To add to the stockpile of data gleaned so far from the Japanese Grand Prix, we have the qualifying times, which is the first set of truly comparable laptime data. We also have some statements from Pirelli reported on the Autosport website that the options are good for 10-12 laps (so no two stoppers with more than one stint on options, unless something odd happens), and that the primes will probably do 20 laps.

So what we need to do in order to get an idea of how the strategy can play out is to get an idea of the comparison of pace between the options and the prime tyres. From the last couple of races, the pace difference is greater on the ‘golden lap’ than over a stint, but at Monza we had 0.7-1s in the race (1.5s in qualifying) and in Singapore 0.5s in the race (2s in qualifying). As we are on the soft and medium compound tyres (same as Monza), the best guess is that the pace difference will be around half that seen in qualifying.

The second piece of data to try to get from qualifying is to figure out who has new tyres left, and what compound they are. This might give some idea of what plans are afoot. I understand that the top teams ran used options (I watch the BBC) for the first Q3 attempt – I’ll assume that is true. Finally, I’ll have a go at making sense of the free practice long runs in the light of the qualifying performance in order to see whether any more light can be shed on the relative performance of the cars in race trim.

The gap between the prime and option looks to be about 1.5s – interestingly a little less for the Ferraris and the Renaults (1-1.2s). Trulli found 1s but as he was the only Lotus to do two runs, this could be due to a mistake. Kobayashi found a full 2 seconds – the Sauber is always different! So I think that we should be looking at around 0.7-0.8s pace advantage on the options as the best guess for the strategy prediction. The pace of each car in qualifying (based on average to Vettel across all laps), and the relative fuel load difference to Vettel (assuming the race pace is equivalent to the qualifying pace) are given below – there is a big margin of uncertainty in the differences between the race and qualifying performance, especially with the DRS. All drivers have 2 sets of primes and 1 sets of options left new unless stated.

**Driver: Gap Long run FP Fuel Load Tyres Left**

Vettel

Webber +0.2s +0.0s 0 laps

Hamilton -0.2s no time no time

Button -0.1s +1.7s 18 laps

Alonso +0.4s (less on primes) +1.6s 12 laps

Massa +0.4s +3.0s 26 laps

Schumacher +0.7s +2.2s 15 laps

Rosberg no time 3 option, 3 prime

Senna +1.4s +3.5s 21 laps

Petrov +1.4s (not consistent) +3.6s 22 laps

Barrichello +2.4s no time no time

Maldonado +2.4s no time no time

Sutil +1.9s +2.5s 6 laps

Di Resta +2.4s +3.4s 10 laps

Kobayashi +1.8s no time no time

Perez +1.9s no time no time 2 option, 2 prime

Buemi + 2.2s +4.2s 20 laps 2 option, 2 prime

Alguersuari +2.7s no time no time 2 option, 2 prime

Kovalainen +4.6s +4.1s -5 laps 2 option, 3 prime

Trulli + 4.7s no time no time 2 option, 2 prime

Ricciardo +7.0s +5.5s -15 laps 1 option, 3 prime

Liuzzi no time 3 option, 3 prime

Glock +5.7s +4.1s -16 laps 2 option, 2 prime

D’Ambrosio +5.6s +5.2s -4 laps 2 option, 2 prime

So the first thing to note is that either the Red Bulls will be very fast in the race, or they were testing on a light fuel load. If Massa’s times are consistent, this would suggest that the Red Bulls were on half the initial fuel load (53 lap race). This seems unlikely to me, so I think that the Red Bull pace in free practice is likely to be real, at least to some extent. The evidence points to the Red Bulls being the cars to beat in the race – I won’t be surprised to see Vettel disappear into the distance. This is supported by the fact that the tail-enders were quicker in free practice relative to the Red Bulls, suggesting that the Red Bulls were not the lightest cars on the long runs.

It looks like Renault are back in a comfortable 5th place, and that the battle between Sauber and Force India will continue – the intelligentF1 model has suggested that Sauber has been quicker in the races at the last two events. With the upgrades they have here, they should be looking to finish ahead of the Indian cars, and score some points. At the back, things are fairly consistent, except Ricciardo is slow. Did he run primes, rather than options, in qualifying?

