2019 F1 season: The pit crew
- Like
- Digg
- Del
- Tumblr
- VKontakte
- Buffer
- Love This
- Odnoklassniki
- Meneame
- Blogger
- Amazon
- Yahoo Mail
- Gmail
- AOL
- Newsvine
- HackerNews
- Evernote
- MySpace
- Mail.ru
- Viadeo
- Line
- Comments
- Yummly
- SMS
- Viber
- Telegram
- Subscribe
- Skype
- Facebook Messenger
- Kakao
- LiveJournal
- Yammer
- Edgar
- Fintel
- Mix
- Instapaper
- Copy Link
I will start my end of season review by writing about the incredible job done by the pit crew over the course of a Formula 1 season. This year, we saw the fastest pit stop ever, and I have no doubt that the record will be broken soon once again. Pit crew members show great dedication, and without them, Formula 1 wouldn’t be what it is today. Let’s get started by taking a look at the numbers.
Methodology
Let’s begin by explaining my methodology. I collected the data from every single pit stop done over the season and did a variety of analyses to obtain meaningful results.
Take a look at the chart shown at the top of this page. 94.7% of all the pit stops done over the season are shown in the chart. As you can see, most of the pit stops took between 2 and 4 seconds, with some of them taking longer than that average time.
Every stop shown in the chart was added together, and the average result was obtained. This number was of 3.27 seconds. That is the average pit stop time for 94.7% of the data points. The remaining 5.3%, however, has an average time of a stunning 18 seconds. What does this mean? Well, that on average the pit stops were very fast, taking less than 3.5 seconds each. The slowest stops, however, were very slow, and include blunders like the one done by Mercedes in Germany.
If I were to take an average of all the stops, including the extremely slow ones, the average would be extremely skewed. In order to correct for that, I decided to use the geometrical mean, instead of the arithmetic mean (or average as it’s usually called). This geometrical mean is not as sensitive to outliers as the arithmetic mean and shows a more meaningful representation of the data.
The chart shown above clearly explains my point. The black dotted line shows the average, taken by considering every single pit stop time done over the season. The green dashed-dotted line shows the geometrical mean, which considers every pit stop time, but isn’t as sensitive to outliers as the regular average calculation. The red dashed line shows the average, but taken after removing the outliers, meaning the blunders and challenging pit stops that took an abnormal time to get done. Finally, the blue line shows a locally weighted regression. This calculation represents the rolling trend of the pit stop times after removing the outliers.
You may think that the red line is more representative of the data points. I mean, it seems like most points are actually around that line right? You would be right with your conclusion. Then why not use it, instead of the geometrical mean? There’s a good reason for that.
While the average calculated after removing outliers is more representative, it doesn’t consider the slow times at all. This calculation basically pretends that the 58.24 seconds pit stop done by Toro Rosso in Abu Dhabi doesn’t exist. Is that fair? I don’t think it is. At the end of the day, blunders are part of the game, and must be taken into consideration when analyzing our data.
The geometric mean seemed like a good middle ground between the arithmetic mean with outliers and the arithmetic mean without outliers. The geometric mean punishes teams for making costly mistakes, but a single mistake doesn’t destroy their whole-season average. At the same time, while it isn’t as accurate as the mean after removing outliers, it still takes into consideration the previously mentioned mistakes.
It is because of the previous reasons, that I decided to use the geometrical mean as the basis of this calculation. If you have any questions about how it works in detail, just post a comment below and I’ll give you more in-depth information about how it is calculated.
Pit stop times per team
This chart shows exactly what I was just talking about. The Mercedes’ pit crew was in general very fast, but they had a horrible pit stop in Germany after Lewis Hamilton went unexpectedly into the pits. That single pit stop would have been enough to send them towards the bottom of the rankings if they were ordered by using the arithmetic mean. The geometric mean punishes them for their mistakes, but still shows their speed over the course of the season.
The chart needs a bit of zoom though. I mean, we can see everything here, but there’s not enough detail to understand why Red Bull was the fastest pit crew.
Ah, things make more sense now. Red Bull made several of the fastest pit stop times of the season, and was incredibly consistent. Most of the pit stops took less than 2.75 seconds, with several of them getting closer to the 2 seconds mark.
