Simple way to rank F1 seasons
- This topic has 12 replies, 6 voices, and was last updated 10 years, 6 months ago by pH.
- AuthorPosts
- 18th October 2013, 20:56 at 8:56 pm #133758AnonymousInactive
I’ve been ruminating about a one number to quantify how competitive the past F1 seasons are. Here’s what I have come up with. Dominance % is just Points gap/Total points scored. How does this feel intuitively. Would you say these seasons below were the most competitive? Any qualitative thoughts welcome.
Season Driver Team Engine Tyres Poles Wins Podiums FL Points Races Pt. Margin Domination
1984 Niki Lauda McLaren* TAG M 0 5 9 5 72 16 0.5 1%
2007 Kimi Räikkönen Ferrari* Ferrari B 3 6 12 6 110 17 1 1%
2008 Lewis Hamilton McLaren Mercedes B 7 5 10 1 98 18 1 1%
2012 Sebastian Vettel Red Bull* Renault P 6 5 10 6 281 20 3 1%
1994 Michael Schumacher Benetton Ford G 6 8 10 8 92 16 1 1%
1976 James Hunt McLaren Ford G 8 6 8 2 69 16 1 1%
2010 Sebastian Vettel Red Bull* Renault B 10 5 10 3 256 19 4 2%
1981 Nelson Piquet Brabham Ford G 4 3 7 1 50 15 1 2%
2003 Michael Schumacher Ferrari* Ferrari B 5 6 8 5 93 16 2 2%
1958 Mike Hawthorn Ferrari Ferrari E 4 1 7 5 42 11 1 2%
1964 John Surtees Ferrari* Ferrari D 2 2 6 2 40 10 1 3%
1999 Mika Häkkinen McLaren Mercedes B 11 5 10 6 76 16 2 3%
1986 Alain Prost McLaren TAG G 1 4 11 2 72 16 2 3%
1961 Phil Hill Ferrari* Ferrari D 5 2 6 2 34 8 1 3%
1988 Ayrton Senna McLaren* Honda G 13 8 11 3 90 16 3 3%
1983 Nelson Piquet Brabham BMW M 1 3 8 4 59 15 2 3%
1997 Jacques Villeneuve Williams* Renault G 10 7 8 3 81 17 3 4%
1974 Emerson Fittipaldi McLaren* Ford G 2 3 7 0 55 15 3 5%
1979 Jody Scheckter Ferrari* Ferrari M 1 3 6 0 51 15 4 8%
1990 Ayrton Senna McLaren* Honda G 10 6 11 2 78 16 7 9%
2006 Fernando Alonso Renault* Renault M 6 7 14 5 134 18 13 10%
1967 Denny Hulme Brabham* Repco G 0 2 8 2 51 11 5 10%
1950 Giuseppe Farina Alfa Romeo Alfa Romeo P 2 3 3 3 30 7 3 10%
1956 Juan Manuel Fangio Ferrari Ferrari E 6 3 5 4 30 8 3 10%
1970 Jochen Rindt Lotus* Ford F 3 5 5 1 45 13 5 11%
1982 Keke Rosberg Williams Ford G 1 1 6 0 44 16 5 11%
2009 Jenson Button Brawn* Mercedes B 4 6 9 2 95 17 11 12%
1959 Jack Brabham Cooper* Climax D 1 2 5 1 31 9 4 13%
1998 Mika Häkkinen McLaren* Mercedes B 9 8 11 6 100 16 14 14%
2005 Fernando Alonso Renault* Renault M 6 7 15 2 133 19 21 16%
1987 Nelson Piquet Williams* Honda G 4 3 11 4 73 16 12 16%
2000 Michael Schumacher Ferrari* Ferrari B 9 9 12 2 108 17 19 18%
1953 Alberto Ascari Ferrari Ferrari P 6 5 5 4 34.5 9 6.5 19%
1951 Juan Manuel Fangio Alfa Romeo Alfa Romeo P 4 3 5 5 31 8 6 19%
1980 Alan Jones Williams* Ford G 3 5 10 5 67 14 13 19%
1996 Damon Hill Williams* Renault G 9 8 10 5 97 16 19 20%
1978 Mario Andretti Lotus* Ford G 8 6 7 3 64 16 13 20%
1960 Jack Brabham Cooper* Climax D 3 5 5 3 43 10 9 21%
1989 Alain Prost McLaren* Honda G 2 4 11 5 76 16 16 21%
1973 Jackie Stewart Tyrrell Ford G 3 5 8 1 71 15 16 23%
2004 Michael Schumacher Ferrari* Ferrari B 8 13 15 10 148 18 34 23%
1977 Niki Lauda Ferrari* Ferrari G 2 3 10 3 72 17 17 24%
1968 Graham Hill Lotus* Ford F 2 3 6 0 48 12 12 25%
1991 Ayrton Senna McLaren* Honda G 8 7 12 2 96 16 24 25%
1965 Jim Clark Lotus* Climax D 6 6 6 6 54 10 14 26%
