Overall the model has made 4.5% betting on totals. It seems to be relatively consistent on all probabilities and odds, however a minimum 10% overlay might be preferred. There are as many overs bets as unders, which bucks most of my models trends, and both have performed equally well. What is interesting is that betting over 170 or under 145 both prove losses. And this is significant with 568 bets at 8.9% ROI.
This could be due to team injuries, or a conservative model, but it could be a strategy to eliminate these bets from your portfolio, although it is a relatively small sample size.
A minimum 10% overlay has produced consistent profits over the years and when looking at the results by league as shown on this link, we see no major big winning or losing leagues, thus strengthening the power of our totals European Basketball model