Right before the start of the Tokyo 2020 Olympic Games we released our event by event podium prediction for every single competition at The Games of the XXXII Olympiad.
We were not the only ones to launch an event by event prediction. The Associated Press did theirs as well. How did we compare?There were a total of 339 events. The Sportspedia predicted 281 medalists with the exact name and color, and 565.5 medalists without the correct color. We did a bit better than AP, who predicted 266 medalists with the exact color, and 550.5 in total.
In terms of the overall tally, we correctly predicted the top 3 nations in the United States of America, the People’s Republic of China and Japan, in that order. The prediction was of 38 gold medals for China, which is very close to the 40 they had. We, however, also predicted 56 for the US and they ‘only’ won 39, which made the medal tally be decided in the very last day of the Games.
Our prediction was that they would win 128 medals; they had 113, which is relatively close.Japan was also a bit disappointing. We thought they would nail 34 gold medals and 76 in total. The real result was 27 gold medals and 58 in total.Another company, Gracenote, also predicted the overall number of medals for the top countries at the Olympic Games. We also fared a bit better in the medal totals compared to them.
Gracenote released medal numbers for 30 countries. Compared with them, we had a closer number in 15, for 9 of Gracenote. Considering gold medal accuracy as the tiebreakers it would be a 19-11 scorecard for The Sportspedia.
Now, this is not just to gloat for our accuracy -although partly it is- but also to analyze if the medal tally was predictable. Considering the premise that our predictions were good, then we draw some conclusions:
- The United States had more chances for gold that were not won. They decreased from 46 to 39 gold medals, although they looked set to do a lot better.
- Russia, as expected, did better as ‘ROC’ than they did as ‘Russia’. Even though some of their competitors were barred, they still aced events where they were expected to earn a lot of gold, such as in Wrestling and Fencing.
- China was very predictable. We expected them to take 40 gold medals and took 38. While they do not dominate every sport, they clearly dominate the ones in which they are best. As we expected, every Chinese weightlifter won gold. We also expected them to take all 8 gold medals in diving and al 5 in table tennis. They dropped one in each. Still, predictable.
- South Korea also disappointed. We expected them to take 12 gold medals and be n the top 10 of the medal tally, but just managed to win 6 events. They took the 4 gold medals we predicted in Archery and the 2 we predicted in Fencing, but disappointed by not taking any of the 4 we predicted in Tae Kwon Do and the 2 we predicted in Shooting. Ironically they led the medal standings at some point.
- Great Britain did well. Although they did not repeat their excellent 27 gold medals from Rio 2016 and took 22, we had expected them to take way less, and they won 65 medals compared to the 59 we expected.
- Ethiopia also disappointed. Its athletics team looked stronger on paper than they actually performed. The Sportspedia predicted 5 gold medals but they only took the Men’s 5,000.
- Canada did way better than we expected. We thought they would only take 2 gold medals -they took both- but also took other 5 in different sports.
- The club of 10. Netherlands, France, Germany and Italy took home 10 gold medals, which was close or exactly as our estimates of 10, 10, 13 and 9 that we expected for them.
- Favorites won. We predicted every single gold medal winner for 3×3 Basketball, Artistic Swimming, Baseball, and Basketball, and only missed one in Weightlifting.
- Unpredictable sports. Contrario sensu, we only got 2/10 right in Shooting, 2/10 in Sailing and 1/12 in Fencing. This is also not surprising considering they are very unpredictable sports in general.
It is always fun to do this exercise. Fortunately sport it not very predictable, which is one of the reasons why we like it so much. But assessing results compared to the expectations allow for countries and teams to evaluate how they performed when it mattered most.