# 2018 Winter Olympics Predictions: Behind the Scenes

Print

Our team developed a mathematical formula to determine seedings based on athletes' performance in World Cups and Championships. The predictions are obtained by assigning medals to the top three seeds. The formula incorporates performances from three sources: 2016-17 World Cup season, 2017-18 World Cup season (until 24 Jan, 2018) and the 3-5 most recent events. The idea behind the formula is to grab the overall performance of the participant in the past two seasons, while also quantifying their current form going into the 2018 Olympics by looking at their performance in more recent events. A more detailed post about the formula and different edge cases will soon be posted on this page.

Our team developed a mathematical formula that quantifies performance of the participants in the the 2018 Winter Olympics. The medal contenders are obtained after sorting the athletes based on the score assigned by our formula.

The formula consists of two parts: historical performance and current form. The historical performance consists of each athlete's performance in the World Cup for the 2016-17 and 2017-18 (until 28 Jan 2018) seasons. The current form is captured through their performance in the last 3-5 World Cup events.

Our team developed a mathematical formula to determine seedings based on athletes' performance in World Cups and Championships. The predictions are obtained by assigning medals to the top three seeds. The formula incorporates performances from three sources: 2016-17 World Cup season, 2017-18 World Cup season (until 24 Jan, 2018) and the 3-5 most recent events. The idea behind the formula is to grab the overall performance of the participant in the past two seasons, while also quantifying their current form going into the 2018 Olympics by looking at their performance in more recent events. A more detailed post about the formula and different edge cases will soon be posted on this page.

## Data Collection

The data was collected from the official governing bodies for different sports. Following are the sources:

• FIS: Alpine skiing, cross-country Skiing, freestyle skiing, Nordic combined, ski jumping, snowboarding
• IBU: Biathlon
• IBSF: Bobsled and skeleton
• ISU: Figure Skating, short track speed skating and speed skating

Our team developed a mathematical formula to determine seedings based on athletes' performance in World Cups and Championships. The predictions are obtained by assigning medals to the top three seeds. The formula incorporates performances from three sources: 2016-17 World Cup season, 2017-18 World Cup season (until 24 Jan, 2018) and the 3-5 most recent events. The idea behind the formula is to grab the overall performance of the participant in the past two seasons, while also quantifying their current form going into the 2018 Olympics by looking at their performance in more recent events. A more detailed post about the formula and different edge cases will soon be posted on this page.

## Quantifying Performance

We used the following three ways to quantify athletes' performance:

• Official Points: Many federations maintain an official points list which takes into account various factors like their standings, penalties (if any), and timingso rank every athlete on the same scale. We came across FIS point list which is available for Alpine Skiing, Cross-Country Skiing, Freestyle Skiing and Snowboarding. Using these points, the athletes can be easily sorted and ranked.
• Timings/Score: For the sports that are either time-based or score-based, we came up with a formula to quantify the 'dominance' of an athlete. We apply min-max normalization relative to the best timing/score and then sum up the normalized values. This very useful in cases where an athlete shows a very dominant performance and beats everyone else by a very convincing margin. Finally the athletes can be sorted and ranked using this score.
• Standings: In this method we initially apply a Rank Score function which is an exponentially decaying function. This function is discussed later on. Finally the scores of all athletes are summed up and then sorted to get final standings.

Our team developed a mathematical formula to determine seedings based on athletes' performance in World Cups and Championships. The predictions are obtained by assigning medals to the top three seeds. The formula incorporates performances from three sources: 2016-17 World Cup season, 2017-18 World Cup season (until 24 Jan, 2018) and the 3-5 most recent events. The idea behind the formula is to grab the overall performance of the participant in the past two seasons, while also quantifying their current form going into the 2018 Olympics by looking at their performance in more recent events. A more detailed post about the formula and different edge cases will soon be posted on this page.

## Rank Score function

It's an exponentially decaying function with a maximum value of 100 and minimum value of 1. The function gradually reduces the difference as the rank increases. Here are the order of scores: 100, 85, 70, 60, 55, 50, 46, 42, 39, 36, 34, 32, 30, 28, 26, 24, 22, 20, 18, 16, 15, 14, 13,...,1,1,1,...

This is a very common scoring strategy amongst winter sports to decide the overall winner at the end of the world cup season.

Our team developed a mathematical formula to determine seedings based on athletes' performance in World Cups and Championships. The predictions are obtained by assigning medals to the top three seeds. The formula incorporates performances from three sources: 2016-17 World Cup season, 2017-18 World Cup season (until 24 Jan, 2018) and the 3-5 most recent events. The idea behind the formula is to grab the overall performance of the participant in the past two seasons, while also quantifying their current form going into the 2018 Olympics by looking at their performance in more recent events. A more detailed post about the formula and different edge cases will soon be posted on this page.

## Final Score and Predictions

Depending on the availability of data, we applied one of our three strategies to quantify the performance of the athletes in these three areas: World Cup season (2016-17), World Cup season (2017-18) and the last 3-5 events. After ranking the athletes in those three areas, we apply the rank score function and then take a weighted sum of the three scores. Following weight values were used for the three scores:

• World cup season 2016-17: 0.5
• World cup season 2017-18: 1.0
• Last 3-5 events: 1.0

We gave more weight to the most recent season to get a better proxy of athletes formheading into the 2018 Winter Olympics. Our final predictions are the top three athletes according to our formula.

Our team developed a mathematical formula to determine seedings based on athletes' performance in World Cups and Championships. The predictions are obtained by assigning medals to the top three seeds. The formula incorporates performances from three sources: 2016-17 World Cup season, 2017-18 World Cup season (until 24 Jan, 2018) and the 3-5 most recent events. The idea behind the formula is to grab the overall performance of the participant in the past two seasons, while also quantifying their current form going into the 2018 Olympics by looking at their performance in more recent events. A more detailed post about the formula and different edge cases will soon be posted on this page.

## Interesting Figures

• Germany is expected to win most number of medals (42) which will be an all-time record for total medals won by any country in a single Winter Olympics. The sliding events are going to be Germany's strength in 2018 Olympics.
• Norway is expected to win 16 gold medals which will be an all-time record for most gold medals won by a country in a single Winter Olympics. Norway expected to win 36 medals in total which will mark 2018 as their best performance in Winter Olympics. Cross-country skiing and Nordic combined will be their biggest contributors in medals, according to our predictions.
• Canada will again be a dominant force in all team sports with all three gold medals in Curling. They are expected to win 29 medals which should be their best performance in Winter Olympics yet.
• Japan is expected to be most the improved country with eight gold medals and 14 total medals. Speed skating and figure skating will be their best events in the 2018 Winter Olympics.
• Martin Fourcade, a biathlete from France, is expected to have a stellar performance this year and win four individual gold medals in biathlon. It will be the best individual performance in the 2018 Winter Olympics.