Kiss my Score is a social app to play against friends predicting football scores. It’s pretty easy to play, you just need to input your guesses for matches of the World Cup and other national and European competetions. Besides, the app allows you to follow matches live, create private leagues with your friends and even having the chance to win some prizes.
With more than 260.000 players and over 8 Million guesses registered the portuguese Startup decided that they should do something else with all the data points collected from worldwide users guesses’.
Will collective intelligence combined with traditional football data allow us to predict football scores accurately?
What is the challenge’s objective?
The main objective is to predict football scores, with data from the Kiss My Score users’ guesses.
Having that accomplished, a secondary objective would be comparing the quality of those predictions with a traditional algorithm, such as: https://projects.fivethirtyeight.com/2018-world-cup-predictions.
How can you share the results of your research and ask questions related to this challenge?
You can send your notebooks and any questions related to this challenge to [email protected].
What does the dataset look like?
There are +260.000 users and +8.000.000 guesses, with an average of 100.000 guesses per World Cup 2018 match and 1.200 per national/European competition match.
The dataset is composed of 6 tables:
Bets: The users’ guesses, with the date of creation of the guesses and the date of the last update made to those same guesses;
Clubs: Clubs and National Teams competing on the available competitions;
Competitions: Competitions included in the app and the country of origin;
Countries: Countries of origin of the included competitions;
Matches: The matches, with the home and away teams, final score and date and time of the matches;
Users: Gender, users’ first login date, favorite national team, favorite club and nationality.
Have we ever used KMS players’ guesses to predict football scores?
Relying only on the most frequent bet for each match we simulated the World Cup Group phase and tried to predict which teams would proceed to the next phase. You can learn more about that here.
I agree with the Terms and Conditions
