If we take the statement as a statistical prediction, we have to clarify what it means, exactly. I’m going to assume it means that 90 % of the time, one of the cars starting in the first row (the first two positions) will win the race. Very little research is needed to prove that false. In the past 25 years, the Monaco race has been won by someone in the first row only 80 % of the time. The data do not support the hypothesis.1¹ 20 out of 25 races is just over 1.645 standard deviations away from what would be expected if the true fraction was 90 %. Thus it is outside the significance threshold I use for casual analysis.

But maybe the driver didn’t mean it with statistical accuracy. Maybe they were just trying to say that a first-row winner is more common in Monaco compared to other tracks. I went into the Wikipedia pages for a few randomly selected long-running Grands Prices2² I’m well aware the plural of prix is prix., and gathered the following statistics.3³ At the time I’m finishing this article, this data collection happened a couple of years ago, so I don’t remember how far back I went for this data. Possibly the same 25 years.

Sure, Monaco is up there, but it’s not the most extreme of these tracks. At this superficial level, there is nothing about the Monaco number that makes it any different from any of the other tracks.

In Swedish, the word ekipage comes from equestrian sports and means “the horse-and-rider team”. For this article, we’ll convert the spelling to the more English-sounding equipage and use this term to mean the driver-and-car team in a Grand Prix.

This matters because while we sloppily talk about “driver skill”, even the best driver needs a good car to perform well. And not all drivers are equally good on all tracks. So to avoid this mistake, in this article, we’ll talk about equipage capability to discuss the potential performance of the driver-and-car team on a specific track. If it makes it easier in your head, feel free to substitute that with “driver skill” but remember that other factors play a role.

The assignment of starting positions in a race is not random. Instead, drivers take turns trying to set the fastest lap around the track the day before the race, and the two equipages with the fastest laps in the final round get to start on the first row in the race. This introduces a very annoying confounder: naturally, higher equipage capability improves the chances of winning the race, but it also improves the chances of starting in the first row through qualifying well.

If a first row equipage wins the race, is that because they started in the first row, or did they start in the first row because they were high in capability on that track, and that’s also why they won? The causal graph looks like this.

In order to measure the effect of a first row start in this system, we need to control for equipage capability. One way to do that is to include it as a separate predictor in a regression analysis. The idea is that the equipage capability coefficient will eat up most of the effect of equipage capability, and that leaves the first row start coefficient to contain just the effect of the first row start.

But that requires being able to measure equipage capability. One way to do that is to take the driver’s championship points at the end of the season, but the drawback of that is that it doesn’t tell us anything about differences in equipage capability across different tracks, or as it varies over a season. We calso cannot use the qualifying results to measure equipage capability, because the reason we are doing this in the first place is to separate out the effects of equipage capability from qualifying results.

That’s where I got stuck for a while. Then a couple of years later I had a flash of insight!

Qualifying proceeds in three rounds. If we take the worst result from each round, that might represent a kind of capability baseline of that equipage. And it turns out it is a decent proxy for equipage capability, too: if we order drivers based on their average of this measurement, and compare to the driver’s championship results, the correlation is +0.82. This means, loosely speaking, that around 70 % of the variation in the driver’s championship result is determined by which equipage can drive fast. The other 30 % is luck and race dynamics such as being skilled at overtakes, etc.4⁴ On the one hand, I’m surprised as much as 70 % of the driver’s championship is about driving fast. On the other hand, maybe that’s not so strange if “driving fast” is what gives drivers a start in the first row, and a start in the first row is what wins races? I have not tried to tease those effects apart.

Now that we have a way to measure equipage capability on specific tracks, we can start modeling. I found easily accessible data for both qualifying and race results covering seasons 2022–2025.5⁵ I planned on delaying the publication of this until the week after the 2026 Monaco race to get one more year of data in there, but by the time I’m finishing this article up the official results from the weekend are still not available. So the below analysis proceeds without that. Fitting a logistic regression model to this data, we get the following coefficients out:

Here, equipage capability is measured in seconds of qualifying delay6⁶ Meaning how much slower an equipage’s slowest round’s fastest lap is compared to the fastest equipage on that track in that season. This means an equipage who’s baseline qualifying performance is one second slower than the best is 0.86 log-odds less likely to win.. It’s not super interpretable, but it doesn’t matter anyway because in the rest of this article, we’ll hold it constant at the average equipage capability. This is how we subtract away the effect of equipage capability to zoom in only on the effects of starting position.

To compute the probability of winning from e.g. the pole position, we add the log-odds of the intercept, the pole position, the first row (because pole position is part of the first row), and the average equipage capability. We get a log-odds of 0.124. We convert that log-odds into a probability, and we get 53 %.

We repeat that calculation for all the starting positions we can given the coefficients above. To reemphasise, equipage capability is held constant at the average for all these positions, meaning the results below are purely effects of starting position. It’s as if we took the average equipage and cloned them into all 20 positions of the starting grid, and had them race themselves. This is how often they would win from each position:

• Pole position: 53 %
• P2 (still first row): 19 %
• Second row: 16 % in total (8 % per equipage)
• Further back: 13 % in total (0.8 % per equipage)

This means that the odds of any of the first-row equipages winning is 17 times the odds of anyone from behind the second row winning. I hear the Formula One fans shouting already: “Upset victories are far more common than that! There’s no way winning from the back of the grid is that uncommon!”

But it is – when we hold equipage capability constant at the average. When someone from the back of the grid catches up and wins the race in the real world, it’s always a high-capability equipage that happened to get a bad qualifying result. Great equipage capability can make up for some of the starting position disadvantage, but the disadvantage is real and massive.7⁷ Concretely, to counter the effect of starting in the second row instead of the first row, the equipage capability must be 2 seconds better. That’s a huge difference in capability. It’s basically the span from the 7th percentile to the 93rd percentile. The average equipage practically never ends up winning from the back of the grid.

Aaanyway, we were talking about Monaco. We would like to perform the same model fit on only the Monaco races. The problem is I only have data for the four seasons 2022–2025, and since there are five coefficients to fit, it’s not going to happen. In fact, even with just four coefficients it doesn’t converge sensibly.

This means we cannot measure the effect of starting in each position individually. We have to be content with looking at only one of (a) pole position, or (b) the entire first row. When I tried one at a time, the effect of starting in pole position was more stable with the little data there is. So we go back and do over, except only with pole position, for non-Monaco tracks:

• Pole position: wins 42 % of the time
• Rest of the field: wins 58 % of the time (3 % per equipage)

and for Monaco:

• Pole position: wins 70 % of the time
• Rest of the field: wins 30 % of the time (1.6 % per equipage)

Remember, this is when we hypothetically race the average equipage against 19 clones of themselves. The pole position confers a larger advantage in the Monaco race than on the average track, only due to the position itself. It has nothing to do with driver skill, car performance, etc.

However, we weren’t interested in the average of all other tracks. We wanted to see where Monaco sits among the other tracks. The limited data I have doesn’t allow comparing all tracks, but these are the ones for which the model converges sensibly.

Is Monaco decided at qualifying? Yes, sure, one could say that. But so are many of the other races.

Is Monaco decided by starting position nine times out of ten? No. Australia gets close, but Monaco is far from it. Driver skill still helps in Monaco.