Mercedes and Williams had the same mean time after 21 races. Pretty incredible if you ask me. While Williams did faster stops that Mercedes ever did, they also made more mistakes. We will see more about that in the next section of this article.
Ferrari finished fourth in this category, edging Renault by just 4 hundredths of a second. The data once again shows something similar to what happened with Mercedes and Williams. Ferrari had some pit stops that were faster than Renault’s, but Renault made less lengthy stops.
Haas might have been disappointing with their season, but their pit crew was quite commendable. They struggled with some pit stops, especially at the beginning of the season, but corrected course and finished as the 6th fastest crew.
Detailed analysis
General information
The detailed analysis gives us more insight into the numbers that were are seeing. The table shows, from left to right, the following:
- The name of the team
- The number of pit stops performed by that team over the season
- The fastest pit stop of the season
- Columns 4 to 7 represent the 25th, 50th, 75th and 90th percentiles. They indicate the value below which a given percentage of observations in a group of observations falls. For example, Red Bull’s 25th percentile (2.17), means that 25% of their stops were of less than 2.17 seconds. Their 50th percentile (2.52), shows that half of their 67 stops, took at less than or equal to 2.52 seconds. The same applies to all of the other percentile numbers shown.
- Column 8 portrays the slowest pit stop of the season for each particular team.
- The 9th column shows the geometric mean for each team, in seconds, as well as their rank based on this calculation.
- The 10th and final column represent what I considered to be “slow” pit stops. The threshold for this calculation was 3.58 seconds. This number was decided as it is the 75th percentile of all the pit stops done during the year. This means that every lap that took more than 3.58 seconds was among the 25% slowest laps of the season.
Analysis
We see some very interesting trends in this table. Take for example Mercedes and Williams. Both of them have the exact same geometric mean (3.11), but their numbers are in fact very different. Williams performed 25% of their laps with a time of 1.97 second or less, while Mercedes’ 25th percentile was of 2.11 seconds. This means that when Williams was fast, they were much faster than Mercedes.
How come they have the same geometric mean? Well, take a look at their 90th percentile. Williams’ did 90% of their stops in a time of 5.47 seconds or less, while this number is of 3.87 seconds for Mercedes. The meaning of this? Mercedes’ crew did less slow pit stops than Williams. The “slow stops” column tells us exactly that. Williams had 20 pit stops, the most of any team, that took over 3.58 seconds. Mercedes, on the other hand, only did 9 pit stops over the full season that were 3.58 seconds or more.
Another interesting piece of information comes from Racing Point and McLaren’s data. The Pink Panthers narrowly edged McLaren, even after having worse 25th, 50th, and 75th percentile numbers. Both teams did 19 slow stops, so why did Racing Point comes out on top of McLaren? The 90th percentile column, as well as the column showing the slowest pit stop of the season, tell us the answer. The Racing Point crew members were usually not as fast as the McLaren ones, but their slow stops weren’t horribly slow. McLaren was faster in general, but their slow stops were very slow. Racing Point average stop time in those slow stops, was of 6.32 seconds. McLaren’s time? 10.8 seconds per stop. This comes to show you how being fast is not everything. Consistency, and limiting costly mistakes, can take you a long way.
Toro Rosso and Alfa Romeo had the two slowest pit stop crews of the season. I will talk about them in one of the following sections.
Fastest pit stops of the season
The analysis of the 20 fastest pit stop times of the season is dominated by Red Bull and Williams. The Austrian team did the fastest five pit stops, all coming after the 10th race, Silverstone. How fast were they? Well, all of their 5 fastest stops were of less than 2 seconds. In total, 11 of the 20 fastest stops recorded this year, were performed by Red Bull’s pit crew. Just astonishing.
Williams had a season to forget, but if there’s a silver lining, it is the performance of their pit crew. In total, Williams did 7 of the 20 fastest stops of the year, with 3 of them being of 2 seconds or less.
