Biathlon is a notoriously variable sport: besides skiing, there’s the shooting aspect which can ruin a race in a matter of seconds. Sometimes an athlete like Lowell Bailey wins a World Championship for the first time after 15 seasons competing; sometimes the most reliable guy in the sport, Martin Fourcade, finishes 46th. It’s a wild world.
Norwegian PhD student Øyvind Skattebo set out to quantify just how variable the sport is, and what amount of improvement is needed to make the signal of better performance come through that noise and lead to more medals.
“The trend is for more athletes on the podium in biathlon compared with cross-country skiing,” Skattebo, who recently published one of his studies with Thomas Losnegard at the Norwegian School of Sports Science in the International Journal of Sports Physiology and Performance, wrote in an email. (The paper is available for free/open-access here.)
“The reason for this is not that it’s a ‘tougher’ field than in cross-country skiing,” he explained. “If anything, there is less spread between the best athletes in cross-country skiing than in biathlon. However, the biathletes vary more in performance from race to race due to the combination of several components: skiing, shooting time, and shooting result.”
Skattebo reviewed data from World Cup sprint competitions from 2005 to 2015, making up 109 races for women and 110 races for men. That represented more than 10,000 performance observations, and more than 1,000 observations of athletes who were ranked in the top 10 during a World Cup season.
Looking at the variability of performance within and between individual athletes, he could find some interesting patterns. For instance, men were no more consistent than women, but the top athletes were better matched in terms of ski speed.
Another finding was that while shooting times are highly variable between athletes, in terms of statistical significance, the magnitude of this variation didn’t have much effect on the final result of a sprint competition. Athletes spend much more time skiing than shooting, and missing a shot has a greater effect than simply shooting slower than the next athlete on the range.
“There has been a great development in shooting time the last 10-15 years, and I believe a further reduction in shooting time (for the best athletes in the field) is more or less a waste of training time,” Skattebo asserted. “They have more to gain by getting more accurate or skiing faster. Understand me right, there are still a lot of senior competitors out there that can improve performance by reducing shooting time, but I think more or less that the best athletes don’t have anything to gain.”
It’s possible that shooting speed would have a bigger effect on total race time in head-to-head competitions like pursuits and mass starts, but Skattebo didn’t look at these results because sprints are much more straightforward to analyze – and there are more of them on the World Cup schedule, too.
“Maybe if we had looked at the last shooting in mass start or pursuit, where a quick shooting with several competitors entering the shooting range at the same time can definitely give an advantage on the last lap, [there would have been something],” he wrote.
Meanwhile, in terms of total race time, there was more variability in athletes’ overall performance than in either skiing or shooting separately.
“The shooting and skiing performance varies moderately and accumulates into a large variability in overall performance,” Skattebo explained.
Strategies to try to make teams and athletes more competitive should, at an organizational level, take this into account. In biathlon, having a solid, but not star, athlete — one ranked between 10th and 20th, for example — can lead to medals, but not reliably.
“The annual top ten cross-country skiers vary 1.8% in performance from race to race (in a 15 k skate race), compared with 2.5% for the same standard of biathletes,” Skattebo wrote. “Therefore, at one biathlon event the apparently best athlete can miss the top-10, and if several of the best performers fail a normally ~15th-ranked athlete can win. The biathletes are generally less reliable in performance than athletes in several other endurance sports and, therefore, it is less easy to predict the winner.”
In some ways, that’s a selling point of biathlon: the excitement of a race, the anything-could-happen feeling in a mass start, the lead changes and penalty loops.
“This makes it more exciting to watch,” Skattebo wrote. “Of course, [I’m being] completely objective since this is coming from a former biathlete.”
To get reliable medals instead of waiting for a chance for that 15th-ranked athlete to seize the day requires a targeted approach to improvement. Skattebo applied a calculation used in other sports to try to find the minimal “worthwhile improvement” needed for an top-10 athlete to get one more medal out of every ten sprint starts.
The strategy is to take the variability in performance of the top-10 athletes – and in this case, take the standard deviation as the measure of variability – and multiply it by 0.3.
“Previous statistical analysis of performance times and data simulations have shown that a performance improvement of 0.3-0.4 times an athlete’s SD will lead to about one more podium placement per ten competitions, for an elite athlete normally finishing among top 10,” Skattebo explained. “These analyses and simulations were originally conducted in track and field, but the derived thresholds have later been checked and used for other sports, such as rowing, swimming, orienteering, cycling, and cross-country skiing.”
For a biathlon sprint, that measure corresponds to about 12 seconds for both men and women if the athlete is already in the top 10.
In other words, that variability in performance will always be there for each athlete. But if one of them improves their performance on average by 12 seconds – skiing 1.2 seconds faster per kilometer for me, or hitting one more shot every other race, perhaps – then the best end of that variability will be medal-worthy 10% more of the time.
For athletes ranked further down the standings, bigger gains would need to be made.
“An improvement corresponding to 0.3 x an athlete’s SD that normally finishes 30th place would not get him/her to the podium,” Skattebo wrote. “This since the 30th-place male on average finishes ~96 seconds behind the winner, and an improvement of 12 seconds would ‘only’ get him to 24th place, on average. An improvement of six ranks is substantial, and would of course be meaningful for the athlete. But not enough to aim for the medals. So, even though the thresholds are calculated to evaluate how much a top-10 athlete need to improve to be at the podium more often, an athlete in the middle of the field would definitely benefit from a similar improvement.”
But for a top-10 athlete where 12 seconds will make a big difference, the next question is where those 12 seconds will come from.
One interesting result that Skattebo saw was that simply targeting greater resources for championship events — where medals matter the most — might not cut it. Although such a strategy might raise the performance of a team’s athletes, performance actually improves substantially across the board, so it might not be enough.
And showing up at the Olympics and turning in simply an average performance from the rest of the season will not lead to impressive results.
“It looks like all the athletes (on average) improve 2-3% towards championships compared to their mean performance at the World Cup,” Skattebo said. “So, to even finish equally as [they would at] World Cup they need to improve. And to improve their rank they need to improve their performance even more.”
To take a top-10 World Cup athlete and make an improvement worthy of an Olympic medal, then, Skattebo suggested that a 3-percent improvement would be needed — more like 45 seconds compared to the athlete’s standard previous World Cup time.
A number of strategies are possible to try to get to such a level, and probably they should all be pursued.
“I believe it is not a single explanation, but a combination of … better prepared skis and physiological peaking by the athletes,” Skattebo wrote.
Chelsea Little is FasterSkier's Editor-At-Large. A former racer at Ford Sayre, Dartmouth College and the Craftsbury Green Racing Project, she is a PhD candidate in aquatic ecology in the @Altermatt_lab at Eawag, the Swiss Federal Institute of Aquatic Science and Technology in Zurich, Switzerland. You can follow her on twitter @ChelskiLittle.