The ability to shine in wet-weather conditions is highly prized in Formula 1. This is partly because of the increased difficulty for drivers in slippery conditions with poor visibility, and partly because wet weather is considered a leveler of mechanical differences (although there are of course some chassis that are much better suited to wet-weather conditions than others). Legendary drives in the wet are often cited as the greatest driving performances of all time, such as Senna’s drive at Donington 1993, Schumacher’s drive at Catalunya 1996, and that wet qualifying session at Watkins Glen in 1979 when Gilles Villeneuve set the fastest lap by 11 seconds.
After looking around a bit, I didn’t find any detailed analyses of wet-weather driving skills across the history of the sport, so I figured I’d do my own.
In performing any analysis of wet-weather driving skills, it’s first necessary to define what one means by a “wet” race. If it doesn’t rain during the race, but the track is wet at the beginning, is it a wet race? Or what if it only rains in the last two laps? The dryness/wetness of a race could be considered a continuous variable, ranging from totally dry to monsoonal rain. Some drivers may really excel at certain levels of dryness/wetness more than others. Unfortunately, the frequency of wet-weather races is relatively low, meaning that most drivers experience less than 20 wet weather races in their entire career. Any attempt to break wet races down into further categories (e.g., mixed weather, full wet weather) for comparing performances under these different conditions is thus doomed to fail, as there simply aren’t enough data to make any meaningful conclusions.
I therefore chose a binary definition: races are either wet or dry. I defined a race as being wet if there was a wet surface at any point during the race (including the start). Overall, 15% of Formula 1 races since 1950 have been wet, by this definition. On the left hand side, you can see a chart of all the wet races in history. Also, because it’s hard to actually find a comprehensive list of all wet races, I have included the list I compiled at the bottom of the article.
There have been 6 seasons where at least one third of the races were wet: 1952 (3/8), 1954 (3/9), 1966 (3/9), 1981 (6/15), 2000 (6/17), and 2008 (6/18).
There have also been 8 seasons with no wet races at all: 1957, 1959, 1964, 1969, 1970, 1982, 1986, and 2014 (so far).
Generally, wet races are considered more entertaining by fans. When F1Fanatic analyzed the user ratings of 100 races from 2008-2013, 17 of the 21 wet races from that time period appeared in the top 50. Wet races were much higher rated than dry races on average (28th vs. 56th) with a statistically significant difference (p = 0.00007, Wilcoxon rank-sum test).
Now, how might one try to quantify wet-weather driving performances? The simplest approach, which others have looked at, is to check what percentage of wet races each driver won across their career. This statistic can potentially be skewed by mechanical failures, so for each driver I first excluded all races with mechanical or other non-driver DNFs, just like in my last post. The remaining starts (including finishes as well as races where a driver crashed or otherwise ended the race by their own hands) were analyzed. For this analysis, I used only the 50 drivers who scored at least 5 wins in their career.
When you do that analysis, here is what you find.
|#||Driver||Percentage of wet starts won|
It’s worth noting that Ascari’s incredible 80% is based on a very small sample — he won 4 of the 5 wet races he started. Senna, in 2nd place, won 14 of 21 wet races he started (without non-driver DNFs), putting him well ahead of any other modern drivers. This analysis provides a measure of absolute performances in wet weather, although still partially confounded by car performance. While some drivers scored very well in absolute terms, wet weather was not necessarily their forté. For example, Jim Clark won 42.9% of all wet races he started, but he won 56.4% of all dry races he started! In other words, his odds of winning a wet race were less than his odds of winning a dry race. For many other drivers, the opposite is true.
Here is what happens if we rank drivers by the difference between their wet weather and dry weather winning percentages.
|#||Driver||Wet versus dry difference in percentage wins|
Senna now emerges on top, with Ascari close behind. Very few drivers have much higher win rates in the wet than in the dry. Button, Villeneuve, and Moss all stand out as exceptional wet-weather drivers. Interestingly, Schumacher’s win rate in the dry was almost the same as his win rate in the wet; he simply excelled in all conditions.
An alternative method of comparing wet and dry performances is to take the ratio of the winning rates in wet and dry conditions (an alternative method that many statisticians prefer is to take the “odds ratio”, but many people find odds less intuitive to think about than probabilities). This tells us the factor by which a driver is more likely to win in a wet race than a dry race. A value of 1 means a driver is equally likely to win under wet and dry conditions. A value greater than 1 means a driver is more likely to win a wet race than a dry race. A value less than 1 means a driver is less likely to win a wet race than a dry race.
|#||Driver||Wet versus dry winning probability ratio|
Button stands well ahead by this metric — he is an astonishing 4.75 times more likely to win a wet race than a dry race. Of Button’s 15 wins to date, 8 are from his 179 dry starts, while 7 are from his 33 wet starts. Massa is ranked second of the currently active drivers. Although many remember his disastrous performance in the wet at Silverstone 2008, he has 8 wins from his 153 dry starts, versus 3 wins from his 30 wet starts. Among other current drivers, Hamilton’s ratio is 1.19, Alonso’s ratio is 0.75, and Vettel’s ratio is 0.50.
