The strategy blunder that sums up Ferrari's crisis

Heads have not yet rolled in public at Ferrari, yet assuredly they will. But as we digest news of the latest reshuffle in the technical department this past week, let's not kid ourselves that the malaise is confined to the car itself and those who produce it.

Ferrari's performance at the Hungarian Grand Prix last weekend was a prime example of a team in crisis trying to extract a result through alchemy rather than science, and in doing so making terrible decisions. Charles Leclerc's fortuitous second place in the season opener now appears even more of an outlier.

Red Bull's balance problems in Hungary enabled both Ferrari drivers to qualify higher than they might have expected. Then, at the end of the first lap, Sebastian Vettel and Leclerc were running fourth and fifth, each having gained a position at the start.

But it was already abundantly clear that the track, soaked by a pre-race downpour, was drying rapidly given the mild-to-warm ambient temperature of 22C - indeed, both Haas cars had pitted to get off the intermediate rubber at the end of the formation lap (for which both would later be penalised).

PLUS: Where common sense faltered to F1's letter of the law

Putting all its faith in predictions that more rain was due shortly, and therefore seeking to get as many quick laps in as possible before swapping back to wet-weather rubber, Ferrari pitted Leclerc for soft tyres at the end of lap two and signalled Vettel to do likewise next time around.

If Ferrari thought this might yield some advantage when fresh rain fell - perhaps gaining track position during short-order pitstops for wet-weather rubber - it was dead wrong. Vettel saw this calamitous decision for what it was, and vetoed it on the grounds that the track was still drying. By doing so, and going onto medium Pirellis instead, he was able to salvage a sixth-place finish as Leclerc laboured to 11th, a lap down.

It's an established strategic wisdom in Formula 1 that you make decisions based on the data to hand - what's happened in the past and what's happening right now - rather than what might happen. Teams can and do roll the dice, but the most successful ones do so with a firm grasp on probabilities, based on that data.

We often talk about strategic gambles in F1 as "inspired" - when they pay off. This is an example of the historian's fallacy, where those analysing a decision fall into the trap of assuming that the person who made that decision had access to all the information the analysts now have to hand.

Ferrari arguably compounded its error by keeping Leclerc out, in the increasingly forlorn hope that the rain would come, even as he demanded that the failing strategy be abandoned

Take the 2007 European GP, and the much-lauded decision by the Spyker team to place Markus Winkelhock on full wet tyres at the end of the formation lap in anticipation of rain. Had that rain not arrived - by the bucketload, as it turned out - Winkelhock would not have had his moment in the sun, as it were, enjoying a brief period of being classified in the lead as chaos elsewhere resulted in a red flag.

It was a gamble, pure and simple - and by a team with nothing to lose. The car was based around a 'tub' at least two seasons old, and had been qualified solidly at the back of the grid by a rookie driver who had never seen rain in an F1 car before.

Neither had Winkelhock driven the Spkyer before that weekend. Starting from the pitlane in a car which qualified dead last is a different order of risk to starting there in a frontrunner.

What the rose-tinted spectacle brigade also forget is that Spyker then gambled on more rain, leaving Winkelhock out on full wets long after it was obvious intermediates were the right rubber for the job. Was this a bad decision following on the heels of a good one? Assuredly not: they were both equally bad, being gambles on what might happen.

In Hungary, Ferrari arguably compounded its error by keeping Leclerc out, in the increasingly forlorn hope that the rain would come, even as he demanded that the failing strategy be abandoned.

He was not the only driver to run on softs in that phase of the race - George Russell, Antonio Giovinazzi and Daniil Kvyat did too - but those drivers had either started near the back of the grid (Kvyat and Giovinazzi) or dropped back from a midfield qualifying position (Russell), so they had less to lose by taking on more risk.

Also, Leclerc was one of the first to go onto softs and the last but one to get off them. This is a classic case of the sunk cost fallacy, in which a person (or team) persists with a poor decision because they've already invested in it and cannot bear to write off the 'cost'.

Leclerc was a net sixth after the initial pitstop phase shook out but couldn't even summon the speed of the Haas cars that were now ahead. Before the race was 10 laps old he was lamenting over the radio the wisdom of going onto softs. And yet Ferrari remained on rainwatch until lap 20, by which time the unfancied rubber was exhausted and Leclerc had dropped six places, only one of which he would recoup before the chequered flag.

