I use the Sagarin Power Ratings. Not the predictor or elo chess. I also do not factor in home court.
I look for an underdog according to vegas spread who Sagarin has winning outright. Tonight Wizards are underdog to Timberwolves but Sagarin has Wizards rated higher in power rankings. So make Wizards a ML play.
I use ML because then the system doesn't have to hit over 50% to turn profit. Hell, it can hit 40% and make good money.
But this is more of an experiment and I have never used it before. CFB I am having great success so thought maybe it work in NBA with couple tweaks. NCAA hoops would be interesting too but that is way too many games to look over. Especially on weekends.
My whole theory when coming up with a system is put parameters on it (you get select few plays each week/night) and have it make logical sense. So for this NBA system, I am taking a team + ML who Sagarin has rated out better. Makes sense to me. I have 3 outcomes. My team wins outright, my team covers spread but loses, or my team doesn't cover spread and loses. In football underdogs cover spread but I figure in NBA more times than not my team will win outright or not cover spread. So ML made logical sense to me.