Why F1 teams are wrong about playing catch-up

Usain Bolt is a slow starter. It's all relative of course, as he's still fast enough out of the blocks to outsprint anyone over 100 or 200 metres, but he doesn't have to be ahead from the start to win spectacularly.

Some will have you believe the same is true of a grand prix season. A bad Australian Grand Prix? Don't worry about it, it's an unusual track, strange things happen. A bad first four races of the season? No matter - wait until the Spanish Grand Prix and all the upgrades change the game.

Already there is talk from some teams that even if they are not on the pace this weekend they can turn things around when they really get going later in the year. But does the game ever really change? And how powerful a predictor of what is to follow will 2017's season opener at Albert Park be?

Well, for the last three years, the winning constructor in Australia has kept winning. And in two out of those three years, the winning driver has gone on to take the title. But there's far more to the competitive order than just who wins, and it's possible to dig a little more deeply into the predictive power of the first race.

And, of course, a sample set of one race out of up to 21 races is desperately limited and you might as well look at any single race in this context, so if you are going to test the orthodoxy that things don't start as they mean to go on, it's sensible to look at the block of four races at the start of the year as well.

The past eight seasons is the logical cutoff. This is because 2009 was the year in-season testing was banned. Although strictly limited testing has subsequently returned, it's nothing like the magnitude that existed even in 2008, when mileage limits were already in play. For example, in-season testing in 2008 at group tests stretched to a smidgen over 100,000km. Last year's number was a fifth of that. And around half of that was using rookie drivers.

It also makes sense to look at constructor performance more than driver performance, as the two-car sample set at each race makes it more reliable.

During that period, the only time a team wasn't top of the constructors' championship after four races, but went on to win it was in 2010. That year, Red Bull threw away a couple of victories in Bahrain and Australia, so it sat third behind McLaren and Ferrari after four races, but it was emphatically the quickest team.

Drifting outside of the period in question, the last time before that was 2007 when McLaren was top after four races and only didn't win the constructors' title thanks to it being excluded for the spy scandal. Prior to that, 2003 when McLaren led but lost the championship as it campaigned the venerable MP4-17D throughout the year while waiting for the MP4-18 that never came.

And before that? It was 1994 when Williams had a tragic start to the season with a car that didn't work well and did manage to overhaul Benetton - but even then aided by disqualifications and exclusions for Michael Schumacher! And before that, 1987 when McLaren was ahead but Williams had lost Nelson Piquet for round two at Imola. You get the picture...

Going back to 2009-2016, there are several factors we can look at. First, a look at constructors' championship position. Taking the first race only over that period, on average the difference between constructors' championship position after the first race compared to the end of the season is 1.795. So that's not unreasonable, it means that, on average, there's a little under two places gained or lost for each team over a year. That would tend to suggest that there's a reasonable amount of flexibility.

But that is somewhat skewed by the fact it's a single data point, and it should be noted that that if you use the mode average (the figure that most often occurs) rather than the mean, the most common move is just one place.

Apply that methodology to the block of the first four races and the mean drops to 0.864 places - so on average less than one placed move. And the mode is 0, meaning that the most common outcome is that where you are in the constructors' championship after four races is where you finish. So much for Spanish GP upgrades changing the game.

Championship position is a useful little starting point, but the more interesting thing to look at is the pace of the cars. So it's interesting to look at what Autosport calls the 'supertimes'.

These are generated by taking the quickest single lap overall of the weekend for each team, expressed as a percentage of the single fastest lap.

This makes it possible to normalise for circuit length (so for a 21-race season like last year, each race weekend would contribute exactly a twenty-first of the data used to generate their performance figure). This gives you a performance figure for each car for each race, which can be averaged out over the season.

Comparing race one supertime to full-season averages, the variation is up to 0.672%. So, for the sake of argument let's say the pole time this weekend is a 1m25.000s - that equates to a gain or loss of 0.571s per lap in relative terms potentially. That is not insubstantial and can make a big difference to the competitive order but judge it over the first four races and that number plummets to 0.267% - 0.227s per lap.

Daniel Ricciardo suggested Red Bull was within half-a-second of Mercedes based on testing. Well, even if Mercedes has a disastrous development programme and slips back by the average, and Red Bull gains by the average, that still doesn't make up for the gap if you're at the upper end of within 0.5s.

It's a crude and simplistic way of looking at it, but when you bear in mind that the figures tend to be skewed by a few outliers who make big gains, usually those lower down the field, it tells a story.

This is not meant to be anything like an exhaustive mathematical investigation, as there are many more depths to the figures that can be explored beyond this surface scratch, but those figures do give a pointer to how things usually go.

The first point is that, when there's an unusual result in Australia, it's fairly obvious. The last time the winning team didn't go on to win the constructors' championship was when Kimi Raikkonen won for Lotus in 2013. At the time, that was obviously an unusual result driven by superior tyre management, and talk of title pushes was always predicated on Lotus being able to make performance gains to stay in the hunt. The team didn't gain enough, and didn't win again.

So if the front three rows wipe themselves out at the start and Felipe Massa wins for Williams, that doesn't mean it's going to win the championship. Likewise, if a fast car falls apart while leading, there's a fair chance it'll be leading a lot in the following 19 races.

Secondly, the idea of slow-burning starts is rooted in a different time, one long past. It was not unusual for teams to start the season with their old cars. While there are a couple of examples of it this century (Ferrari in 2002, 2003 and 2005 and McLaren with the MP4-17D that was ultimately used for all of '03), this is now an obsolete practice.

Even if you do end up with a faster car, the power of being ahead cannot be underestimated. For a classic example of that, look at the 2009 season. The Brawn was mighty at the start of the season, winning six out of seven. But for the rest of the year, the car only won twice, champion Jenson Button only twice even got onto the podium and the Red Bull was the superior car. But when you are playing catch up, it's extremely difficult.

Another example of this was 1991, when the Williams-Renault FW14 proved unreliable early on and McLaren's success at the start of the campaign meant it was never caught despite falling back on pace.

So while you could well see some big moves lower down the order, with an underachieving team like McLaren ripe for making gains this year, the die is cast early on.

Even if the Australian GP is not overwhelmingly powerful in its predictive power, the four races as a set are hard to argue with.

So when you hear people talking about game changing Spain packages, more often than not it's just convenient bluster. Last year, Force India did it with what was characterised by some as a b-spec car, but it wasn't in the title fight. After all, all the teams at the front will find time, and to catch the leader you have to find everything they find and your own on top to get the swing.

It's true that bigger changes in the order are more likely this year because of the significant aerodynamic rule changes, and just because it doesn't usually happen doesn't mean it never does.

But the start of the season is critically important, and it has been for some time. And don't think that the teams, for all their bluster, don't know that.

Nobody plans a slow start because they've got some heavy artillery to deploy. Red Bull is starting the season playing catch up because it and 'wrist-watch' engine supplier Renault haven't collectively done as good a job as Mercedes or Ferrari.

That won't have been a strategic decision, it's just circumstances. And playing catch up, as history shows, is desperately difficult.

Statistics compiled by Joao Paulo Cunha of FORIX