1972 Emerson Fittipaldi Lotus* Ford F 3 5 8 0 61 12 16 26%
1993 Alain Prost Williams* Renault G 13 7 12 6 99 16 26 26%
1985 Alain Prost McLaren* TAG G 2 5 11 5 73 16 20 27%
1962 Graham Hill BRM* BRM D 1 4 6 3 42 9 12 29%
1975 Niki Lauda Ferrari* Ferrari G 9 5 8 2 64.5 14 19.5 30%
2011 Sebastian Vettel Red Bull* Renault P 15 11 17 3 392 19 122 31%
1995 Michael Schumacher Benetton* Renault G 4 9 11 8 102 17 33 32%
1952 Alberto Ascari Ferrari Ferrari F P 5 6 6 6 36 8 12 33%
1966 Jack Brabham Brabham* Repco G 3 4 5 1 42 9 14 33%
1957 Juan Manuel Fangio Maserati Maserati P 4 4 6 2 40 8 15 38%
1963 Jim Clark Lotus* Climax D 7 7 9 6 54 10 21 39%
1954 Juan Manuel Fangio Mercedes Mercedes C 5 6 7 3 42 9 16.86 40%
1955 Juan Manuel Fangio Mercedes Mercedes C 3 4 5 3 40 7 16.5 41%
1969 Jackie Stewart Matra* Ford D 2 6 7 5 63 11 26 41%
2002 Michael Schumacher Ferrari* Ferrari B 7 11 17 7 144 17 67 47%
1971 Jackie Stewart Tyrrell* Ford G 6 6 7 3 62 11 29 47%
2001 Michael Schumacher Ferrari* Ferrari B 11 9 14 3 123 17 58 47%
1992 Nigel Mansell Williams* Renault G 14 9 12 8 108 16 52 48%18th October 2013, 22:11 at 10:11 pm #243061Max JacobsonParticipantInteresting that Sebastian Vettel twice appears in the top 7, despite the notion that he has been super dominant!
My one question would be have you just taken all results on the basis of the points system at the time, or have you converted them all to on system? Winning points of course were a higher percentage above second place previously, and there were various dropped points schemes, which may make them appear more or less dominant respectively to those since the new points system was introduced.
Very insightful however!
19th October 2013, 2:06 at 2:06 am #243062MichaelParticipantThat’s a nice easy way of seeing how competitive the season was.
I think it’s a good way of grouping the competitive seasons together.
In 2010 for instance there were 4 drivers fighting for the championship until late in the game. I don’t know if that makes it better or worse than 2008 when there were 2 drivers or 2007 when there were 3 drivers fighting.
So the number of drivers could be used to finesse it slightly.
@Max_Jacobson, the same tune over and over again:-)
19th October 2013, 4:09 at 4:09 am #243063AnonymousInactiveThese are just straight points system in place at that time, since that would be exactly the conditions faced by those champions at that time. so, no revisionist history.
One way to see if these are any good would be, the most dominant season ever was 1992 with winner scoring 48% more points than second placed driver vs. 2007 where there was less than 1% separating the winner v.s. runner-up.
* indicate that winner’s car also won constructor’s championship that year as well.
Here’s the formatted view of same info
http://farm4.staticflickr.com/3699/10356047433_222178b222_b.jpg19th October 2013, 6:32 at 6:32 am #243064MichaelParticipantThanks for including the WCC info as well.
It also shows how the WCC and the WDC go hand in hand. Is it 45 out of 55 times that the WDC winner also won the WCC since 1958? That’s 80 percent (excluding 2007).
Even a 10% difference means the championship went all the way to the last or 2nd last race in the championship.
