“I’m happy to qualify 11th because we get a free choice of tyres tomorrow. We also start from the clean side of the grid.” How many times did you hear a driver say words to that effect when the tyre rules were in place? Were they just putting a brave face on things, or is there genuine evidence that qualifying 11th was superior to qualifying 10th? Let’s look at the data.

Firstly, the issue of where a driver qualifies can be surprisingly messy. Grid penalties are regular, sometimes drivers are excluded from qualifying, and semi-regularly drivers in Q3 will not complete a Q3 lap. The clean side of the grid is also not consistent on different tracks. For this investigation I’m comparing qualifying positions and nothing else (i.e. not starting position). This approach has some weaknesses (as does every other approach), but it keeps the problem simple at least.

Different criteria will be looked at from qualifying results of 9th-12th and data between 2017 and 2020:

1. Amount of points finishes.
2. Percentage of races finished ahead of others.
3. Points scored.
4. Average finishing position.
5. Chance of DNF.

Fig 1: Percentages of races in the points, and the top 7, respectively.

This first graph summaries the data well. The data for all four starting positions is fairly similar, meaning we need to be careful in our conclusions. This is one reason why several other metrics have been looked at too. However, this data suggests that qualifying 9th is slightly superior to 10th or below. In fact, qualifying 9th comes out on top of every metric I looked at except one. Qualifying 10th and 11th seem to yield similar results to each other, whereas 12th has a similar rate for a points finish, but has a significantly lower chance of getting a good haul of points, which I’ve defined here as 7th place or higher.

Fig 2: Percentage of races that the qualifying position ended up as the lead car of the four.

Another way of measuring success is comparing how often each starting position yields a superior finishing position. This is useful because it is directly comparing the cars we want to compare, without having to worry about the effect of several cars ahead of us retiring, for example. Here 9th is clearly leading again, with 11th slightly ahead of 10th.

One surprising result is that qualifying 10th has yielded more points at the end of the race than qualifying 9th. This is partly due to freak results falling in favour of drivers starting 10th (i.e. good/bad luck), but is also partly a result of how the qualifying system works: if a driver does not set a time in Q3 they will typically end up 10th. This means that a driver in a top team is more likely to qualify 10th (due to crashing or unreliability in Q3) than 9th. The median result still falls in favour of 9th place, emphasising that it is still the preferred option. I included DNFs to see if there was a significant difference due to being further in the fact, but I don’t think there’s enough data to draw a proper conclusion. Note that data with the words “finishing position” (e.g. “Median finishing position”) assumes a race finish, whereas most of the data does not.

One more thing to consider is how random the data as a whole is. For just one example of this, let’s look at the yearly changes that occur in median finishing position:

Fig 3: How the median finishing position varies year-to-year.

The variation year-to-year is greater than the differences between positions! This example is not cherry picked, and the fact is true for other metrics too. This is the reason to use several years of data. It’s also why we are looking for consistent trends across different metrics. Of course, if race results were totally predictable from qualifying they’d be no point in watching the race, so the randomness is by no means unwelcome, even if it does make understanding the data a bit more difficult.

As a side point, pouring through the data reminded me of just what a crazy season 2020 was. Drivers who qualified 8th-11th scored significantly more points in 2020 than in 2019, 2018 or 2017, despite the season having fewer races. In addition, there were 4 podiums in 2020 for those drivers, compared to just 1 for the three previous seasons combined.

Conclusions

• Qualifying 9th is, on average, better than qualifying 10th, 11th of 12th. They’re more likely to score points and finisher higher up.
• It’s not clear if qualifying 10th or 11th is superior. 10th is ahead for points scored, but this is mostly due to just a few results. In other, probably more representative metrics they are very similar. This suggests there is some weight (although not much) to the idea of 11th being a good starting place, considering that the natural assumption is that it’s obviously inferior to 10th.
• Both 10th and 11th are significantly better than qualifying 12th. Out of he four positions 12th is the worst for everything.
• There is a lot of variation in results race-to-race, and even year-to-year.

Hope you enjoyed the read. Apologies the conclusions were not particularly definitive on the 10th vs 11th question, but that’s the way it goes sometimes!