We predict tennis matches and we do this by using a form of Artificial Intelligence: Machine Learning. 

With Machine Learning, in short, we're capable of identifying patterns in past matches and can provide predictions for fixtures (future matches).

We analyzed 400.000 past tennis matches by looking at their statistics like surface, player's ranking, head to head, win percentage and more. We then took this data, along with the market outcomes and created a model: the model was learning to identify winning selections better and better. 

After the model is created, we are now capable of predicting markets for future match outcomes by assembling the same data for those matches and asking the model for its prediction.

Because we use Machine Learning, we're capable of not just making a prediction, but we get a certainty probability for each prediction too. This helps us to see how strong a given prediction from the machine is and we can use data for further processing.
​