All F1 teams run pre-race strategic simulations to varying degrees of thoroughness depending on their resources. These might begin with an almost infinite number of permutations, but the range of options and potential outcomes narrow during the simulation, enabling the strategists to arrive at a small selection of theoretically optimum choices.

But rain naturally introduces an element of randomness into even the most carefully researched scenario. Any competent F1 strategist will tell you that when they review a race, they subject positive outcomes to the same degree of scrutiny as the negatives. They might also add that they view the weather radar's prognostications as being of the same order of reliability as those you might source from a crystal ball.

In reviewing a race, the best F1 strategists look at the outcome dispassionately, instead judging the quality of the decisions based on the information available at the time. Viewed through these optics, a poor decision doesn't become a good one simply because luck and randomness caused a positive result to eventuate. It's still evidence of a systemic failure.

By the same token, for instance, aerodynamicists are just as alarmed by new components which generate unexpectedly good performance as they are by new parts which underdeliver. Both are evidence of bad correlation, and represent warning signs to be acted upon.

This isn't the first time Ferrari has gambled and lost on weather conditions. Vettel had already done much to torpedo his own championship ambitions in the second half of the 2018 season when Ferrari arguably adminsistered the coup de grace by sending its drivers out on intermediate tyres during Q3 at Suzuka, figuring that the rain which had disrupted Q2 would return. It didn't, leaving Vettel chasing a result from ninth on the grid.

Last year's Mexican Grand Prix was decided by an undeniably ballsy call by Mercedes that required Hamilton to be at his brilliant best, but it was underpinned by solid data

Mercedes is generally held to be a risk-averse team, and even then it has made a number of high-profile blunders in recent years (for example, the 2018 Australian Grand Prix, when it miscalculated the safety car delta). But it has been known to embark upon risky strategies, particularly when trying to salvage a result from unpromising circumstances.

Last year's Mexican Grand Prix is a powerful example of Mercedes' oft-maligned strategists precisely calculating risk versus reward by using all the data to hand. Lewis Hamilton and Valtteri Bottas qualified third and sixth, but ran fifth and seventh after an opening-lap contretemps in which Hamilton damaged his car's floor in contact with Max Verstappen. Up front, polesitter Leclerc led from Vettel, and both Ferraris settled into tyre-management mode in a race that everyone had expected to be a two-stopper.

What had led to this certainty? Friday's running established that many teams had got their tyre allocations wrong, favouring the softer end of the rubber spectrum. The soft tyres were lasting but a handful of laps and the mediums not much longer. Of the top three teams only Mercedes completed any meaningful running on Pirelli's usually unfavoured hard compound.

Mercedes' initial struggles persuaded Ferrari that the third-placed Red Bull of Alex Albon was its greatest threat, so when Albon pitted unexpectedly early it brought Leclerc in for new mediums too, on lap 15. Vettel, reporting his tyres were good, opted to extend his stint.

Hamilton hadn't been able to pass Albon and even now, running second behind Vettel, he didn't appear to have the legs to overtake - then Mercedes pitted him for hard tyres at the end of lap 23. This wasn't checkmate, but it was the preface to it.

Ferrari, thinking Mercedes had been planning a one-stop but had now bailed out on it - for surely even Hamilton couldn't make the hards last another 47 laps - left Vettel out until the end of lap 37, enabling Hamilton to undercut him royally.

As Hamilton kept his tyres alive, against even his own expectations, Leclerc and Albon had to stop again, eliminating them from contention for the win as Hamilton cruised home ahead of Vettel in the final laps.

It was an undeniably ballsy call by Mercedes, and it required Hamilton to be at his brilliant best, but it was underpinned by solid data: 21 laps of running on the hard tyres by Bottas on Friday, followed by keen observation of laptimes being set by Daniel Ricciardo (who started outside the top 10, on hards) and Verstappen (who pitted early for hards after a puncture). All these data sets pointed towards the hards being durable enough and capable of setting competitive laptimes throughout.

This is the difference between a serial championship-winning team and a former rival sliding ever further away from credibility and competitiveness: the one making sure-footed decisions based on knowns, the other throwing ideas at the wall to see what sticks.