Several drivers never won a race in the wet and so they have a ratio of zero. Every driver with 12 or more wins won at least 1 wet race in their career. The least proficient wet-weather driver of this group was Hakkinen, with a probability ratio of 0.15. He won 19 of his 86 dry starts, but only 1 of his 30 wet starts. Conforming to reputation, Prost was also typically less successful in the wet than the dry, with a probability ratio of 0.40.
One of the greatest challenges of wet racing is simply getting to the finish. After excluding races where a driver did not finish for a reason beyond their control, I analyzed all starts of the 50 drivers with 5 or more wins. Overall, 12.5% of the dry starts ended in a driver-related DNF (crashing or retiring for another driver-related reason, as defined in my last post), whereas 20.5% of the wet starts ended in a driver-related DNF.
A handful of drivers paradoxically had lower driver-related DNF rates in wet races than dry races. Most notably, Senna had 16 driver-related DNFs in his 100 dry starts, but only 2 driver-related DNFs in his 21 wet starts. Button, on the other hand, follows the typical trend of crashing more frequently in wet races, with 14 driver-related DNFs in his 179 dry starts, and 5 driver-related DNFs in his 33 wet starts.
Points per race
So far, I have only considered wins as a performance metric. For most drivers, wins are a relatively rare event, so this will not be the most reliable metric, nor the most inclusive metric. An alternative is to consider points scored in dry versus wet races. To do this, we need to score points in the same way for all drivers, so I retroactively applied two different points scoring systems to all races: the 10-6-4-3-2-1 system that was used from 1991-2002 and the 25-18-15-12-10-8-6-4-2-1 system that has been used since 2010.
For this analysis, I included all drivers who scored at least 10 wins worth of points across their career — 100 points under the old system and 250 points under the new system. I again excluded non-driver DNFs for each driver. I calculated the average points scored per race under both wet and dry conditions. I then took the difference between the wet and dry scoring rates to assess how much better each driver performed in the wet than in the dry.
First, these are the rankings under the old 10-6-4-3-2-1 scoring system.
|#||Driver||Wet versus dry points bonus per race|
Many of the usual suspects are at the top of the list, although we now also see some wet-weather specialists who didn’t make the cut under the 5 wins criterion: Boutsen, Frentzen, Heidfeld, Kubica, and Alesi. Button is again the highest ranked active driver. Vettel is surprisingly well down the list, with -2.02 points per race in wet vs. dry conditions.
The next table shows the rankings under the modern 25-18-15-12-10-8-6-4-2-1 scoring system.
|#||Driver||Wet versus dry points bonus per race|
This time, Gilles Villeneuve rises to the very top of the list. Notably, Massa also eclipses Button as the highest ranked active driver. Again, it emerges that Vettel, on average, scores at a considerably lower rate in wet races, to the tune of 5.27 points. In his 92 dry starts, he has scored 1475 ‘modern’ points (16.03 points per race) and 35 wins, with only 4 driver-related DNFs. In his 21 wet starts, he has scored 226 ‘modern’ points (10.76 points per race) and 4 wins, with 3 driver-related DNFs.
These metrics provide a sense of who were the greatest wet-weather drivers in the history of the sport. Certain names consistently appear near the top of each list. However, none of the metrics are ideal. Relative measures of wet vs. dry performance — like relative win rates or points scoring rates — tell us whether a driver performs relatively better in wet or dry conditions, but do not tell us their performance level in an absolute sense.
Absolute measures of success — like percentage wins in wet races — are still greatly affected by car performance. Wet races only partially level the playing field. In dry weather, 53% of all races since 1950 have been won by the year’s highest scoring constructor. In wet weather, 44% of all races since 1950 have been won by the year’s highest scoring constructor. To comprehensively answer who was the greatest wet-weather driver, we need some way of estimating both driver and car performances, so that we can adjust drivers’ results relative to the competitiveness of their cars. That is a very difficult problem, but one that I will address in my next post!
List of all wet races (year and round numbers)
1952: 3, 4, 7
1954: 1, 5, 7
1956: 4, 8
1961: 5, 6
1962: 6, 7
1963: 2, 4
1965: 3, 9
1966: 2, 4, 6
1968: 5, 6, 8
1971: 4, 10
1972: 3, 4, 12
1974: 2, 4, 11
1975: 5, 8, 10, 12
1976: 10, 16
1977: 7, 12, 15
1979: 3, 15
1981, 2, 4, 5, 8, 10, 14
1985: 2, 13
1988: 8, 9, 15
1989: 6, 11, 16
1991: 2, 3, 14, 16
1992: 4, 8, 12
1993: 1, 2, 3, 15
1995: 3, 7, 11, 14, 16
1996: 2, 6, 7
1997: 5, 8, 12
1998: 3, 9, 13
1999: 7, 14
2000: 6, 8, 11, 13, 15 16
2001: 2, 3
2003: 1, 3
2004: 2, 15, 18
2006: 13, 16
2007: 10, 15, 16
2008: 6, 8, 9, 13, 14, 18
2009: 2, 3
2010: 2, 4, 7, 13, 17
2011: 7, 9, 10, 11, 16
2012, 2, 20