Moving to the 2017 pre-season data, we have almost no prior information. Perhaps they were bluffing? Fuel-correcting lap times in 2017 For unknown reasons, Renault looked considerably faster in pre-season testing than they did once the season started last year.
The one notable exception was Renault, where both estimates were far from the actual data. Nonetheless, in almost all cases, the long run estimates came closer to the actual data than the best lap estimates. Neither method was perfect, because even when we are correcting for tyres and fuel, we may be missing other factors (e.g., engine modes), and some teams may have improved between pre-season testing and the races. Note that I wasn’t able to obtain enough long runs for Haas or Manor to make any estimates last year. I chose the first four races, because the return to Europe usually corresponds to the first major car upgrades. In the graph below, we can see how my estimates from long runs (red) and Andrew Benson’s BBC estimates from single laps (blue) compare to the actual average qualifying gap (green) across the first four races of 2016 (using best times set by each team). I showed a method for doing that last year, which involved “anchoring” fuel-loads for certain stints that were known to be from the teams running race simulations. Besides incorporating more data, this approach potentially allows us to accurately fuel-correct the lap times. This was exactly the approach taken by Andrew Benson for the BBC analysis last year when he estimated the team hierarchy at the end of pre-season testing.Ī superior approach is to consider performance on long runs, consisting of many consecutive laps. The best we can do with single lap-times is consider the length of the stint (i.e., how many laps the car kept running after setting its best time) as a lower bound on the amount of fuel the car was carrying. Fuel loads and other factors that teams may use to disguise lap-times are more difficult, because they are largely invisible. Making time corrections for tyre compounds is relatively easy, because they are visible. Correcting for these factors is therefore critical for any meaningful comparison.
SAN MARCOS SOLAR ROOF TILES FULL
The difference between a fresh ultrasoft and a fresh medium compound tyre can be over 2 seconds, and the difference between a full fuel load and a qualifying fuel load can be over 3 seconds.
SAN MARCOS SOLAR ROOF TILES DRIVER
These alone, however, can be unreliable estimators, because times are affected by many factors, including tyre compounds, fuel loads, set-ups, engine modes, traffic, and how hard the driver is pushing.
When it comes to testing data, media focus tends to be on the headline times, including the absolute fastest times set by each car. What is the form guide this year? Reviewing 2016: The importance of long-run analysisīefore starting, it’s worth revisiting last year’s pre-season analysis to demonstrate the value of this data analysis approach. As I showed in a previous analysis of team dominance, major rule changes have a tendency to give a large advantage to one team, whereas static rules tend to close up the field over time. Not only do we want to quantify how much faster the cars have become, but also assess how Pirelli’s new harder tyres are performing, and see how the new rules have changed the team hierarchy, if at all. With that in mind, examining 2017 pre-season testing is a fascinating proposition. There is a chance they will be the fastest circuit-racing cars ever created. Even from the first days of testing, the 2017 cars have looked viscerally exciting compared to those that came before. Instead, we have become accustomed to changes designed to curb speeds and improve safety, which led to the 2014-2016 breed looking somewhat prosaic in corners by the heady standards of the previous era. Not since 1966, when engine sizes were doubled, have the sport’s rule-makers introduced full-format changes with the primary intent of making the cars faster and more exciting.
2017 marks one of the largest changes in the history of Formula 1.
0 kommentar(er)
