Seasonal Linear Predictivity in National Football Championships

Predicting the results of sport matches and competitions is a growing research field, benefiting from the increasing amount of available data and novel data analytics techniques. Excellent forecasts can be achieved by advanced statistical and machine learning methods applied to detailed historical data, especially in very popular sports such as football (soccer). Here, we show that despite the large number of confounding factors, the results of a football team in longer competitions (e.g., a national league) follow a basically linear trend that is also useful for predictive purposes. In support of this claim, we present a set of experiments of linear regression compared to alternative approaches on a database collecting the yearly results of 746 teams playing in 22 divisions spanning up to five different levels from 11 countries, in 25 football seasons, for a total of 181,160 matches grouped in 9386 seasonal time series.