Which specialty matters most for the world ranking?
Cyclists with different specialties can make it to top of the world ranking. See for instance this table, where the UCI Ranking is combined with the Cycling Elo’s for nine skills. In this article I make an estimation which specialties are more important to explain the variance in number of world ranking points. The hill quality is the most important determinant to explain the variance. The climbing skill is getting more important and the time trial skill is getting less important.
I use a regression analysis with five different skills that define for a big part the different qualities of professional cyclists, controlled for age, team quality and team leadership. To calculate the relative importance of the different variables, the R2 of the regression analyses are decomposed using a method proposed by Lindemann, Merenda and Gold. The relaimpo package in R is used for this. The skills are based on the Cycling Elo. For the world ranking i use firstcyclings ranking, since it has a comparable ranking system for a longer period. 2007 is the first year in this analyses, because it’s the first year the Cycling Elo is calculated.
Figure 1 Decomposition of variance (R2 in a regression analyses) in points First cycling ranking
In total, these 5 skills and 3 control variables explain around two third of the variance in total number of points collected by individual cyclists. Out of the race skills, the hill Elo is the most important: it explains between 15 and 20 percent of the variance in points. Partially, this is caused because different types of riders can be top cyclists in different types of hill races. Sprinters with some climbing ability – like Sagan or Matthews - can be world class in the easier hill races. And climbers with some explosiveness – like Roglic or Joaquin Rodriguez – can win hard hill races like Liege-Bastogne-Liege or Il Lombardia.
The cobble specialty has the least explanatory power: only around 2 percent. This is probably because a big part of the peloton hardly participate in those types of races, so it’s hard to get a reliable estimation of their skill in cobble stone races. But also because the number of cobble stone races is low compared to the other types.
Between 2007 and the last two years, mountain skills have been getting more important, while time trial skills have become less important. Not a big surprise, since in the World Tour the number of time trial kilometers has dropped. For instance, it’s now even more important to be a top climber to get on top in the final classification of Grand Tours than in the past, when time trialers like Olano could win Grand Tours.
Looking at the control variables we can see an interesting change. The quality of the team has been getting much more important, while being a team leader has been getting a lot less important. In earlier blogs we saw the inequality between teams has risen. So the best cyclists are more in the better teams. Because of this they can help each other beat the other teams. But they probably also alternate more among each other as team leader during the year.