This was the simplest approach that I tested back in 2018. It was based on laying over 3.5 match goals where the market was expecting goals. This was done on only 41 selections.