Participation Rates

One really simple question that occurred to me while playing with all this skiing data is, How many races do these athletes actually do?

Now, I can’t really answer that completely, since I don’t have every single cross country ski race in existence in my database.  But what I’m really getting at is the notion that we’ve seen an ever increasing number of races on the international scene:

The stage race refers to the newly adopted Tour de Ski event (and the use of the same format at the World Cup finals this past year).  But each individual stage in these events is counted among either the distance or sprint events.

With more and more races to choose from each season, are skiers racing a smaller and smaller proportion of them?  I mean, there’s only so much travel and racing that a person can handle over the course of a season.  So really what I’m interested in is participation rates.  What proportion of the available major international races do skiers actually do?

This is actually harder to pinpoint than it seems, as it requires a brief discussion of some nitty gritty details of World Cup racing.  In general, for World Cup races the number participants is allocated by nation (Olympics and World Championships have analogous but slightly different systems).  Each nation gets a certain number of spots, based roughly on how good that nation is.  Norway can start a ton of athletes; Botswana not so much.

However, in addition to these “regular” spots, the host nation for a particular race will get some number of extra spots for it’s athletes.  These tend to be filled by lower tier skiers who aren’t actually World Cup regulars.

What this means for us is that when you look at every skier for a season and then look at how many races they’ve done, you have a lot of skiers who’ve done only a handful of events.  Many of these skiers aren’t really in the group I’m interested in, since they aren’t really “World Cup skiers” per se.  I’m more interested in the people who are actually on their nation’s national team and getting regular opportunities to start a race, even if they don’t take all of them.

So the first thing I’m going to do is to drop skiers who’ve done fewer than three races.  Arbitrary, I know.  And technically I’m contaminating things since it’s possible to have some truly elite skier who for whatever reason only does two races (injury, illness, etc.).  But otherwise these skiers just swamp everything else in the data.

Here’s a graph of participation rates over time by quantile (don’t worry, I’ll explain):

First, I’ve actually smoothed the data slightly, so these are trend lines, not the raw data.  I’ll return to the raw values in a moment.

Now, there are a lot of lines floating around here, but don’t despair.  The vertical axis means just what you think it does: the proportion of races done.  Now, I’ve included 9 different lines to represent the distribution of participation rates for each season.

I don’t just want to know what the average participation rate is, because in any given season there will be people who do nearly every race, and some people who do only a few.  I want to track whether each subgroup of participating in more or fewer races.  For instance, the top green line tracks the participation rate of the 90th percentile group.  These are the people who do more races that 90% of the skiers that season.  The level on the y axis is the actual, absolute participation level for that group.

Let’s look at the distance panels first.  The bottom third with respect to participation rate (red lines) appear to have been slowly dropping.  So has the middle third (blue lines).  The top third (green lines) has been more steady, although you see some waving up and down for the women.  Only 10% of skiers participate in more than 80% of the available distance races in a given season, on average, and that has remained roughly constant for 19 seasons.

The sprint panels are a bit messier, but that is to be expected given the newness of sprinting.  Early on, pretty much everyone did nearly every sprint race.  There weren’t very many of them, and there may have been a perception that they were “easy” compared to distance races in terms of how taxing they are on your body.1

Now I want to return to the un-smoothed, raw data for a moment.  The reason is that it turns out that the trend lines, while useful for picking out general trends, particularly when I’m plotting 9 lines all on one panel, have obscured something really interesting:

First, now you can see why I smoothed these.  With this many lines bouncing around, it’s tough to make sense of.

But if you look closely, you can see the hint of some periodic behavior in the distance panels, particularly with the top third green lines.  In fact, it looks to be dropping every four years or so.  Hmmmm.  Wonder why that could be?

The Olympics.

Check the years when the green lines drop; they are Olympic years.  That’s when many of the top skiers will skip large portions of the World Cup circuit to rest and prepare for the Olympics.  I could do some fancy autoregression and calculate the lag estimate to confirm this, but really, just look at the graph.

  1. I am most certainly not saying this is true.  In fact, I suspect it’s just plain false.  But I’d bet that that perception was fairly common early on with sprinting.

Related posts:

  1. Gadflies & Punching Bags
  2. How Well Prepared Are World Cup Rookies? (Part 1a: Distance)
  3. Is It Panic Time For The Norwegian Men’s Distance Skiers?
  4. Podium Heartbreak
  5. Week In Review: Friday Oct 1st

About Joran

Comments

3 Responses to “Participation Rates”
  1. Brayt says:

    It’s worth nothing that the peaks on the graphs correspond exactly to the years without any major championships (OWG or WSC). Do medals at major championships count more than an overall WC title, or is it just that you can prepare for those events away from the World Cup, while it’s rather hard to accumulate WC points racing or training elsewhere.

    • Joran says:

      Could money be a factor? Are WSC/OWG medals worth more $ than WC points/podiums? (Factoring in possible bonuses from national ski federations or sponsors?)

Trackbacks

Check out what others are saying about this post...
  1. […] This post was mentioned on Twitter by Joran Elias, statskier. statskier said: Participation Rates http://goo.gl/fb/Nkd24 […]

    [WORDPRESS HASHCASH] The comment’s server IP (208.74.66.43) doesn’t match the comment’s URL host IP (74.112.128.10) and so is spam.



Speak Your Mind

Tell us what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!