So to strategy. What can be done with 2 sets of new primes and one set of new options, given a start on old options. The teams seem to be limited to two or more stops based on the lifetime of the option tyres, and 15 laps is what would be ideal for a three stop strategy. So, let’s start by assuming that the primes are 0.8s slower than the options. For the front runners, the blue dashed line represents three stops (option-option-option-prime), the red dashed line two stops (option-prime-prime) and the grey dashed line four stops (option-option-prime-prime). It does not matter which order the tyres are used – the answer is the same. So the intelligentF1 model gives the following chart:

Nominally then, three stops is better, with the two stopper beginning to feel the pace on the primes before the end. Four stops is 15s slower – unless there are real problems, it looks like this is to be avoided. This case has stints of 8, 8 and 10 laps on the options. With everyone having a new set of options, it should be expected that all the cars should be able to avoid this, but it depends how much qualifying takes out of the tyres. I woudn’t be surprised to see the first stops within the first eight laps, and cars struggling slightly at the end to eke out the prime tyres to the end of their three stop plan.

If the primes are slower, say the 1.5s seen in qualifying, then this works in favour of three stops over two or four due to the amount of time which is spent on the slower tyre. This is clear in the below figure. The four stop and two stop are close, but 25s behind the three stop strategy which maximises the time on the faster tyre.

So is it worth hanging on to the options when they are starting to go, in order to avoid going on to the primes? Well, it depends how much they degrade, but say the options degrade at 0.5s extra per lap past 6 laps (race) lifetime. Then we get something that looks like the figure below, where the three stop stays ahead, but the gap is cut to 5s. A bit less degradation, and three is OK, a bit more and it’s time to make the call between getting on the primes or going four stops. Note that this is still the case where the primes are 1.5s slower.

If the primes are only 0.8s slower, and we have the same degradation of the options after 6 laps, then we have a final sting in the tail. The chart below from the intelligentF1 model shows that two stops, with two stints on primes would be the way to go here. The strategists are going to be very busy in the first half dozen laps working out what is happening, and then will have about two minutes to make a call if the options are struggling. And hoping not to have to feed back into traffic. There could be some tough decisions to make, and some big errors, if the degradation of the options is as bad as feared.

Case study summary table:

Option tyre life Prime tyre pace 2 stops 3 stops 4 stops

1o laps +0.8s +7s optimum +15s

10 laps +1.5s +26s optimum +25s

6 laps then 0.5s/lap deg. +1.5s +6s optimum +5s

6 laps then 0.5s/lap deg. +0.8s optimum +10s +8s

Whatever happens, it is likely to be very interesting. History suggests that the bigger teams will cover each other – perhaps with four stops. Maybe this is a chance for those reputed to be soft on tyres to make a three stop work where others can’t. It is also likely that an extra set of new options would go that bit further – I would guess at Rosberg going that way, although perhaps option (to clear the tail)-prime-option-option to be out of sync with the cars he is trying to pass.

So much to anticipate. Although I am a little concerned that it might be another Turkey – not in number of pitstops, but that Vettel will be so quick that it doesn’t really matter what he does.

*Strategy Predictions, Underlying Pace Analysis*

raymondu999

October 8, 2011

Mate; raymondu999 from F1T forums here.

May I suggest including a table? That summarizes the overall race simulation time difference; based on the time delta between primes/options? So basically somewhere where we can just look and see; if options are X tenths faster, then the optimum strategy is Y stops; which is Z seconds faster.

intelligentf1

October 8, 2011

It’s a good suggestion. At the moment these are a set of case studies based on best guesses. I can summarise the case studies no problem, but you got me thinking that what might be helpful in watching the race might be something like a set of decision criteria.

So 2 stops if …

3 stops if …

4 stops if …

I can write a bit of code to do that, but it won’t happen before the next race.

raymondu999

October 8, 2011

Yeap that would be fantastic. Obviously safety cars are a wildcard and I don’t think hobbyists such as ourselves can ever write an algorithm complex enough to factor in safety cars or “Trulli Trains.”

Also here’s something I also PMed you in F1T;

Hey mate. Just been reading your latest blog post. I was wondering; about fuel loads; rather than comparing drivers’ long runs to one another; which would still leave us in the dark on fuel loads; would it not make sense to compare each driver’s Q3 lap vs their long run pace; perhaps as a percentage? i.e. Button’s long run average pace was x% of his Q3 lap time; and compare this value of “X” amongst the drivers?

Just something that I was thinking of.

So basically; you get to know how much of their true pace they were running in FP long run; as a percentage of Q3 pace; and then use that to compare between the teams. Surely those whose long run pace is at a higher % than others are running more fuel then.

intelligentf1

October 8, 2011

If we add fuel effect to Q3 time, we get something around 3s shy of the time difference, due to golden lap, DRS, tyre management, …

Once we get the race times, we can look at this effect, per car and see if it works as a constant. If so we might just be able to get the fuel loads out.