How we class cars
We start with every car’s fastest qualifying lap time plotted on a timeline. Below is a portion of what one of those timelines might look like, each circle represents a car. We then look for natural groupings of the cars. You can see how the cars that will become Class 5 and Class 4 form groups that are distinct from each other. In those cases it’s clear where to class those cars. It’s not as obvious where to draw a line between Class 3 and Class 2. This is when we look deeper into the qualifying data for each car near the area that we suspect the line will be drawn. For cars 22,1,15,14,99,111,22 and 7 we would look at the fastest driver’s second fastest lap and average lap time. We would also look at the second fastest driver’s fastest lap, average lap time as well as the average of all lap times for the car. When we look at all of the data in this way it becomes much more clear where to class those cars.
How we monitor classing during a race
During races we generate classing analysis pages every hour. Below is a sample of one of those pages after 5 hours of racing. Each column contains all of the cars in the race sorted a different way. That sort mode is listed on the top of each column. The cars are also color coded to match the class they are currently in. From this example some cars are starting to stand out as possibly not being classed correctly: #110 (currently yellow indicating Class 2) is clearly running in Class 3 (green) by most indicators. Based on this data #110 would be moved to Class 3.