Statistical Skier

Continuing on from yesterday as promised, we have two rather large graphs showing the head-to-head results of some top US skiers on the domestic scene.  These graphs are, well, big and to be honest I’m not particularly proud of them.  I’ve been struggling with how to display the head-to-head results from a large number of pairs of skiers in an appealing way and quite frankly I haven’t had much luck.  This is the least bad option I’ve found so far.  Even so, it clearly won’t work for a truly large number of skiers.

In any case, here they are:

These may be a bit confusing at first, so let me explain.  You should read these by focusing on rows of panels.  For example, in the men’s graph we have a row of panels for Lars Flora.  Each of these panels displays his results versus the person in that column using the difference in percent back.  Red dots mean the person on the row won, blue means the person on the row lost to the person on the column.  So the large number of red dots in the Lars Flora row is good (for Flora).

It’s not the best, but you can at least scan along a row and get a sense for how often someone is winning versus a particular person.

I haven’t posted very much regarding domestic racing here in the US or Canada.  There are several reasons for this:

• Useful and meaningful performance measures are harder to come by.  FIS points start getting somewhat wishy-washy when you start having to assume that the penalty system is accurately capturing the strength of the field.  Not saying it’s terrible, but I’ve just been cautious about it, is all.
• Generally speaking, I have a lot less data on domestic racers.  I only really capture events that are FIS sanctioned, and that omits a large number of races here in the US/CAN with fast skiers.
• I know (or are friends of friends) with a some of these folks, and it feels weird sometimes talking about them “analytically”.

Still, I ought to pause for a bit after US Nationals and post something.  So here it goes.  Distance FIS point trends for a selection of top US men and women (excluding US Ski Team members): Continue reading ›

A commenter asked what would happen if we removed the bonus seconds awarded during the distance events of the Tour de Ski.  If we add the bonuses back in we get the following results:

[table id=75 /]

[table id=76 /]

While the actual top 12 have remained the same, the order has not.  The women see relatively few changes, but not so with the men!  The top five men end up completely re-arranged.  Of course, there’s always the caveat that the entire Tour may have played out differently without the bonuses.  Different time gaps along the way could potentially influence people’s strategies and decisions about whether to drop out and how to ski particular stages.  So in some sense, just adding the bonus seconds back in doesn’t literally tell us what would have happened.

Now let’s take a more detailed look, stage by stage.  This is an instance where looking at the time back from the median skier doesn’t make much sense, since the bonuses are happening at the front.  So it probably makes more sense to look at how adding the bonuses back in influences each skier’s time back from the leader:

My apologies for not sorting the panels in a more sensible manner.  On the one hand, without the bonuses we end up with a rather thrilling spectacle of Bauer, Perl and Clara (and to a lesser extent Northug) chasing down Cologna over the final few stages.  But of course, without the bonuses, Cologna (and everyone else) would likely have skied differently, so we can’t really say that Curdin Perl “would have won” without the time bonuses.  Still, it’s an interesting thing to think about.

Despite there being fewer changes on the women’s side, you still end up with a pretty exciting “alternate ending”, with Majdic and Kowalczyk trading the lead mid-Tour and then Johaug rushing forward on the last day to claim the overall title.  Many of the other top twelve women won very few time bonuses (or none at all) so their lines are virtually unchanged.  But the same caveat applies here: if the top twelve women were really all within 20 seconds going into the final two stages, those races would have played out far differently than they actually did.

With the women, we see first that without the distance event time bonuses Justyna Kowalczyk would actually have surrendered the overall lead very briefly after Stage 5 to Petra Majdic.  That would have created some excitement, at least.  Here’s a closer look at how the Tour “might” have played out between some key players.  I’ve used seconds behind the median skier again simply because it creates more separation between the lines making them easier to distinguish: Continue reading ›