The closing ceremony of the 2018 Winter Olympics signified the end of the Games last week in South Korea. While the athletes, their coaches and families are still basking in the glow of winning 20 medals at the games, there is also cause for celebration in the Faculty of Economics and Business at the University of Groningen.
The speed skating program in the Netherlands is one of the most developed and best financed in the world, but since 2010, RUG professor Gerard Sierksma has been providing the secret weapon for the Dutch skating team’s consistent success: statistics. Sierksma is a mathematician and an avid speed skater himself, and about 8 years ago, he realized that using statistics could make the selection process for which athletes compete at the Games more objective.
“I went to the technical director of the speed skating squad [Arie Koops] and explained it to him, and he was very enthusiastic from the beginning”, Sierksma says. In an interview with the Wall Street Journal, Koops explains the so-called “Performance Matrix”: an algorithm that takes into account the athletes’ results from recent years, adjusts for location (i.e. altitude and ice conditions) and then determines which athletes are best suited to success in any given event. “The problem with Dutch speed skating is that there are very many good speed skaters, but the number of starting positions is limited”, Sierksma explains.
Performance matrix
So what is the process? Sierksma initially wanted to fill in the 10 men and 10 women in the team by looking exclusively at their past performance at other championship events, but coaches and skaters wanted a trial for the athletes to quality. Prior to the qualification trials, a ranking was made of the athletes’ results in relation to their starting positions in their events. Sierksma’s performance matrix, which was outsourced to ORTEC/Sports to calculate the probabilities, used the ranking system to identify which disciplines had the best potential for a medal and then filled in the athletes and their starting positions based on performance in the qualification trials.
Koops announced in December that he would be leaving the team later this season, but Sierksma expects that his successor will also see the appeal of applying math to the selection process. “We started slowly during the Olympic games in Vancouver”, Sierksma explains – in the 2010 winter games, the Dutch team won 8 medals. “We applied it to a full extent before the Olympic games in Sochi”, where the team won 24 medals.
Professor Sierksma says that with 8 gold, 6 silver and 6 bronze medals at the PyeongChang games, it’s obvious that the Dutch team performed incredibly well – even though he concedes that Norway did even better. But, he is quick to add, “We only do it for more or less one discipline: on skates, speed skating.”
Discussion about this post