7th 
2 
6 

8th 
1 
4 

9th 
2 

10th 
1 

Total 
26 
39 
101 
Note that for each system, the total number of points available in a race is different. It could be said that each pointspaying position can be expressed as a percentage of the total points available. If we express our values in this form, this allows us to compare each of the formats.
So as an example, winning a race between 1991 and 2002 scores 10 out of a possible 26 points, or 38% of the possible total points available in a race. Applying this logic to each of the points scoring format gives the following table on the true value of each pointspaying position:
So as an example, winning a race between 1991 and 2002 scores 10 out of a possible 26 points, or 38% of the possible total points available in a race. Applying this logic to each of the points scoring format gives the following table on the true value of each pointspaying position:
Finishing position 
19912002 
20032009 
2010present 
1st 
38% 
26% 
25% 
2nd 
23% 
21% 
18% 
3rd 
15% 
15% 
15% 
4th 
12% 
13% 
12% 
5th 
8% 
10% 
10% 
6th 
4% 
8% 
8% 
7th 
5% 
6% 

8th 
3% 
4% 

9th 
2% 

10th 
1% 

Total 
100% 
100% 
100% 
This now allows us to compare between all three points scoring formats. Straight away there are some interesting observations, notably how Format 1 is very much skewed in favour of the victory, while Format 2 has the smallest margin between first and second which makes second a very acceptable finishing position in that era. Format 3 remains unchanged, since this points total was out of almost one hundred already.
Since we are now able to compare all three formats as they are all now measured as a percentage, we can now create a model that compares championship results from each of these different eras. For this the seasons of 1999, 2008 and 2010 have been selected as our case study. These selections are arguably the closest of each of their respective eras, so these seasons are most likely to deliver interesting data to compare.
In order to find each drivers’ true points score, we will add the total points they would have scored as a percentage value. We will then take the average score over the total number of races in the season, and compare our results to find the best driver out of those three seasons [footnote 2]. For this initial analysis, we will only compare the top 6 drivers in each of the three championships. This gives us a wide spectrum of data to work with, while not overstretching our initial use of this model.
The full tables of data are given in the Appendix, but the final outcomes for each of the drivers are given in the tables below (use the tabs to switch between years):
Since we are now able to compare all three formats as they are all now measured as a percentage, we can now create a model that compares championship results from each of these different eras. For this the seasons of 1999, 2008 and 2010 have been selected as our case study. These selections are arguably the closest of each of their respective eras, so these seasons are most likely to deliver interesting data to compare.
In order to find each drivers’ true points score, we will add the total points they would have scored as a percentage value. We will then take the average score over the total number of races in the season, and compare our results to find the best driver out of those three seasons [footnote 2]. For this initial analysis, we will only compare the top 6 drivers in each of the three championships. This gives us a wide spectrum of data to work with, while not overstretching our initial use of this model.
The full tables of data are given in the Appendix, but the final outcomes for each of the drivers are given in the tables below (use the tabs to switch between years):
Position 
Driver 
Total percentage points 
Average points per race 
1st 
Hakkinen 
292.31 
18.27 
2nd 
Irvine 
284.62 
17.19 
3rd 
Frentzen 
207.69 
12.98 
4th 
Coulthard 
184.62 
11.54 
5th 
Schumacher, M 
169.23 
16.92 
6th 
Schumacher, R 
134.62 
8.41 
Position 
Driver 
Total percentage points 
Average points per race 
1st 
Hamilton 
251.28 
13.96 
2nd 
Massa 
248.72 
13.82 
3rd 
Raikkonen 
192.31 
10.68 
4th 
Kubica 
192.31 
10.68 
5th 
Alonso 
156.41 
8.69 
6th 
Heidfeld 
153.85 
8.55 
Position 
Driver 
Total percentage points 
Average points per race 
1st 
Vettel 
253.47 
13.34 
2nd 
Alonso 
249.50 
13.13 
3rd 
Webber 
239.60 
12.61 
4th 
Hamilton 
237.62 
12.51 
5th 
Button 
211.88 
11.15 
6th 
Massa 
142.57 
7.50 
The 1999 season is a good example of the benefits of taking an average points score, rather than the total points accrued. Michael Schumacher broke his legs in an accident that season and missed half a dozen races. As a result he finished 5th in the championship. However, by taking his average result over the 10 races he competed at rather than the full 16 of the season that year, our data suggested Schumacher was a better driver than Frentzen or Coulthard that year. And history would certainly seem to agree with that theory.
Combining all our data together to find the best driver out of these three years provides the following result:
Combining all our data together to find the best driver out of these three years provides the following result:
Pos 
Driver 
