Singapore In Depth Data Analysis

Posted on October 2, 2011

0


Using the intelligentF1 model of the race, the race data can be analysed to assess the performance of each car/driver in an objective manner. All data analysis is subject to some level of judgement, but the data analysis here is based primarily on the data, and not from input from team personnel. There are places where input from the teams can assist in the understanding, but it does not determine what is considered important in the intelligentF1 data analysis.

The Singapore Grand Prix was a complex race, and is thus a rich source of data for this type of in-depth analysis. In contrast to the previous race in Monza, the key strategic consideration was the lifetime of the tyres, especially the option (super-soft) tyres, rather than the relative pace of the two tyre types. If the tyres had lasted as long as desirable, then a two-stop race would have been the fastest strategy, with a one stop competitive in the case that the prime (soft) tyre was almost as fast as the option (which was the case in reality). A three stop would be competitive in the case that the prime tyre was slow. As most drivers stopped three times, and the prime was competitive (about 0.5s slower per lap on average), it shows that the tyres were not lasting. The option tyres were lasting about 10 laps before degradation became significant, and the primes about 20 laps.

The expectation before the race from intelligentF1 was that most teams at the front would do 3 stops (option-option-option-prime) as done in the race by Mercedes, and that midfield teams would try to go with 2 stops (option-prime-prime or prime-option-prime). Pirelli suggested that the option was 0.8-1s faster than the prime, and the golden lap in qualifying was a full 2s faster.

The opening laps showed that the option tyres were not lasting long,  Ferrari and Mercedes stopping within 10 laps. The opening few laps are shown in the Race History Chart zoom below.

Added to the chart in dashed lines are the intelligentF1 race model for Sutil and Di Resta. The stint pace of Di Resta was about 0.4s per lap slower according to the fit, but the prime tyre was lasting well. Di Resta lost time early in the stint due to the early pace of the option tyre, but as the race settled down, it quickly became evident that that the prime was going to be an effective race tyre. The Williams (blue lines) held on for a few laps, but were rapidly dropped once the intial grip from the options went away. This stint from Di Resta defined the race – most of the teams reacted to this by fitting primes for the second stint, and again when the safety car was deployed.

The pace of the cars of the first stints is shown by the fits in the race history chart below.

Relative to Vettel, Button is 0.9s slower, Alonso 1.2s, Rosberg 2.5s, the Force Indias 2.8s, the Williams 3.5s, Senna 4s and Trulli 5s. The first cars to feel the tyre wear on the options were the Mercedes (normal) and the Ferraris (less expected). The issue for the Ferraris is that the limiting factor on tyre wear is traction from the acceleration out of slow corners as opposed to the lateral loads from high speed corners at a circuit like Barcelona or Istanbul Park. The Ferrari has serious traction, and this contributed to the quick wear of the tyres. Neither Ferrari was to use the option tyres again.

It was clear at this stage that Renault were slow, and that Perez was looking strong – and he was only just losing performance on the options after 10 laps. The second stint is shown in the race history chart below. There is a lot of data on this chart, so it needs careful analysis.

The first thing to note is that Alonso, Button and Vettel are slightly slower than would be predicted from the first stint – likely due to fuel saving. Rosberg is right on the pace shown in the first stint, but is 0.5s slower than Schumacher, who was showing some very strong race pace. Perez was quick on primes – Sauber show the smallest difference in pace between the prime and the option tyres, and was quicker than the Force Indias in this stint. Indeed, were he able to pass, he had the pace to pull away – Sauber were faster than Force India in the race. Interestingly, the safety car was bad news for Mercedes and Ferrari as they were the teams who had already made their second stops – they lost out most. Hamilton gained, but not by as much as it seemed at the time as he was within 16s of the Mercedes/Perez/Force India battle ahead of him and was 1.5-2s quicker in real pace. He would most likely have finished fifth anyway. Toro Rosso were able to show that their pace was about that of the Williams with clear air for the first time, and Renault’s worst fears were realised. Also interesting is the pace of Webber, who is a full 1.2s slower than Vettel in this stint – only part of this can be explained by the time spent behind Alonso’s Ferrari.

