MaldiniNet — A proprietary Neural Network for football match result prediction.

Leveraging advanced expected goals metrics to better predict the 1x2 market for top european football leagues.

Table of Contents

• Introduction
• Feature Exploration
• Feature Engineering
• Neural Network Creation
• Match Prediction Process
• Betting Strategy Simulation
• Conclusion

Introduction

Goals are like gold dust when it comes to a football match, for fans of multiple sports a try or touchdown score is celebrated fondly, but arguably not as joyful as a solidtary goal scored late in a 1–0 win in an important game in a football match. Football is low scoring, most leagues will average between 2.2–3 goals, if your unlucky you could witness many games over your fandom where goals are non existent and games do not provide that one euphoric blast of relief and happyness. But as a football fan that wont stop you ever coming back for the next game because when those moments come back they are truly special. Football, like any competitive sport is based on scoring more than your opponent, having a better defence to keep goals out is just as important as forward attacking play. Goals are random, we don’t ever know when they will arrive. Everything about the game is stochastic…