I have written previously how I used Azure Machine Learning, R-language and Power BI to calculate and visualize betting odds for Finnish ice hockey league games. Now I decided to widen my experiment to include NHL games.
I’m using neural network model behind the scenes. I’m not using directly goals, but calculated expected goals, which include shots on goal. Because there are so few games played in this season, I’m still including last season games when training the model. Maybe after week or so I will drop off those last season games when I’m retraining my model.
I’m using 5 € bets per game to follow how my model is performing. At least first round was quite unsuccessful and I ended up losing 26 € in 11 games. Let’s how this continues. Note, there are two pages on below report – Liiga and NHL. Use arrows symbol in the right lower corner to toggle report to full screen mode.