After the safety car, the strategy became a choice of one stop or two to the finish. Most of the teams fitted primes (Alguersuari, Kobayashi, Liuzzi and  Hamilton the exceptions) and hoped not to stop again. The battles for the podium and for sixth covered one another, but with differing solutions to the problem. The intelligentF1 analysis makes it clearer why these choices were taken. The chart in the first stint after the safety car is shown below.

Vettel is clearly right back on the first stint pace, but this time the challenge is more serious with Button only 0.3s slower, and Webber about 0.1s slower after the initial field spread due to the lapped traffic. Alonso’s pace is 1s down. Di Resta’s pace is consistent with the previous stints, which suggests that Rosberg had about 0.3s in hand (although it is impossible to say how much pace the Mercedes had on primes as it never had clear air) and Perez about 0.2s based on his pace in the second stint. The strategy dice was rolled early by Massa having been stuck behind Perez, which shaped the strategy of the battle for sixth, and by Webber at the front which shaped the strategy at the front. Hamilton was committed to stop due to choosing option tyres under the safety car, but always had the pace to come fifth. The intelligentF1 model can predict the pace of the cars based on making the opposite call in reaction to the stops of Massa and Webber – this is done in the chart below.

The intelligentF1 model traces (dashed lines) of Vettel  and Button assume not stopping again and a tyre degradation similar to that of Di Resta. The model for Di Resta is also shown, and the increasing tyre performance loss is modelled well enough to be able to assess the impact. Vettel had about 0.5s pace in hand in the final stint, and would have been pushed had Webber shown the pace of the third stint – given the fast first laps, it looks like Webber was bringing it home. So, if he pushed, he would have been in a position to catch a non-stopping Button a couple of laps from the end – so McLaren were right to cover the Red Bull stop. Vettel had pace in hand, and would probably have won either way, but would have been much worried at being caught on the wearing primes if he had not stopped. Hamilton showed that there would be little resistance had the freshly tyred cars caught those on very tired primes in the final laps.

The battle for sixth is even more interesting. Di Resta just had to keep going as long as no-one else stopped. However, Massa had the pace to catch him, but blew in in not being able to pass Barrichello for three laps. Massa, ironically, was the major reason for those ahead not to stop, as they would have come out behind him – they may have had fresher tyres, but the Ferrari was a much (2s) faster car. The intelligentF1 model for Perez is shown, and he would have had the pace to catch Di Resta (and indeed Massa). This was a no-lose strategy for Perez, and it would have been interesting to see how it would have played out . It was Massa, however, who had the most to gain by doing something different. Had Massa saved his stop until lap 49, he would have had a larger pace advantage and could have gone onto options worth another 0.5s for the final laps. Assuming he remained stuck behind Perez to this point, his final laps could have looked like the chart below.

Matching Perez position on Lap 49 (assuming Massa is right behind) and stoppint for prime gives Massa the sort of pace advantage over the Force Indias/Rosberg/Perez that Hamilton enjoyed. With this pace advantage, it seems unlikely that he would have been stuck behind Barrichello, and he must have had a very good chance of leapfrogging the Rosberg/Sutil/Perez train, and may well have caught Di Resta in the very closing laps. Had Massa stopped at this point, he would probably have undercut other stoppers, and so would have gained substantially. It looks like Ferrari got this one wrong – sixth place was on.

A few other notes are worth making. Alguersuari attempted to finish the race on options from the safety car, and started dramatically losing pace with 10 laps to go. He stayed out, and crashed – probably due to having no grip left and Senna having caught him. Indeed Senna went past both Kovalainen and his team mate after the safety car, but Petrov was unable to pass the Lotus, and eventually gave up after failing to undercut and then getting stuck behind D’Ambrosio despite the data suggesting that he had about a 0.6s advantage over the Lotus in clear air. This can be seen in the chart below. The pace of the Virgin is about 1.7s slower than the Lotus.

So, a fascinating race from the data. Vettel again with much in hand, but with strategy able to make a big difference in the closing laps. IntelligentF1 will be back for an in-depth analysis of the Japanese Grand Prix in Suzuka.

Advertisements