The remaining two stops in this chart were done by (1) McLaren, in Spain, and (2) Ferrari, also in Spain. Those two stops may not look as impressive after seeing the incredible work done by Red Bull and Williams, but a sub 2.10 stop is remarkably impressive.Â
Slowest pit stops of the season
We reach the section that teams like Alfa Romeo and Torro Rosso don’t want to see. Here you go, the 20 slowest pit stops of the season.
Mercedes had the slowest pit stop of the season, of 50.39 seconds, until the last race of the year. Toro Rosso, however, took away from them the not so honorable distinction of performing the slowest pit stop of the season. After Pierre Gasly received front wing damage during the first lap of the Abu Dhabi Grand Prix, he went into the pits to replace the whole wing. Unfortunately, the mechanics struggled to remove the piece, and it took them almost a full minute to release Gasly.
McLaren did 5 of the slowest 20 pit stops of the season, with 4 of them taking 20 seconds or more. As I said in the previous section, when McLaren’s crew was fast, they were really fast. When they were slow, they were, unfortunately, quite slow as well.
Alfa Romeo finished the season as the slowest pit crew. They had 25 pit stops of more than 3.58 seconds, the most of any team, and are represented in this chart with 3 of the slowest stops of the season. Unlike McLaren, Alfa Romeo was the slowest team in most of the metrics considered in this analysis. Alfa Romeo has the worst 25th and 50th percentile times, meaning that they were never very competitive, even in their fastest pit stops. The Swiss-based team also had the second-worst 75th and 90th percentile times, meaning that when they were slow, they were also quite slow. Their mechanics will have to work hard during the off-season in order to catch up before the beginning of the next season.
Final remarks
This year, 50% of the pit stops took 2.92 seconds or less. Just think about that. 2.92 seconds. That’s more or less what took you to read this paragraph. That is absolutely astonishing.
Teams like Red Bull and Williams have set an exceedingly high bar for the rest of the teams, and themselves, for next year’s season. That is what Formula 1 is all about. Innovation, surpassing the limits, and always moving forward.
It is very easy to forget about the unsung heroes of the season, which include the pit crew members. Races were decided because of their performance, and they deserve recognition for their wonderful contributions to Formula 1.
Thank you, pit crew member, for your commitment, discipline, and passion. People like you shouldn’t go unnoticed, and in this blog, you will always be recognized as the great performer that you are.
I hope that you have enjoyed this article. If you did, please share it with your friends and let me know what you think in the comments below.
- Like
- Digg
- Del
- Tumblr
- VKontakte
- Buffer
- Love This
- Odnoklassniki
- Meneame
- Blogger
- Amazon
- Yahoo Mail
- Gmail
- AOL
- Newsvine
- HackerNews
- Evernote
- MySpace
- Mail.ru
- Viadeo
- Line
- Comments
- Yummly
- SMS
- Viber
- Telegram
- Subscribe
- Skype
- Facebook Messenger
- Kakao
- LiveJournal
- Yammer
- Edgar
- Fintel
- Mix
- Instapaper
- Copy Link
Tremendous job! Wow! Thank you!
Thanks
I’ll be posting more long articles during the next few days. I mean, the season is over, but there’s still so much to analyze right?
What about nose changes etc. They take longer but the crew can still perform very good on them.
Hello Sieper
I grouped all the pit stops together, that includes the ones in which the crew changed the nose. You are correct by saying that they were still very good on them, but they do affect the overall time average.
By using the geometric mean (instead of the average), those slow-ish pit stops caused by changing the front wing are still considered in the analysis, but they don’t ruin the mean number. For example, Mercedes’ crew took around 50 seconds to change the wing in Germany. If I had used the mean(average) for the analysis, then that stop would have ruined Mercedes’ time. Basically, it would have said that the Mercedes’ crew was slow, which isn’t true at all. By using the geometric mean, it shows that their crew is actually very good.
I hope that kinda makes sense.
By the way, I want to thank you for your kind message and your donation. I really both of those things, I really do.
I will create a section to thank all the people who have donated to the site, so don’t be surprised if “Sieper” shows up there soon =)
Hello, I was wondering if you get the pitstop datas somewhere or if you collected it yourself in a sheets ?
Hello Dorian
If I remember correctly, I got this data from the DHL website and just compiled it into a single data frame.