Context As part of a research that I made at the University about trends of sports betting agencies (bookmakers), I collected data about winning odds for 2 years. The data collected from a global bookmaker and it includes winning odds of football matches around the world. The goal of the research was to learn and identify trends in matches (lines) that is possible to bet, in order to make a smart decision (bet). Many gamblers see the odds in the moment they bet (between the first opportunity to bet and match start date) but they don't know the history of the changes in winning odds. The bookmakers change the winning odds according to several scenarios: players injury, weather at the time of the match, rumors about players, opening players, the wisdom of the crowd, stadium match and many more factors. The gamblers don’t have this kind of information or the ability to analyze it to the right odd. A change in the winning odds due to these factors can help the gambler make a better bet (for example, a team that was a favorite at the beginning of the line but few minutes before the start match the team is underdog - has a high probability to lose). Content The data set has been collected for 2 years from a global bookmaker and includes all the winning odds of football matches from all over the world. The data set contains more than 35,000 matches, in which more than 200,000 changes in the winning odds that were made up to the beginning of the match. The matches from all over the world – from an amateur league to a professional’s leagues. The database is divided into two tables (CSV files): Matches_Odds - Describes the odd's change for a winning (and the time that it happened) by the bookmaker for every match till the start match date. Matches_Results - Describes the result for every match. Inspiration Is it possible to predict the result of a match and get higher accuracy from the betting agency? If we want to predict tie – Yes or No, how the model will be change? Is it possible to achieve high accuracy in matches that have reverse line movement? Does the date of odd’s change and the start date of the match affect the result of the match ? Develop and check which independent variables affect the match result prediction. What is the optimal time a gambler should bet on a match? Which leagues/teams can achieve a higher accuracy thanks to the winning odds? Should we make a cluster to the leagues/teams? Is the expected value according to the bookmaker is positive? Relevant Links [The basics of reverse line movement][1] [Why do betting odds change?][2] [Beating the bookies with their own numbers][3] [1]: https://www.pinnacle.com/en/betting-articles/educational/basics-of-reverse-line-movement/QAH26XGGQQS7M3GD [2]: https://help.smarkets.com/hc/en-gb/articles/214559905-Why-do-betting-odds-change- [3]: https://arxiv.org/ftp/arxiv/papers/1710/1710.02824.pdf My Research In my research, I developed independent variables and examined their impact on matches that were reversed movement of winning odds. In addition, a statistical model was adjusted to predict the result of a match, when the final accuracy was more than 5.43% of the accuracy of the betting agency (bookmaker). The development environment was R Studio. # I hope this data set will help further studies :) Elad Silvas