Average points 
Pos 
Driver 
Average points 
1st 
Hakkinen ('99) 
18.27 
10th 
Hamilton ('10) 
12.51 
2nd 
Irvine ('99) 
17.79 
11th 
Coulthard ('99) 
11.54 
3rd 
Schumacher, M ('99) 
16.92 
12th 
Button ('10) 
11.15 
4th 
Hamilton ('08) 
13.96 
13th 
Raikkonen ('08) 
10.68 
5th 
Massa ('08) 
13.82 
14th 
Kubica ('08) 
10.68 
6th 
Vettel ('10) 
13.34 
15th 
Alonso ('08) 
8.69 
7th 
Alonso ('10) 
13.13 
16th 
Heidfeld ('08) 
8.55 
8th 
Frentzen ('99) 
12.98 
17th 
Schumacher, R ('99) 
8.41 
9th 
Webber ('10) 
12.61 
18th 
Massa ('10) 
7.50 
There are many points of interest to expand on from this result. Firstly, the data suggests that in 1999 Irvine and Schumacher were both better drivers than Hamilton and Vettel, both in their respective World Championship winning years. The data also suggests than Hamilton and Massa both had far superior seasons in 2008 than they did in 2010. These are areas of future analysis that could be expanded upon.
But returning to the original hypothesis, we have been able to compare drivers of different racing eras and we can conclude that the most successful driver out of the three years of F1 racing analysed in our model was the eventual 1999 World Champion Mika Hakkinen [footnote 3]. We have taken our first step towards a model that could provide us with mathematical evidence of the greatest driver in Formula 1 history.
Appendix: full tables of data calculations
But returning to the original hypothesis, we have been able to compare drivers of different racing eras and we can conclude that the most successful driver out of the three years of F1 racing analysed in our model was the eventual 1999 World Champion Mika Hakkinen [footnote 3]. We have taken our first step towards a model that could provide us with mathematical evidence of the greatest driver in Formula 1 history.
Appendix: full tables of data calculations
Footnotes
[1] The points systems used pre1991 did not include all results. Instead, a designated number of the best results were the only results that counted towards the championship. In addition the period in the season when these results were picked were varied. For example, in the 1967 season championship points were scored with the top 5 result from the first 6 races, then the top 4 results from the last 5. This points system has affected the outcomes of World Championship, and in the early stages of this model this is a scenario worth exploring separately.
[2] As of yet, this model does not factor in halfpoint races. This is when less than 75% of the race distance has been completed, usually due to adverse weather conditions or a major incident, and thus half the standard points are awarded. Fortunately this is a rare occurrence, and did not occur in the three seasons in our case study, so this did not affect our model and can be accounted for at a later point.
[3] This model does not differentiate between the success of a racing team and the success of the individual. So, the completely accurate conclusion is that the best driver in 1999 was Mika Hakkinen driving for the McLaren team. However, it takes a combination of both driver and team to compete for a World Championship, so this fact should not invalidate any deductions from the data. The implications of a driver’s average points score when driving for a different team is certainly an area worth exploring in future, as having a superior car to everyone else (such as Schumacher in the early 2000s or Hamilton in the 201415 Mercedes) or indeed an inferior car has a significant impact on the success of the driver.
[1] The points systems used pre1991 did not include all results. Instead, a designated number of the best results were the only results that counted towards the championship. In addition the period in the season when these results were picked were varied. For example, in the 1967 season championship points were scored with the top 5 result from the first 6 races, then the top 4 results from the last 5. This points system has affected the outcomes of World Championship, and in the early stages of this model this is a scenario worth exploring separately.
[2] As of yet, this model does not factor in halfpoint races. This is when less than 75% of the race distance has been completed, usually due to adverse weather conditions or a major incident, and thus half the standard points are awarded. Fortunately this is a rare occurrence, and did not occur in the three seasons in our case study, so this did not affect our model and can be accounted for at a later point.
[3] This model does not differentiate between the success of a racing team and the success of the individual. So, the completely accurate conclusion is that the best driver in 1999 was Mika Hakkinen driving for the McLaren team. However, it takes a combination of both driver and team to compete for a World Championship, so this fact should not invalidate any deductions from the data. The implications of a driver’s average points score when driving for a different team is certainly an area worth exploring in future, as having a superior car to everyone else (such as Schumacher in the early 2000s or Hamilton in the 201415 Mercedes) or indeed an inferior car has a significant impact on the success of the driver.