Barcelona Test 2: Day3 – Midfield to the Fore

Posted on March 4, 2012


The interest on day 3 of the Barcelona test may have been taken up with the goings-on at the rear end of the Red Bull RB8, but as far as the long run data is concerned, the real data is provided by the midfield teams, with Sauber, Force India, Williams and Toro Rosso doing mulitple stint runs, and there is a tantalising long run from Caterham which may show that they are also not too far from joining the party.

The key advantage of looking at the multiple stint runs, especially where the data supports a run without refuelling, is that the uncertainties associated with fuel load can be massively reduced. If a car has run for a race distance without refuelling, the uncertainty in the fuel load is between the amount is takes to do the race distance and the fuel tank capacity. For Singapore, the fuel required is about 165kg which is about as high as it gets, and for Barcelona the load is in the region of 150kg. Even if some allowance is made, this gives a maximum uncertainty of about 20kg (say) which is about 0.9s according to the intelligentF1 model (I think Gary Anderson said 1s in the Autosport article yesterday – so I’ll take that). So then, if we can be confident that the teams have not been refuelling between stints, and they have done a full race distance, then this is about the uncertainty we have. Chances are that the midfield teams are doing something more representative of a race, so the uncertainty could well be lower than this. From where we sit, however, we have to assume that they have all done pretty much the same thing, and keep in mind the possible level of uncertainty.

So, following the standard intelligentF1 practice of creating fictional in/out laptimes and constructing a representative race history chart, we get the chart below. I have included Vettel’s ‘race’ from the third day of the first Barcelona test as a ‘control’ and labelled the traces – Kovalainen is the green trace with only one stint. Massa and Grosjean’s traces from Friday are also included to give the fullest picture we have of the relative pace of the cars.

From this we see that the Sauber seems to be fast (which can be argued to correlate to the shorter run), Force India are quicker than Toro Rosso, and Williams are behind. If the one stint we see is representative of a first race stint (by no means certain), Caterham are in the same race as Williams. The traces of Massa (or Alonso) and Grosjean are just ahead of the Force India trace.Given the closeness of the traces, these look to be reasonably representative and provide an indication of what we might see. There are always caveats – Vettel did his run a long time ago, and on hard tyres.

The two teams which are conspicuous by their absence are McLaren and Mercedes, Neither have done a run without (as far as I can tell) refuelling, so it is not clear that their fuel loads are as high at the start as would be realistic. Hamilton’s run last week matched Vettel in the first stint, and both Rosberg and Button have run multiple stints which started faster than Vettel, but involved toying with the fuel loads. Both these teams are clearly at the quicker end of the pack, but where they fit (and indeed the pace at which Red Bull can really run) is not known.

I have performed fits on each of the traces using the intelligentF1 model, and the results are below. I will provide graphs of the fits for the new runs only.

Vettel: The first stint (on softs) is 0.6s faster than his next two stints (on mediums) but shows higher degradation. On mediums the degradation is around 0.1s per lap (compensating the fuel load) and on softs 0.15s per lap. He is a further 0.6s slower in later stints on the hard tyres, where the degradation is similar to the mediums.

Massa: The pace on softs is 0.6s from Vettel, but with degradation more like 0.22s per lap. The pace on the hards is about 0.9s slower (which is good in comparison with other teams), and the degradation is similar to Red Bull. Alonso is similar, but his pace on mediums is 0.6s down on the pace on softs, and the degradation on mediums is also higher than the Red Bull at about 0.18s per lap. Ferrari’s long runs do not look good on tyres.

Grosjean: The pace on softs is about the same as the Ferraris, and with similar degradation. On the hard tyres, the Lotus is 1.2s slower, but the degradation is more in-line with Red Bull than Ferrari on these tyres. If pushed, I would guess that the Ferrari is a faster car, but the Lotus is clearly towards the front of the midfield.

Di Resta: (soft-soft-medium-soft-medium) The fit for Di Resta is below. I get his pace on softs to be about 0.2s per lap slower than Lotus/Ferrari and the degradation to be slightly better at 0.18s per lap. On the mediums, the Force India is then 0.6s slower, with similar degradation. The interesting thing to note in Di Resta’s trace is that the last two stints are disproportionately slow, by 1.5s in both cases. This could be fuel (would have been quick) or could be testing of fuel saving modes.

Perez: (soft-medium-?-?) I have used mediums for the third and fourth stints as these fit the data better. What is interesting is that Sauber have stated that they are happy with the tyre wear and they were the standout kings of tyre preservation last year, but the fits of the testing data suggest that they have higher degradation than most. The pace on softs is about 0.7s faster than Vettel (high 0.25s per lap degradation) and 0.6s slower on the mediums (same as for most people, but with worse degradation than Ferrari at 0.22s per lap. As I don’t believe that Sauber are quite this quick, and the simulation was short, the best guess is that they started with a fuel load representative of a second stint. This would cost something of the order of 1-1.5s depending on how long they imagine the first stint to be. If I add this on to their times I get them at best just ahead of Ferrari/Lotus, and at worst level with Force India, but with worse tyre degradation. Either way, they seem to be in a pretty good place. The trace + fit is below.

Ricciardo: (?-soft-soft-soft) The data for the first two stints is OK, so I think that the results for the soft tyres are OK, but the analysis for the last two stints is not so good – similarly to the Force India, the pace is 1s slower for the final two stints. The first stint seems to be on much newer tyres than the second stint as it lasts much longer, and with lower degradation. It seems that the Toro Rosso has the lowest degradation on the softs at about 0.1s per lap, although the stints 2, 3 and 4 show more like 0.15s per lap. The pace? About 0.5s down on the Force India. The trace and fit are below.

Senna: (soft-soft-medium-?) I have also used mediums to simulate Senna’s mystery compound as it fits the data better. The Williams is a little adrift in this simulation at 1.7s behind Vettel (on softs), which is 0.4s off the Toro Rosso. The situation is worsened by the 0.22s per lap degradation which is among the worst. The situation on mediums is even worse as the pace difference is 0.9s (compared to 0.6s for most teams) and the degradation is similar to that seen on the softs. Based on this race simulation, Williams seem to be off the back of the midfield pack, much like last season. Trace + fit again provided below.

Kovalainen: His one long stint on unmarked tyres is about 0.1s per lap slower than the Toro Rosso and 0.3s faster than the Williams. If they are on race fuel, then they look like they have just about made the midfield. I think Kovalainen suggested they needed another 0.3s, and on this evidence I would think that they will be thereabouts.

So, in conclusion, the intelligentF1 interpretation of these race simulations gives us something like:

Red Bull  +0.0s (but almost certainly with more to come)

McLaren  unknown

Ferrari +0.6s

Mercedes unknown

Lotus +0.6s

Force India +0.8s

Sauber +0.3s to +0.8s depending on fuel level

Toro Rosso +1.3s

Williams +1.7s

Caterham +1.4s but higher uncertainty on fuel level.

I have read suggestions that the field is closer this year, with perhaps 2s covering most of it. That’s pretty much what I have…