Abstract
Synopsis: Create a fair ranking system for the 2018 FIFA World Cup LCAV20171130
Level: BS, MS
Description: Imagine that you are given the scores of all the 32 teams participating in the 2018 FIFA World Cup and you are asked to rank them based on their performance. Would you agree with the actual standings?
The World Cup consists of two stages, a group stage and a play-off round, and the typical approach to choose the teams from the first to the second stage is to rank them based on the number of games they won and lost. However, it neglects the number of goals that the teams score in each game, and how strong the opponents were. By binarizing the output of a game, we throw away valuable information which results in deteriorating the quality of the ranking.
In this project, we propose to study and provide answers to the two following questions: First, given the scores of all teams during one competition, estimate the most fair ranking of the teams. Second, given a set of a games played during the last few seasons, design a tournament such that a) we minimize the number of games played in the tournament, b) we get the most informative set of games that allows us to estimate the ranking fairly.
Deliverables: While doing this project, you will be part of a research team, where collaboration and interaction between its members are essential. You will get an experience with the code review, acquire the best programming practices and learn how to efficiently do the research within a group. The deliverables will be a working and well-documented code base and a report.
Prerequisites: experience with Python (and optionally SQL), background in linear algebra and statistics
Number of students : 1 or 2
Type of Work: 50% theory, 50% programming
Level: BS, MS
Description: Imagine that you are given the scores of all the 32 teams participating in the 2018 FIFA World Cup and you are asked to rank them based on their performance. Would you agree with the actual standings?
The World Cup consists of two stages, a group stage and a play-off round, and the typical approach to choose the teams from the first to the second stage is to rank them based on the number of games they won and lost. However, it neglects the number of goals that the teams score in each game, and how strong the opponents were. By binarizing the output of a game, we throw away valuable information which results in deteriorating the quality of the ranking.
In this project, we propose to study and provide answers to the two following questions: First, given the scores of all teams during one competition, estimate the most fair ranking of the teams. Second, given a set of a games played during the last few seasons, design a tournament such that a) we minimize the number of games played in the tournament, b) we get the most informative set of games that allows us to estimate the ranking fairly.
Deliverables: While doing this project, you will be part of a research team, where collaboration and interaction between its members are essential. You will get an experience with the code review, acquire the best programming practices and learn how to efficiently do the research within a group. The deliverables will be a working and well-documented code base and a report.
Prerequisites: experience with Python (and optionally SQL), background in linear algebra and statistics
Number of students : 1 or 2
Type of Work: 50% theory, 50% programming