Canadian Grand Prix 2012: Gary Anderson’s technical review
Lewis Hamilton's performance in winning the Canadian Grand Prix was exceptional.
It was a great drive - Hamilton did exactly what he had to do when he had to do it and still won despite a couple of pit stops that were not quite as quick as they should have been.
But his McLaren team-mate Jenson Button was never in the race - just as he hadn't been in Monaco or in Spain.
McLaren really have to look at why that is. It's a very strange situation to have one of your drivers in the quickest car and the other, also a world champion, struggling so much.
Hamilton looked strong throughout the race, but Ferrari did have the chance to beat him with Fernando Alonso had they been a bit sharper. And they should at least have been second.
To me, it always looked as if the best strategy was going to be stopping on lap 15-20 and then on lap 50. McLaren planned for that and Hamilton drove to whatever lap time they thought was a fast race.
Ferrari had already heard Hamilton saying on the team radio in his second stint that his rear tyres were going away - that should have been the first signal Alonso would not make it to the end on that set either.
If Ferrari had reacted to that and stopped Alonso a lap or two before Hamilton, they might have had a chance of beating him.
Even after that, they had a number of chances to hold on to second place.
Firstly, they could have reacted immediately to McLaren's stop.
Secondly, two laps after Hamilton's stop you could already see Alonso was in trouble. There were 18 laps to go - he was a second-a-lap slower than Hamilton and he was only 14 seconds ahead.
Ferrari might have thought Sebastian Vettel's Red Bull would block Hamilton for a while - but that didn't happen.
Finally, if you look at lap 58, 12 laps from the end, Alonso's times started to drop off by nearly a second.
If you take the lap times Alonso should have been able to do on a fresh set of tyres and compare them with what he actually did, he lost 25 seconds in those 12 laps.
A pit stop is 14 seconds, so that means Alonso would have been 11 seconds further up the road at the end had he stopped. Which would have had him battling with Romain Grosjean's Lotus for second place.
The teams have computer programmes that take a driver's lap time and predict where he is going to be by the end of the race. They constantly update throughout the race. And that's not even counting the blokes on the pit wall making the decision.
Vettel was a slightly different case. His lap times in the middle of the race, from lap 40 to 60, weren't quick enough. He was clearly trying to get to the end by nursing the tyres.
Unlike Ferrari, with the messages Red Bull were getting from their lap times, they couldn't have done anything different from what they did.
The only question is whether Vettel was driving that slowly because he was saving fuel - I suspect everybody thought there was going to be a safety car, which gives a few laps of grace, but it never happened.
Unlike Hamilton, Alonso and Vettel, Grosjean and Sauber's Sergio Perez were able to make a one-stop strategy work.
The leading cars are that bit quicker and that puts that bit more energy into the tyres and using them up faster, whereas the slower cars put less energy into the tyres so they last longer.
It's an unusual situation - quite often in the past, with other manufacturers' tyres, the faster cars with more downforce have looked after them better.
But the Pirellis don't work like that. If you put energy into them they simply wear out faster.
Despite their strategy errors, Ferrari were fighting for victory in a dry race on a normal circuit. They have come on in leaps and bounds since the start of the season.
In Australia at the start of the season, they were more than a second a lap slower than McLaren; in Canada, they were more or less the same pace. So Ferrari have definitely improved their car a lot.
They introduced their first major upgrade package of the season in Spain two races ago. Canada saw another - and they made another big step forward.
Ferrari have made some very visible aerodynamic changes - and I'm sure there are some mechanical modifications we have not seen that are adding to the package.
In Spain, they temporarily abandoned their attempts to use the exhausts for aerodynamic effect, adopting a more benign exhaust lay-out that allowed them to make better use of the important 'coke-bottle' area at the rear of the car.
In my opinion, compromises in both the shape of the rear bodywork and the radiator air exits meant they had exploited only about 50% of the possibilities of that approach. Making it better again, by getting more of the disruptive bodywork and low-speed air out of that area, would have been a good solution.
Instead, they have gone for what is effectively a copy of McLaren's approach to the rear end - one that Ferrari tried in pre-season testing only to find they were over-heating their rear tyres.
This has more bodywork sticking out into the coke-bottle to house the exhaust exits, which are directed at the area where the rear tyres meet the floor.
For a circuit such as Montreal where traction is critical, this is a good solution, as it allows the team to blow exhaust gases on to the aerodynamically shaped brake ducts to increase downforce on the rear tyres.
But there is still a lot of blockage in the coke-bottle area, so this might not be such a good solution at a track such as Silverstone - which hosts the British Grand Prix next month. Aerodynamic efficiency is critical there.
To that end, the solution of Red Bull and Sauber, who also blow the exhausts in that area but have much less bodywork obstruction in the coke-bottle, is probably the best compromise.
Another neat feature on the Ferrari in Canada were new front brake ducts which more or less eliminated the hole that takes in the air.
They simply had a shield in front of the tyre and the air is taken into the brakes between the tyre and the turning vane inside the wheel, through the gap you need because of tyre deflection.
The shield acts as a wheel faring, which solves the problem of the tyre disrupting the airflow by keeping it attached to that surface and then feeding it into the low-pressure area behind the tyre.
That means you get both a drag-reduction effect and it stops the turbulent air getting into the underfloor.
That's a good, positive step - and if you can run it in Montreal, where you need the best brake cooling of the season, you can run it anywhere.