19th October 2013, 11:27 at 11:27 am #243065Rick LopezParticipantstrange that 2004 is so low as schumacher dominated so much
19th October 2013, 12:51 at 12:51 pm #243066AnonymousInactive@Michael – yes, that’s correct and also an observation, if your lead driver manages just a 3% gap with the runner up, you most likely end up winning the WCC as well.
Of those 10 instances where we WCC/WDC are split, the margin is 3% or less for 8 of those seasons. Exceptions are 73 & 82.
Shows in reality both those championships are one and the same from a team goal perspective, just a 3% gap.
Season Driver Team Domination
2008 Lewis Hamilton McLaren 1%
1999 Mika Häkkinen McLaren 3%
1994 Michael Schumacher Benetton 1%
1986 Alain Prost McLaren 3%
1983 Nelson Piquet Brabham 3%
1982 Keke Rosberg Williams 11%
1981 Nelson Piquet Brabham 2%
1976 James Hunt McLaren 1%
1973 Jackie Stewart Tyrrell 23%
1958 Mike Hawthorn Ferrari 2%19th October 2013, 13:05 at 1:05 pm #243067AnonymousInactive@Rick – I’d generally agree, but it shows the recency effect + a strong second scoring from the same team + general perception of total domination from MS.
Here’s the same view from a constructor’s perspective where you can see both ’02 & ’04 score as very strong seasons from a team perspective.
A good example from this view, both 2010 & 2011 RBR were equally strong, but from a driver’s perspective 2010 was very close due to second driver vs. 2011.
Season Constructor Dominance %
1988 McLaren 67%
1996 Williams 60%
1984 McLaren 60%
2002 Ferrari 58%
2004 Ferrari 55%
1971 Tyrrell 51%
1993 Williams 50%
1989 McLaren 45%
1980 Williams 45%
1987 Williams 45%
2001 Ferrari 43%
1992 Williams 40%
1981 Williams 36%
1977 Ferrari 35%
1979 Ferrari 34%
1963 Lotus 33%
1978 Lotus 33%
1986 Williams 32%
1967 Brabham 30%
1960 Cooper 29%
1966 Brabham 26%
1969 Matra 26%
1975 Ferrari 26%
2011 Red Bull 24%
1961 Ferrari 22%
2010 Red Bull 22%
1968 Lotus 21%
1959 Cooper 20%
1995 Benetton 18%
1997 Williams 17%
1958 Vanwall 17%
1965 Lotus 17%
1972 Lotus 16%
1998 McLaren 15%
1962 BRM 14%
2012 Red Bull 13%
1994 Williams 13%
2008 Ferrari 12%
1970 Lotus 12%
1983 Ferrari 11%
1974 McLaren 11%
1973 Lotus 11%
1976 Ferrari 11%
2009 Brawn 11%
2000 Ferrari 11%
1991 McLaren 10%
1990 McLaren 9%
1985 McLaren 9%
2003 Ferrari 9%
1982 Ferrari 7%
1964 Ferrari 7%
2005 Renault 5%
1999 Ferrari 3%
2006 Renault 2%
2007 Ferrari 0%19th October 2013, 14:32 at 2:32 pm #243068MichaelParticipantyes, that’s correct and also an observation, if your lead driver manages just a 3% gap with the runner up, you most likely end up winning the WCC as well.
Yes, winning the WDC by a margin 3% or anywhere from 6-9 points using today’s scoring system almost guarantees the team the WCC.
I wish there was a way to do this analysis factoring for the driver being favored in the team – I know it’s very subjective but even a 52%-48% bias is a 10% advantage for one driver.
I remember a BPL footballer being told by his coach the other day that playing with 5% less effort means the player (he) is 50% less effective. I think that applies to any competitive sport – 1%-2% off and you’re out pretty much out of the competition.
19th October 2013, 17:13 at 5:13 pm #243069pHParticipantI’d like to start by saying that your idea for measuring dominance is very interesting and I appreciate the work you put into it. It did make me think. However, I also think that the question of points systems is _very_ legitimate.
no revisionist history
I understand what you mean, replacing Senna with Prost would be a bit revisionist, but in my opinion, keeping the “official” results has even worse problems.
Here’s an example: Imagine two drivers, A getting all first places and B getting all second places. Obviously these two are going to be #1 and #2 in the standings. If this happens in 2002, the winner appears in your chart with domination percentage 40%. A year later the very same situation yields 20%.
