Fantasy Cross Country Skiing

Adam Mahar, David Ndereba, Hannah Ajmani



Fantasy Cross Country Skiing is a game in the spirit of all fantasy sports, wherein players select a certain number of athletes to their team and are rewarded for their picks with points based on the athletes' performances. The particular game from which our data set emerged follows the FIS Cross Country World Cup, which is the professional racing circuit for cross country skiing, and operates on the following basic rules:
1) Each player may select up to 8 men and 8 women for their fantasy team each week. There is no limit to how many players may have a given athlete, nor is there any kind of spending cap.
2) Players may only change their teams on days when there are no competitions (typically weekdays).
3) Athletes score fantasy points in accordance with the number of World Cup points they score.


Cross country skiing is among the most predictable of sports, as the athletes who win, win often. The players in the fantasy game take advantage of this fact and select their teams on this basis at the start of the season, using only knowledge held over from the previous year. This is evidenced by our first graph (see “Fantasy Cross Country Skiing”), which displays the relationship between the number of players that selected each athlete on the first weekend and the total points that athlete scored for the entirety of the season.
On the other hand, fantasy football has a much lower rate of predictability. The graph on the right (see “Fantasy Football”), looks at the comparable metric of the average draft order of an athlete in a fantasy football league versus the total points they scored in the following season. This graph demonstrates the lack of a correlation between the expected and actual performance of an athlete. While fantasy cross country players were easily able to draft athletes who performed well during the season, fantasy football players had no such luck and were surprised by the performances of key athletes. This inconsistency across the NFL is due to the fickleness of injury in such a high impact sport and the reliance of each athlete's score on the rest of their teammates' performances in any given game.

Results of Predictability

As a result of the predictability of cross country skiing, lots of players pick the same athletes. The graphs in this section represent the relationships between players (gold) and the athletes they’ve chosen (blue). There are gray lines between each player and his or her picks, and these lines exert a pseudo-gravitational force that contorts the graph to show how these relationships compare to each other.
In the graph labeled “Weekend 5”, the high scoring athletes are seen in the center of a ring of selecting players, surrounded by a ring of barely chosen and unselected athletes. The graph labeled “Weekend 6” appears to show a different trend; athletes are instead cast to the outside of the ring while players form a central body. We selected “Weekend 6” for exactly this reason; every other weekend looks very similar to “Weekend 5”.
The fifth weekend of the season is a seven day stage race, which is an unusual event in the cross country skiing world. Most race weekends are only two days of racing, and even then, an athlete will often only compete in one of the days. In a stage race, however, every athlete must compete in every stage (and finish within a time cutoff) or they won't be allowed to continue racing. After such a taxing event, most of the athletes take the next weekend off. Thus, weekend 6 displays the fantasy picks for this "off" weekend when players are unable to pick their normal favorites because those athletes are not racing. Note how there is no single obvious hotspot of athletes as in the previous graph, but instead several different ones distributed on the outside of the mass of players.

Nationality Bias

We have seen that players are generally very good at the game and tend to pick the best athletes. We were curious to see if there existed any nationality bias in how various players chose their athletes. The first two graphs demonstrate the player nationality breakdown of the male and female world cup leaders. Given that almost every fantasy player has the top athletes on their team, we used these graphs to demonstrate the overall distribution of players’ nationalities within the league.
The next six graphs give a similar nationality breakdown for the top scoring US athletes. Note how the fantasy players for each athlete are predominantly from the United States. We can see a similar nationality bias in the next three graphs; the top three Canadian athletes were most often selected by Canadian players.
The final three graphs are a hodge podge of results. The first is the top scoring Swiss athlete; as Switzerland is not represented in our key, the athlete is mostly chosen by players from “other”. The next athlete is the top scoring Finnish athlete, and displays a distinct Finnish bias, though she is still well selected by other nations. The final graph is for a French athlete, Gros, who had a breakout year; he went from middle of the pack (and thus probably only selected by French players) to excelling in his races and attracting a sum of players from other nationalities