Game theoryFeb 13th 2018by J.T.twitter iconfacebook iconlinkedin iconmail iconprint iconFANS of winter sports are used to paying close attention to forecasts. Few would fancy taking to the slopes in howling gusts of 50mph (80kph) or temperatures that have fallen to -26˚C (-14˚F) with wind chill. Such conditions have caused the postponement of several events at…
FANS of winter sports are used to paying close attention to forecasts. Few would fancy taking to the slopes in howling gusts of 50mph (80kph) or temperatures that have fallen to -26˚C (-14˚F) with wind chill. Such conditions have caused the postponement of several events at the winter Olympics in Pyeongchang, South Korea, which opened on February 9th. Increasingly warm winters are threatening the futures of many ski resorts around the world, but in Pyeongchang the artificial snow cannons are firing for the opposite reason. The air is so cold and dry that snowfall is scarce, with just seven days of it in February last year.
Yet weather readings are not the only forecasts that Olympic teams are monitoring in South Korea. The strongest countries have arrived with ambitious medal targets and will be keeping track of their chances of matching those tallies throughout the games. Until recently working out who was likely to win an Olympic event was a guessing game based on hunches and limited data. Some of the most popular sports, like athletics and swimming, have had unofficial world rankings based largely on form in any given season. But generally onlookers have had to rely on the odds produced by bookmakers for a guide of who is likely to win Olympic glory. That is no longer true. Though predicting medals in each sport is made difficult by the multiple players that compete in most contests and the rarity of similar tournaments outside of the Olympics, accurate projections are now readily available.
Jacob Zuma’s last stand
Diesel-powered cars are on their way out
How to predict winners at the winter Olympics
Sculpting a romantic lead for “The Shape of Water”
Why 5G may be both faster and slower than the previous wireless generation
A hamster is the latest victim in the row over emotional-support animals
The most comprehensive publicly available projection belongs to Gracenote Sports, an analytics company owned by Nielsen, an American market-research firm. Gracenote provides a variety of data services to its clients, but it is best known for the predicted medal table that it publishes during each Olympics (see chart). At first glance, this seems little different to the various other projections that you can find before each games. A handful of financial institutions produced them when Rio de Janeiro hosted in 2016, using a mixture of macroeconomic indicators and performances at previous Olympics to forecast total medal hauls for each country. Though no such models have appeared before the current event in Pyeongchang, there has been the usual smattering of predictions from snow-sport experts.
Gracenote’s distinguishing feature is the ability to produce quantitative analysis for each event. The company has created a performance index that tracks around 500 events across the various sports in the summer and winter Olympic programmes. About 25 national Olympic committees use the database to keep track of their teams, according to Simon Gleave, Gracenote’s head of analysis. That might seem unnecessary, since coaches already have plenty of data about their protégés from competition results and training. Yet analysing that information to reliably measure an athlete’s current prospects is tricky.
Take the example of Laura Dahlmeier (pictured), a 24-year-old German who competes in the biathlon, a sport which combines cross-country skiing with rifle shooting. Ms Dahlmeier produced a record-breaking performance at the sport’s last world championships in 2017, winning five of the six events and finishing second in the other. Yet in the months leading up to the Olympics she struggled to regain that form. Biathlon has no world rankings, but Ms Dahlmeier has won just two of fourteen individual events in the current season of competitions, falling to fourth in the annual points table. Which version of langlauf’s leading lady would show up to Pyeongchang?
The best way to answer that question is to take every previous contest in the sport and analyse how past results correlate with future success. That process is relatively straightforward in games which involve two opponents and typically produce a winner and a loser, either as teams (like football or basketball) or as individuals (such as tennis). These sports generally use the Elo rating system, which was developed for chess by Arpad Elo, a Hungarian physicist. The formula exchanges ranking points from the loser to the winner, with greater rewards for beating stronger opponents. The difference in ratings points between two rivals can be easily used to calculate the probability that one will beat the other. The extent to which recent results alter the ratings varies between sports, with bigger swings typically seen in those with fewer fixtures, like the National Football League, which has just 16 matches in each regular season.
Yet only two events on the winter Olympics programme, curling and ice hockey, involve head-to-head contests. The others are open tournaments, with tens of competitors jostling for the podium. To produce ratings for such sports Gracenote devised an Elo-style mechanism with modifications. Rather than simply measuring whether an athlete wins or loses a competition, the system predicts the share of opponents that he beats. If he finishes higher than expected, based on his previous rating and the strength of the field for the competition in question, his rating improves. Athletes that have been absent for a while, usually through lengthy injuries, have their ratings discounted. Those that compete in teams have their scores blended with their compatriots. And for those that participate in a number of events, such as Ms Dahlmeier, results in related disciplines affect multiple ratings. A strong performance in the biathlon sprint, a group race, would boost her ranking in the pursuit, a staggered race, for example.
The system has allowed Gracenote to harness all the results of competitors across its 500 chosen events, using data from tournaments as lowly as junior championships. That provides more depth than the company’s previous method, which predicted only the top eight finishers at major international competitions, though this approach was still effective. Of the athletes that it picked to win medals in each event in Rio, 53% did. That figure was roughly in line with betting markets, whose odds are a public record of which athletes gamblers think are most likely to win.
Gracenote still uses the old system to produce its public medal table, which also deals in absolute forecasts, rather than fractional ones. If a French athlete, say, is the most likely to win an event, France gets awarded one gold medal in the table, even though the true probability of the athlete winning gold is less than 100% and his chances of claiming silver and bronze are greater than 0%.
Nonetheless, this approach has allowed the company to demonstrate the possible outcomes of the Russia’s suspension from the current games, which the International Olympic Committee (IOC) enforced after accepting evidence of a long-running state-sponsored doping programme. The IOC has allowed 169 Russian athletes to compete under a neutral flag and banned 43 others, after finding them guilty of cheating at the Sochi games in 2014 (a finding that 28 of them have had overturned by the Court of Arbitration for Sport, though the IOC has still refused to admit them to the games). Gracenote has predicted that the depleted team will win eight medals in Pyeongchang. That tally would have risen to 20 if all Russians had been allowed to enter. The bans have benefited Norway most, as the country will likely gain of the five of the 12 foregone medals—enough to nudge it ahead of Germany into first place in terms of total medals won.
As for Ms Dahlmeier, Gracenote was bullish on her chances coming into the games, so much so that it made her the favourite in each of her six events in Pyeongchang. Her Elo rating in 2017 rose to the highest peak in the company’s database for the sport, which begins in 2004. She missed the first competition of the new season through illness, which hampered her in the annual biathlon standings but not in the Elo model. Her results in individual events since then have been less than stellar, but biathlon is also something of a lottery. Mr Gleave notes that the favourite only wins about 30% of the time, a lower share than in any other winter sport. Ms Dahlmeier’s rating has dwindled a little, but not by enough to suggest that last year’s record breaker has become this year’s flop.
Just six days into the Pyeongchang games, that analysis looks prescient. Ms Dahlmeier has won gold in both of her events so far. No winter athlete has ever claimed six gold medals in one year, though her chances of doing so are still slim. Gracenote has yet to turn the Elo ratings into medal probabilities for individual events, but betting markets suggest that the German has roughly a 50% chance of beating the field in each of her two remaining individual races. Regardless of her final haul in South Korea, Ms Dahlmeier is likely to win plenty more Olympic medals. Gracenote’s research into age curves for each sport shows that the best biathletes can maintain their peak performance into their early 30s (see chart). Expect to see more event-by-event forecasting at future Olympics, too.
NextWhy football’s goalkeepers are cheap and unheralded