Journal d'Lemons, vol.8, no.2, (2014)

I used to do science stuff. I guess it's still in my blood. Last week and especially last weekend when I should have been working on the race car and tow vehicle I instead started poking around in the MyLaps data from the Vodden race. Then I poked a little more, then jotted down some notes, even uploaded some graphs to Facebook, and hoped that would be the end of it. Then on Saturday I gave in an wrote another paper for the Journal d'Lemons.

Now it's your chance to nerd out.

Paper: JOLv8n2.pdf
QP Spreadsheet: Vodden-2014.qpw
Excel Spreadsheet: Vodden-2014-09.xls

The charts in the Excel spreadsheet will be messed up. Fix them if you want. I work in Quattro Pro and the xls file is an export.

A quick summary:

The paper will be out soon,but I thought you might like to preview some of the results of the analysis of the Vodden 2014 race. Here you see how many laps teams completed. Notice the break in the trend at around p130. Notice our car just past p160. [sigh]

Here's the same graph but showing the cars by class. Class A covers the whole range and dominates the top and bottom positions. There is a suspicious class C car in the top 10 while most C class cars are much farther back, but in general, the classes spread themselves out pretty evenly.

So how fast was everyone? Here's the distribution. The mode is a 4:22 (252 sec), and there is an outlier at just past 6 minutes. Statistically speaking, that outlier, PeugeotDaddy, is statistically impossible, being over 5-sigma from the mean. I expect they'll be proud of that.

Of course, you'll want to see the class breakdown. Class A owns the quick laps territory, with B just behind A and C just behind B. In terms of lap times, the classing of cars is spot on.

Speed isn't everything. You need to stay out on track turning laps. Here's the eRTOT, estimated relative time on track, roughly the percentage of time a car is out turning laps. It looks like you need an eRTOT of greater that 80% to even think about winning overall. Note that the actual RTOT is typically 10% higher than the quick and dirty estimated value.

Here is a measure of success, the Index of Doing Something Right (IoDSR). Basically, it shows who did better or worse than expected, assuming the fastest car should finish first, the second fastest second, the slowest last, etc.

Trekkor is going to have a tough time convincing the judges he has a B Class car. In terms of lap times he's got a 147th place car. In terms of finishing position, a 10th place car. The difference, execution, shown here by a IoDSR of over 1400%!

Skipping to the end...

Let see how this works out in terms of class. Wow, it looks like C Class does a better job of finishing these races than A or B, relative to the size of each class. C Class must be composed of a bunch of turtles, A with hares, and B with, I don't know, Toyotas?

The rest are on the team's facebook page under photos/albums/Vodden The Hell Are We Doing? 2014

Edit - I updated the images here. Sorry about the mistake mixing up the classing of some cars. Apparently I had sorted on team name once when I should have sorted on car number. Three multi-car teams registered under the same name. The corrections are now in the paper and it and the corrected spreadsheets have been uploaded.