Here’s what Lewis Hamilton said after finishing second to his team-mate last time out in Brazil: “I’m here to race. It’s so difficult to overtake here and when you both have to do pretty much the same strategy it’s kind of already set from the beginning. So I’m like, ‘if there are other strategies, let’s do it, let’s take the risk,’ and they’re like, ‘look after the tyres,’ and I’m like, ‘no, I’m racing.’ And I think that’s what people want to see… It would be great sometimes to be able to do something different, see how it plays out. They do so many strategy simulations, pick the best two and that’s what you’re stuck with.”

Here’s what Mercedes team boss Toto Wolff said about the idea of the drivers calling their own strategy. “They really wouldn’t want to do it. I guarantee that they would lose the race pretty much every time.”

Yes, if only one of the drivers was calling it and the other was being guided by the team. That’s on the assumption that we’re talking here of a race like Brazil where no other team had a look in, where it was always only a question of which Mercedes driver was going to win. But what if – in that situation – both were allowed to call it as it unfolded? What if Mercedes, should it find itself in a dominant position again this weekend in Abu Dhabi, let them at it?

What’s the risk? The constructors’ championship has long been won, the drivers have now secured first and second place in the championship. The engines and gearboxes can be used to the max as this is the season’s final race. There’s a challenge.

When Wolff talks of how the drivers would call it wrong every time, he’s talking of the theoretical ideal. That ideal changes between pre-race and during the race as different factors come into play – tyre degradation may be slightly different to that seen in the practices, the track temperature can be different, the pattern of the traffic and where the gaps are, etc. These things cannot be exactly modelled. Keeping track of all that to define the exact right time to come in would indeed be close to impossible for the driver.

But they don’t need to get it that right.

So long as the Mercs have the sort of advantage they’ve enjoyed at many races over the last two seasons and they get away from the grid in 1-2 formation and they pull out enough of a gap over third place, it wouldn’t much matter if their calls were not theoretically perfect relative to the non-Mercs; they should still finish 1-2. The intrigue would be in seeing a driver having to think it through while racing in order to beat his team-mate.

Here’s how those two ideal strategies Lewis was referring to are usually worked out. The difference between the two tyres is tested during practice. There are three critical pieces of data here:

1) The lap time difference with the two compounds at their peak
2) The performance drop-off of each tyre type
3) The number of laps it’s possible for each tyre type to do before it runs out of tread

These all vary according to the track surface, temperature, etc., but a fairly clear pattern can usually be built up from the practices (weather permitting). But that’s just the start of it.

Does the faster tyre over a lap (usually the option tyre) eventually become the slower tyre? And if so, after how many laps? Which, therefore – option or prime – is actually the faster tyre over a stint? Does the combined durability of the two tyre types make it possible to stop only once? Or do you run out of tread too early to make that possible?

And if not, does one or other of the tyre types lose enough performance through heat degradation that even though it has enough tread to keep going, it’s actually slower to do so, and therefore a two-stop is faster? Or even a three-stop? Is it faster to do a two-stop with a bit of lift-and-coast to keep the tyres in shape than it is to do a three-stop running flat-out all the way? If so, how much lift and coast is the minimum necessary?

Then, once it’s decided, for example, it’s a two-stop, is it faster to go option/prime/prime? Or option/option/prime? Or option/prime/option?

Other data established during the practices is the fuel consumption and brake wear. How does this fit in with the tyre patterns? If you are having to do some lift-and-coast anyway to make, say, the two-stop faster than a three, then you’re going to be saving fuel (and brakes). So, is it worth reducing the starting fuel load? And how does that, in turn, impact upon the tyre wear/degradation pattern? If you’re not going to lift-and-coast at all and run the race flat out, how much fuel will be required? Still under the 100kg limit?

How much, if at all, should we include the likelihood of safety car laps in the fuel calculation? After all, if we fuel too light and there are no safety cars we can always bring it back on schedule with lower engine modes. What’s the performance/risk trade-off there? Does the brake wear demand lift-and-coast? Is that the defining limitation – and do you fit in the tyre and fuel strategy around that?

What is the optimal time to use the more aggressive engine modes available? What should the mode mix be throughout the various phases of the race – within the parameters set by the engine guys? How difficult is it to pass slower or lapped traffic? When we build that into the model, is the theoretically faster strategy of multiple stops actually slower in reality – because of the greater amount of time spent in traffic?

So, all that is calculated and the team comes up with a plan. It’s a much more straightforward plan when you’re running from the front row of the grid, but still it has all those calculations and variables within it. One plan will have presented itself as the best – and that’s what both drivers will be on. Except of course they can’t both pit at the same time. So there will be a one-lap offset.

The driver in second place will have that tiny – but often crucial – strategic disadvantage to the guy leading, who gets first call. That usually means he comes in first (the undercut). But not always. If the tyres they are each switching to take more than a lap to come up to full temperature, sometimes it’s a disadvantage to pit first (the overcut). When are the pitstop windows? When are the safety car windows – ie. from which lap is it advantageous to pit? Some of it of course is game theory – the correct answer depends on what the others do.

None of this is left to chance. Pre-race, every team has hard numbers on all of this and makes its choices accordingly. But you know going in that those numbers are going to change. Because reality contains infinitesimal ways of screwing with the plan. So, again, those changes are monitored, everything constantly re-calculated according to what is unfolding out there in the real world. The team – dozens of clever, focused, competitive people – have access to all this data in real time. And it is the picture that they build with this stream of data that informs the driver, through the medium of his race engineer.

So of course he can’t possibly be expected to have the full picture. And if he is competing against a driver in the same team being guided while he is relying on instinct and judgement, he will lose most times. But even then, not always. And if both go off-line?

Sometimes the numbers in the model haven’t accounted for what a real, live racing driver can do. He may intuitively know that if he can just stretch the elastic, run into debit on his resources, he will gain track position and that he can then monitor that debt back into credit once he’s already in front. He may feel he can run a similar pace to the guy ahead while taking less from the tyres – and therefore destroy the other guy’s strategy by having way more pace around the pitstops than expected.

The messages the tyres are giving him through the steering wheel, the way the car is moving around him, the way, say, the front end has got more bite than in practice is allowing him to use a more fuel-efficient harvesting mode, or the way he’s quite relaxed about the entry oversteer the worn rear tyres are giving him makes him know that he can maintain a good enough lap time to make an unusually spaced pitstop or tyre choice strategy work.

All these things are imponderables to the software, but gut instinct to the driver. Plus, the superior strategy on paper may be so by the margin of five seconds over a race distance. What if the driver is totally confident he can make up 15 seconds over a race distance by superior skill and judgement? Then that five seconds doesn’t really matter. Of course, he could be totally wrong. And over a sequence of races if the two drivers are closely enough matched, the software will beat the driver. But every now and again it will lose.

And that’s not something that the strategists and engineers would even dream about. So, Mercedes: why not let them loose in Abu Dhabi? Just for this one race.