I’m back with the pit stop data! This time, I decided to take it to a whole new level. Pit stops are one of the most important elements in modern F1, and the pit stop crews deserve a lot of respect for the job they do week in and week out. Still, one of the most difficult aspects of judging a pit crew is the fact that pit stop times are highly variable, and a single pit stop can make the best team look like the worst. So, I’ve decided to create a robust model to address this issue and answer the question: Which team has the best pit stop crew in Formula 1?

The raw pit stop data

First of all, let’s take a look at the raw pit stop data. It shows exactly how the stops happened during the season—no adjustments for conditions, just all the stops compiled and summarized. As you keep reading, you’ll see that this approach comes with some subtle but noticeable downsides.

Raw performance per team

Looking at the data, grouped and summarized by team, the Ferrari pit crew comes out on top. Unlike the past few years, when Red Bull were dominant in the pit stop department, the battle is much more competitive in 2025. Ferrari and their mechanics are doing a fantastic job, performing at the top of their ability during the first 12 races of the 2025 season. Their average of 2.41 seconds per stop is 0.17 seconds faster than Racing Bulls’. It may not sound like much, but considering that each pit stop takes, on average, less than three seconds, this is actually a significant gap.

Speaking of Racing Bulls, it’s quite interesting to see them ahead of their big brother, Red Bull. Their data is intriguing: on average, they are beating Red Bull by 0.07 seconds per stop, but it appears that their maximum, raw pace isn’t quite on Red Bull’s level. However, they have made fewer major mistakes than the Red Bull crew, which helps lower their average substantially.

Looking down the rankings, Mercedes and Sauber are competing neck and neck, with Mercedes ahead by just 0.02 seconds per stop. The bottom five teams in this ranking are separated by around a single tenth, with sixth-placed McLaren averaging 2.89 seconds per stop, compared to 3.01 seconds per stop for Haas. The raw data, however, tells an interesting story. While McLaren are technically just a little faster than the worst-performing teams, it seems like they have a lot of pace in reserve. They have posted several fast stops—even more than better-ranked teams such as Mercedes and Sauber. Their problem? They have been very inconsistent and have had many slow pit stops, which heavily pulls their average up.

Raw performance per driver

I think it’s fair to say that, normally, you would expect both drivers from the same team to have similar pit stop times. Each team has a single pit crew that works on both cars when they stop for new tires. Still, reality is messier, and it’s rare to see two teammates with exactly the same average pit stop times.

Most teammates are, in fact, very close to each other. Hamilton and Leclerc are separated by just 0.01 seconds; Tsunoda and Verstappen by 0.08 seconds; Russell and Antonelli by 0.04 seconds. You get the idea. Nonetheless, there are a couple of very interesting results.

First, the McLaren pairing: Lando Norris has an average pit stop time of 2.82 seconds, placing him right in the middle of the pack. Oscar Piastri, however, has an average pit stop time of 2.97 seconds—0.15 seconds slower than his teammate—putting him close to the bottom of the grid, just ahead of Alonso and Bearman.

Second, the results at Haas: While neither Ocon nor Bearman boast strong pit stops—their fastest stop of the season so far is 2.49 seconds—it’s clear that Ocon has a faster average time, currently by a whopping 0.2 seconds, which is massive in the world of F1 pit stops.

Our new metric: xPT

This is where things get interesting. We’ve already looked at the raw pit stops, which show us exactly what happened during the season—including all the messiness and chaos that are part of pit stops. However, we can try to separate the true skill level of the pit crews from this chaos and understand the expected, real performance of each crew. To do this, I created my own performance metric, which I’ve titled expected Pit Time, or xPT.

xPT is the output of a statistical model that aims to identify the patterns within pit stops. The xPT metric is our model’s best estimate of how fast a pit crew should be, based on their talent and equipment. It’s calculated by analyzing hundreds of pit stops and statistically adjusting for all the “noise” and context. By removing the factors that obscure our data, xPT gives us a single, fair number to compare the real, repeatable skill of each team’s pit crew. A team with a low xPT is fundamentally fast, regardless of whether they got lucky or unlucky on a particular Sunday.

