The UEFA Champions League is the biggest annual soccer competitions in Europe. Our goal was to predict the outcome of the tournament. We used data from 2004 up to 2020 and we first used Elo rankings to rank our teams based on their abilities. We have also introduced home-advantage and showed how big of an impact the teams have by playing at home. We then ranked the sixteen teams that qualified for the round of sixteen of the competition. We did that by over-weighting the Champions League results of each team, rather than their corresponding league's results. We then used Poisson distribution to predict the goals scored by each team. We showed that the goals scored by each team are approximated by a Poisson distribution. We then used what Dixon and Coles did in their model and introduce a time-decay model that over-weighted the most recent results. We then found offensive and defensive strengths for each of the sixteen teams. Finally, we simulated the tournament and we had results for both models and we compared them with each other, as well as with the results found from fivethirtyeight.com.


Pasteur, R. Drew




Mathematics, Predictive model

Publication Date


Degree Granted

Bachelor of Arts

Document Type

Senior Independent Study Thesis



© Copyright 2020 Zacharias Andreou