Share, , Google Plus, Pinterest,


Posted in:

Babushkas at Le Mans, 2015’s Race Analysed

Looking back at the Le Mans 24-hour race is a bit like looking at one of those Russian dolls. It’s all very pretty with the detailed patterns and flowers, but you always know that there is another one inside, waiting to be discovered. It is in that vein that I have spent much of the time since Le Mans, trying to work out the detail.

In a somewhat simple way, and to be deliberately provocative, Porsche didn’t win Le Mans; everyone else lost it. Nissan simply wasn’t in the fight, Toyota was just plain slow and Audi had problems. All Porsche did was to pound round and round, having learnt their lessons from last year’s race and not break down.


You can’t possibly be reading this without knowing that average lap times are going to be mentioned. Just as Russian dolls are always female, my looking at a race means average lap times. Over the years, I have become convinced that a look at the average of the fastest 20% of laps completed provides the best guide to the outright performance of a car. That’s not to say there isn’t another layer underneath, but it is a fine, if one-dimensional, indicator.

The second thing that has a big impact on the result is the total time spent in the pits. The table below shows both.

Pos. No. Car Drivers Average lap time Time in pits
1 19 Porsche Hülkenberg/Tandy/Bamber 3m 20.0s 34m 15s
2 17 Porsche Bernhard/Hartley/Webber 3m 20.6s 29m 42s
3 7 Audi Lotterer/Fässler/Tréluyer 3m 19.8s 42m 09s
4 8 Audi Duval/di Grassi/Jarvis 3m 20.2s 38m 01s
5 18 Porsche Dumas/Lieb/Jani 3m 20.7s 31m 39s
6 2 Toyota Wurz/Sarrazin/Conway 3m 24.1s 35m 10s


Note that the time spent in the pits includes the time spent taking penalties: remember that the no. 7 Audi and the no. 17 Porsche were both given penalties (the Audi had to serve a drive-through and the Porsche had a one-minute stop-and-go).

Despite this though, the winning car managed to use its 0.6s per lap advantage to overcome a 4m 33s deficit in terms of pitlane time. Multiplying 0.6 (the advantage per lap of the no. 19) by 395 (the number of laps it completed) gives 3m 57s, so some further examination is needed to work out how this happened.

Time to open a layer of the Russian doll.

Not as a result of any clever tactics, but more to do with where the car was in its refuelling schedule, both Tandy and Hülkenberg had to pit the no. 19 car while the Safety Car was out. Tandy came in as he completed his first stint, three-and-a-half hours into the race; and Hülkenberg came in at 8am on Sunday morning, after his ‘moment’ with Roald Goethe’s Aston Martin. Both stops artificially increased the car’s time in the pit lane, as it had to wait at the pitlane exit for the next Safety Car to come past before being allowed to rejoin.

Although this caused the team to lose time, it did enable work to be done on the car when the others were travelling relatively slowly. In overall terms I calculate each stop lost the team around 15s, compared to a stop under normal conditions, but it also means that the Time in Pits figure in the table above should be reduced by about three minutes, in order for it to be comparable with everyone else.


Now, although it is a jolly good way of finding the ultimate pace of a car, the practice of taking the average of the best 20% of laps (in this case 79) obviously ignores the other 80% of the laps. A bit like an opinion poll draws conclusions based a sample, and is sometimes wide of the mark, so the average of the best 20% of laps can throw you off the scent.

If we ignore the laps into and out of the pits and all of the laps spent wholly or partially behind the Safety Car, and look at the all of the remaining laps for the first and second-placed Porsches, we can then split them up into the best 20%, second-best 20%, etc. This gives the results below:

Segment No. 17 Average Lap Time No. 19 Average Lap Time Delta
Best 20%
3m 20.4s 3m 19.9s 0.5s
21% – 40% 3m 21.7s 3m 21.0s 0.7s
41% – 60% 3m 22.7s 3m 21.9s 0.8s
61% – 80% 3m 24.0s 3m 23.2s 0.8s
Slowest 20% 3m 38.0s 3m 35.2s 2.8s


This shows that the fastest 80% of laps for these two cars follow a similar pattern: the Hülkenberg/Tandy/Bamber Porsche’s laps are between a half and one second (on average) quicker than the Bernhard/Webber/Hartley car. It is in that final (slowest) segment where a large difference lies, broadly speaking, these are the laps slower than 3m 25s. Although I have excluded laps spent behind the Safety Car, these laps do include those affected by Slow Zones.

Keeping that all in mind, let’s see if there’s another doll within our Babushka.

For many observers, it was Nick Tandy’s second stint, starting just after midnight and lasting for three hours, that established the advantage for the white Porsche; putting it in an impregnable position as the race entered its second half. Look at the table below, which compares the four cars that were occupying the leading four positions at the time, all of which were on the lead lap.

No. Car Driver From To No. of Laps Lap Time
19 Porsche Tandy 00:14 03:15 52 3m 25.89s
17 Porsche Bernhard 23:45 02:47 52 3m 26.75s
7 Audi Lotterer 00:25 03:26 52 3m 26.28s
9 Audi Bonanomi 00:49 03:51 52 3m 26.79s


Consider also that during this time there were slow zones at Mulsanne, Indianpolis and Arnage, that’s why the average lap time is slower, in force intermittently between 00:54 and 01:22 and again from 02:20 until 02:42.

At the end of his stint, Lotterer complained that the Audi was slow, and that this was why Tandy in the Porsche was able to extend the lead, and although these numbers confirm that, it is interesting that Lotterer was still quicker than either Bernhard in the other Porsche, or Bonanomi in the no. 9 Audi. Certainly, Audi made a change following André’s stint, which enabled Tréluyer to start closing in again, but the damage had already been done.

However, there was something else going on, in my view. The regulations allow for a certain amount of fuel energy to be used per lap. In the case of Porsche, running in the ‘up to 8MJ’ category, that meant ‘only’ 138MJ per lap of petrol energy could be utilised; the others were allowed to use more. Given that a portion of the lap was in a state of neutralisation, where cars had to reduce speed to just 80km/h, a whole chunk of energy, saved during the slow zone, would have been available to be used around the rest of the lap.

This, along with the phenomenal acceleration capability of the Porsche (especially as its batteries would have been fully charged from the heavy braking leading into the slow zone), would have enabled Tandy to make up significant ground on his pursuers: at that time the two Audis in the hands of Lotterer and Bonanomi. Further evidence for this is that despite running more slowly around part of the track, none of the leading runners was able to save enough fuel to enable them to cover an extra lap when a slow zone was in operation. The only time that extra laps were possible was when the Safety Car was out.


Just why Timo Bernhard wasn’t able to pull the same trick is something of a mystery, but maybe it’s necessary to look a layer deeper to find explanations for that.

Without question, the GILL ultrasonic fuel flow meters perform are not 100% uniform. There are variations depending on their temperature, and also on the temperature of the fuel whose flow rate is being measured. Throughout practice and qualifying, the fuel flow meters were being changed regularly, as teams sought to find the ‘best’ ones. Also, as the fuel load goes down, the temperature of the fuel goes up, and different meters react differently.

From what I was hearing in the paddock at Le Mans, the contract with Gill is up for renewal at the end of the year, and as likely as not, there will be an open market soon. The good thing is that there doesn’t seem to be any carping about the problem, but it does provide a possible explanation of why one Porsche was 0.5s per lap (0.25%) faster than the other two.

Paul Truswell