Nerding Out: Is World Cup Downhill Racing Getting Closer?
|I think it’s as hard to get in the top ten now as it was to get on the podium when I first started racing.|
There are so many variables in downhill racing that it makes such a simple comparison a bit useless. It was at this point I spoke to the World Cup’s resident statistician, Eliot Jackson, who jumped on board with the project and offered this to think about:“Is percentage is the best metric to use? Let's take a 1:00 piece of track - if that's true, then being three seconds back at the top of Fort William is the same as being three seconds back on the second sector of Val di Sole. This year the former would put you in 42nd and the latter would put you in 17th. Percentage doesn't take into account the difficulty of the track.“If Windham is in one year that will skew times much tighter on a percentage basis than say a track like Val di Sole. Lourdes is insanely fast in the dry and unrideable in the wet, i.e. really close times and really not close times.”So we decided to dig deeper.Where can we fairly compare racers through the last few years of racing? What piece of race track is the same length every year, with the same features and is unaffected by the weather? The best answer we could come up with was the top section of Fort William. It’s a section of track that lasts roughly one minute every year going from the start gate down to around the first rock garden. Rain doesn’t have a massive impact on this section of the track and, in fact, riders often say it rides better in the wet as it improves traction over the marbley surface. The start hut has been moved backwards a bit over the years, but for our money it was probably the closest we were going to get.So what happens if we do the same test there? Well, it's not the result we were expecting, that’s for sure. While the top ten results show a rough correlation towards being tighter, the top five results show the opposite and suggest it’s getting easier to be in the top five in this sector.It’s interesting that most World Cup results sheets show the same pattern. A large clump of riders in the middle with outliers on either side. At the back of the race, riders drop off due to punctures, crashes, etc, and at the top, well, the level is just that much higher. This probably explains why those guys can command the big bucks. Eliot graphed this for every World Cup round since 2000 and explains: "Each little line represents a race, with the y-axis being time. Just by looking at the graph a conclusion that one could make is that racers are getting faster, but to me, on a macro scale, this graph shows how the variance between tracks throughout the year has gone down and how tracks, on the whole, have gotten a tiny bit faster. The thing to look at here is the slope, or how steep, each of those little lines is. The steeper they are, the more time in between each racer. The flatter they are, the closer that race was. Just from a glance, they definitely seem to be getting a bit flatter, but one thing to note about that is that a lot of races are also shorter now, which means they will inevitably be closer. Super interesting to think/look at! See if you can pick out some of the famous mud races!"If we could measure the rough slope, that could be a better measure of the actual closeness of the race, rather than using a percentage. Eliot used this technique last year to predict winning race times and was able to get within 0.3 seconds, which boded well.So, we went back to Fort William split 1 and did the exact same test. You can see the plots in the animation below, all following the same pattern Eliot described.
0% Loaded prev 1/10 nextTo remove outliers, we took out the top and bottom 20 from each year and measured the slope. Yet again, there wasn’t much of a correlation at all. If downhill racing is getting closer, the top section of Fort William isn’t where it’s happening.
Think you've cracked the code? Create your own World Cup DH fantasy team now - racing begins April 27th in Slovenia.
Support Eliot and his stats website here.