The xPT model is adaptable and gets updated as more data is gathered week after week. The more data the model has, the more accurate it becomes. Its predictions and inferences get stronger as the season progresses.

Using xPT as a metric, we can see which teams are performing better than expected, which are performing at their average, and which are truly struggling. One of the main advantages of xPT is it’s interpretability since it’s expressed in the original units of our data—seconds. With it, we can create a true power ranking of the pit crews, revealing who is consistently the best—not just who happened to get it right once. Let’s take a look at our first results.

xPT by team

Looking at the “xPT by team” chart, we can see what the model “believes” to be the actual expected average pit stop time for each pit crew. Based on 2025 season data, our model indicates that Ferrari’s real, repeatable performance should result in an average time of 2.55 seconds per pit stop. Red Bull, McLaren (yes, McLaren), Racing Bulls, Sauber, Alpine, and Mercedes trail Ferrari, but are highly competitive with one another, while Williams, Aston Martin, and Haas have slower estimated pit stop times of three seconds or more per stop.

While our main chart shows the full range of plausible skill levels for each team, we can also use the model to answer a more direct question: what is the probability that one team is truly faster than another? By leveraging statistical analysis, we estimate that there is a 97.8% probability that Ferrari’s true underlying pit stop skill (xPT) is better than Red Bull’s. This means that, after removing noise from the data, it’s incredibly likely that Ferrari currently has the best pit crew in the business—at least for the time being.

xPT by driver

This chart shows xPT, but now stratified by driver. Why is there a difference between two drivers from the same team? I can’t really say for sure. Our model is good at detecting these differences, but from a statistical point of view, finding the “why” would require data that I don’t have access to. Still, we can observe some very interesting trends.

At Ferrari, for example, our model puts Leclerc at an xPT of 2.53 seconds, compared to Hamilton’s 2.57 seconds. While this is a minor difference—too small to say for certain whether it’s real or just an artifact of our model—it’s interesting to see that there may be slight but noticeable differences between drivers, although calling a 0.01-second difference in a stop significant would be a stretch.

In general, you’ll notice that most drivers have a very similar xPT to that of their respective team, which is what we would expect. The same pit crew works on both cars, so a team’s pit crew has a much greater impact than individual drivers.

Delta to xPT by team

You may have noticed that McLaren has an xPT (expected Pit Time) of 2.68 seconds. This is much lower than their raw average pit stop time of 2.89 seconds per stop. Why is this? Based on their raw pit times, our model believes that McLaren’s crew has the potential to perform an average pit stop in 2.68 seconds per stop. If you look at the raw data, McLaren has performed several fast stops—faster than those of many other teams, such as Mercedes and Sauber. However, they also have many stops that are much slower than you would expect given their underlying quality. At the moment, our model suggests that McLaren is underperforming by 0.22 seconds per stop, meaning that based on their ability, they would normally average 2.68 seconds per stop instead of the current 2.89 seconds.

On the other extreme, we have Ferrari. While our model expects Ferrari to average 2.55 seconds per pit stop, in reality, their average is 2.41 seconds. The delta to xPT shows that Ferrari is not just meeting expectations but far exceeding them, outperforming their xPT by an average of 0.15 seconds.

Currently, Racing Bulls are overperforming by an average of 0.14 seconds per stop. This is why, even though they have a slower xPT than Red Bull (as seen in our xPT by team chart), they are still outperforming their sister team in raw performance.

As for Red Bull, their delta to xPT is 0.00 seconds. This means they are performing exactly at the level the model expects from them based on the 2025 season so far. This is quite interesting, especially since you might expect to see them topping the leaderboard, but our model suggests that the level they have shown in the first 12 races is exactly what we should expect. Remember, however, that the model gets updated as more data is added, so these results can—and likely will—change as the season goes on.

