Motorsport Analytics

I usually do these rapid quantitative previews as Bluesky threads, but my brain’s too fried to write this somewhere I don’t have an edit button today (and also probably too fried to be concise; I’m one of those “this post is longer because I didn’t have time to make it shorter” people). So here we are.

Median Laps

Let’s start with the simplest questions. Who’s been overall fastest, on average, in free practice? In Hypercar, the answer is Ferrari, with Porsche, Toyota and BMW fairly close behind:

In LMP2 it’s the Inter Europol #43, though this mostly seems to be a reflection on the drivers, since their #34 is amongst the slowest cars. Note the Pro/Am (running a Bronze) cars have a much wider range of laptimes than the non-Pro-Am cars (running a Silver). This is not surprising as Pro drivers are quicker than Silvers and Silvers are quicker than Bronzes. The exception seems to be the #34; I might somewhat uncharitably say that Nick Boulle’s a pretty good Bronze and Jean-Baptiste Simmenauer and Luca Ghiotto are weak Gold/Platinums…

For the GTs, it seems like the Lexuses have a slight edge, followed by the Ferraris, and Corvette, McLaren and Mercedes are struggling somewhat.

However, remove the Bronze drivers, who tend to have significantly more spread, to get more at the overall speed of the cars, and Porsche jump forward somewhat (despite the Iron Dames not having a Pro driver for most sessions!). Mercedes move forward a bit but despite their very bad Bronze drivers it’s not as much as you might hope. That might be because the Bronzes have more been crashing the cars than driving slowly in them…

Long Runs

So that’s the basic status of how lap times are looking at the moment, what if we look a bit deeper? First off, how are the cars looking on their longer runs? I’ve thrown out any stint of five laps or less (and out laps, in laps and non-green flag laps) and then looked at the average time each car manages for the n-th lap. Technically it’s a smoothed loess method but that doesn’t really matter it’s basically a way to make the graph look a bit less pointy than taking a straight average. Now this still includes things like getting stuck in traffic, but then so will the race. It looks pretty close in Hypercar, with Toyota, Porsche, Ferrari and Cadillac all being similarly-paced. Cadillac maybe seem to take a bit longer to bring their tyres in.

I’m not going to try this presentation method for the LMP2 cars because there’s just too many of them for it to work. But looking at the graph myself and spending far too long trying to work out who was who, I think it’s Inter Europol who are looking strongest again. GT3 though we can bring it back up…

Well done Corvette, you did a very long run. It’s Porsche and Ferrari who seem to have the best pace though. That said, if the long Corvette run was a bit of an experimental fuel save then it’d probably be slowing the overall average down a bit.

Track Temperature

Le Mans, being a 24-hour race, includes some night-time running (I hope I’m not introducing too advanced a concept here). That means the track temperature varies quite a lot during the course of the event. Fortunately, ACO are sensible enough to organise night-time practices, so we’ve got some idea of how fast the cars run in cooler night-time conditions. Let’s take a look. Here I’m just plotting lap time against track temperature and then trying to fit a linear model to it. (This is because if I use a loess it goes silly in the middle bit where there’s no laptimes)

It seems like in the cooler conditions it’s Porsche, Toyota and especially Ferrari who perform best, though Toyota spent slightly longer on long runs than the other two did. Either way I’d favour the traditional favourites when the sun goes down.

Again this would be a forest of lines if I tried it for LMP2, and with Bronze drivers getting required night-time laps in I’m not sure if it’s even meaningful. It is Inter Europol, once again, who come out looking best, so I’m tentatively calling them favourites given they’re also starting second.

There’s not a lot in it with the GT3s, really. Lexus look fastest on average, but not actually by as much as they do when the track’s at it’s hottest. Seems like if any car really likes the cold it’s the McLaren, but they’re not on the pace overall. Again though there’s an issue here with Bronze drivers getting night laps in.

If I remove the Bronzes the Lexus actually does rather better in the night while it’s pretty even in the heat of the day, and the BMWs are apparently faster in the dark? I mean there are several thousand memes about BMW drivers not looking where they’re going but apparently…

At any rate, it seems like BMW and McLaren ought to do relatively better when the sun sets.

Top Speed

Finally, who’s literally going fastest? Since quite a lot of the Circuit de la Sarthe is a straight line, top speed is more important here than most places. Let’s take a look.

The absolute fastest anyone’s gone is Antonio Fuoco, who clocked 347.8 kph (~217 mph) in FP3, though as that’s 3 kph faster than anybody else has managed he must have had a heck of a tow. On average it’s been BMW quickest through the speed traps, but by barely anything from Porsche:

(The less said about the Peugeot 9X8 and it’s completely uncompetitive BoP the better, I feel). In LMP2, it looks like the secret to Inter Europol’s impressive performance with the #43 is that they’re significantly faster in a straight line.

And in the GT3 class, the fastest car is apparently the McLaren, at least on average, though the Ferrari has managed a higher maximum speed.

Though this is apparently because Ferrari’s Bronzes are slowpokes (actually I’d say it’s more than McLaren’s are both pretty decent) as the Scuderia get the top spot if we leave the Bronze drivers out:

What does it all mean?

Watch the race and find out! Hypercar looks like it’ll be pretty close. The Inter Europol #43 does look like a pretty good bet for LMP2, and in GT3 I think Ferrari probably seem like they have the edge particularly given this is only Akkodis ASP’s second Le Mans with the Lexuses.

But I threw all this together in like three hours so odds are there’s at least one glaring coding error or I accidentally plotted Imola data or something so definitely don’t make any bets based on it!