Now granted, this was an artificial example, so a good question is what difference we can expect in real results. I do not have machine-workable data for drivers, but I can relatively easily check on teams. I tried two cases, what follows is year, gap using official points, and gap using the 9-6-4-3-2-1 system:
2003 9% 14%
2004 55% 66%
So it seems that the difference caused by changing scoring system can easily be about 5%, even exceed 10%. This begs a question whether the comparison using mixed scoring systems is even meaningful, given that a 5% change influences standings in your chart by quite a bit. I think converting to one points system would make the results much more relevant, despite other problems that this would bring.Incidentally, note that the top 5 most dominant seasons in your chart happen to be seasons where the winner was getting almost the biggest share. On the other hand, 2011 rewarded the winner much less. I took the time to recalculate that season (by hand, I might have made a mistake) and came up with the figure 41% instead of 31%, tying Vettel with Fangio.
19th October 2013, 18:43 at 6:43 pm #243070AnonymousInactiveMay I suggest that the winning margin in seconds for each race be included in the system. For example:
Season 1950: (2.6+110+0.4+14+25.7+78.6)/6= 38.55 seconds
Season 1988: (9.8+2.3+20.5+7.1+5.9+38.7+31.7+23.3+13.6+0.5+30.4+0.5+9.6+26.2+13.4+36.4)/16= 16.87 seconds
Season 2011:(22.3+3.3+5.2+8.8+0.6+1.1+2.7+10.9+16.5+4.0+3.6+3.7+9.6+1.7+1.2+12+8.4+8.5+17)/19= 7.4 seconds
Season 2013 till Japanese GP: (12.5+4.3+10.2+9.1+9.3+3.9+14.4+0.8+1.0+10.9+16.9+5.5+32.6+4.2+7.1)/15= 9.5 seconds19th October 2013, 18:45 at 6:45 pm #243071AnonymousInactive@pH – Thanks for your thoughts and I do agree with you in principle.
I will look into how best to do this. Some background that starts to illustrate how confounding it begins to be: http://f1numbers.wordpress.com/2010/02/18/how-fair-have-the-f1-points-systems-been/
I could do a straight 2010 points system and superimpose it on 1950 onwards, but it will just be a theoretical exercise, since those conditions never existed except on this paper. Adding to the effect, that only x races counted during those times, how to incorporate those is also a question.
IMHO, the most purest way (albeit with its own deficiencies) is to just accept that in 1950, Farina scored 30 and Fangio 27, in those conditions that existed, knowing that going into the 1950 season, the points system is such & such, and that they all raced that year in the same condition and that knowledge a win got them 8 points and a second place 6 points (thus, 2/8 = 25% more) allowing the drivers to take just the right amount of risk that the points yielded to win a championship. Contrast with 1991, where there’s a 40% premium for a win (thus more risk a driver can take). Or the current 2012, where the win premium is 28%. F1 has always been about managing risk vs. reward and thus inherently, the driver who own in 1950 had to optimize just as same as the driver in 2013 with those elasticities.
That said, your post convinced me to just recalculate for fun and see how the number line up.
19th October 2013, 21:15 at 9:15 pm #243072pHParticipantThank you for the link, I did some analysis on how much various systems award the winner and top 2, but the guy there took it a few steps further, it was an interesting read.
I agree that point system influences how much risk a driver takes, so even recalculating the results does not make the data “clean”. They are dirty for a different reason, and I feel that recalculated they tell more of the story.
I actually spent some time thinking about it a few years ago when I wanted some convenient way to judge how teams did through the years. This comparison is even more troublesome owing to the fact that – apart from all the factors pertaining to drivers – some teams ran only a few races, some only one driver, some up to six drivers (the 50’s are a nightmare). In the end I decided to use only the best two results from each race for a team (I am aware that this does not solve the participation problem completely, I think it’s not possible to solve it), recalculate all seasons to the 9-6-4-3-2-1 system and then express the achievement as percentage of points earned compared to the total available. This also tells the story of dominance, but from a different angle, I think your and mine PoV’s complement each other very well, it would be great having both using the same scoring system. If you are curious, try
http://math.feld.cvut.cz/habala/misc.htm - AuthorPosts
- You must be logged in to reply to this topic.