Delta to xPT by driver

Just as with the previous section, this chart shows how much a driver is over or under performing based on the expectations (xPT) of the model. Some results are expected, such as Hamilton or Leclerc over performing the expectations of the model. While the model already ranks the Ferrari as the best pit crew in F1, the delta to xPT shows that they are performing even better than expected. Is it down to luck? Motivation? Just a hot streak? Our model can’t answer that, so it’s up to you. If you have any additional information you can send me up, knowing that everything that you tell me is confidential, including your identity.

One of the most interesting findings of this chart, in my opinion, is the discrepancy between the results of Esteban Ocon and Oliver Bearman. While both Haas drivers operate with the same crew, equipment, and have similar baseline potential (xPT of 3.05 and 3.06), our model reveals a significant difference in their race-day execution. Currently, our model estimates with 99.6% certainty that Ocon’s stops have been faster compared to Bearman’s.

Another big discrepancy is found at McLaren. Similar to the the Haas results, our model estimates with a 100% certainty that the pit stops of Piastri have been slower than those of Lando Norris. While the stops of both drivers is worse than expected (+0.12 s for Norris and +0.31 s for Piastri), Piastri’s stops have been consistently worse. This points to a persistent issue tied specifically to his car or his side of the garage. Maybe Piastri is missing his marks in the pit box by a little, or maybe there are subtle issues with Piastri’s car that makes pit stops more difficult. There are many possibilities, but what we can say for sure is that the team’s overall underperformance is not being felt equally by its two drivers.

The search for the fastest stop

One of the most exciting aspects of pit stops is discovering the limit of performance. Formula 1 pit crews are already extremely fast, but they’re always searching for that extra tenth in every stop. Using our model, we can calculate a prediction interval to answer the question: “How fast can a pit stop be?”—based on the current season’s data.

As of the 12th race of the season, the fastest stops have been exactly 2.00 seconds, both performed by the Ferrari pit crew at the Monaco GP and Saudi Arabian GP. For reference, last year saw seven stops completed in under two seconds, with the fastest two clocking in at 1.9 seconds, done by McLaren and Red Bull. So, what about this year? Can the teams break the 1.9-second barrier?

Our model estimates that the most likely fastest stop this season would be 1.95 seconds, which would tie for fifth place in 2024. However, the model also predicts several stops under the 1.95-second mark. To start, 50% of the curve values are under 1.95, which translates to a 50% chance of seeing a pit stop of 1.95 seconds or less.

What about the probability of seeing a pit stop under 1.9 seconds this year? The answer, according to our model’s predictive power, is 22.6%. This means there’s roughly a one-in-five chance of witnessing a 1.9-second stop, which would surpass the fastest from 2024.

You may be wondering about the best pit stop ever—the 1.8-second record set by McLaren in 2023. Can we expect to see that record broken this year? I think it’s unlikely, but not impossible. Our model puts the probability of witnessing a record-breaking pit stop at just 1.72%. While this number is quite low, keep in mind that the model’s predictions become more accurate as it incorporates additional data, so this probability could increase if pit stop times improve in the upcoming races.

Conclusions

To sum it up, this deep dive into pit stop performance using both raw data and the xPT model has offered a much clearer picture of where each team really stands in 2025. Ferrari continues to set the standard, leading in both raw pace and underlying consistency, while teams like McLaren show there’s a lot of untapped potential still on the table. The data also uncovers some fascinating differences between teammates and highlights just how much small details matter when chasing every last tenth. While breaking the all-time pit stop record seems unlikely this year, the model keeps evolving with every race, so the story is far from over. As the season goes on, it will be exciting to see which crews can truly push the limits and rewrite what’s possible in the pit lane.

My plea

This is the most advanced article I’ve written for this blog. Doing inference with statistical models isn’t easy—it takes a lot of time to develop a useful and reliable model capable of producing meaningful insights. Creating this model and writing this article took me around two weeks of hard work. If you’ve enjoyed reading it, please consider becoming a supporter to help me continue writing in-depth articles like this one. Your support truly makes a difference and helps keep the blog going. Thank you!