Steel, have you tried modeling/playing these underdog type picks ATS and/or with teasers? I assume you're system is picking up dogs who have a better than implied odds. Thus wouldn't in reason some of the games will be close even if they aren't moneyline winners?
I found cheap online books with 6 and 7 pt teasers and have done well by picking my spots. I like the Memphis pick tonight for points and ML. It actually went off at +900; I do think Memphis wins games like this more than 10% of the time even sans Conley.
5 and 6 pts are more common and you need to get them cheap and hit the 2 teamers about 75% time or more. I rarely do it since it's not a strategy that scales to volume bets. But can be good when you find an already inflated line and ask yourself will the Warriors (as an example) really be motivated to win this one by 24+ pts off of one day rest. Sometimes I leave the teaser open for a few weeks until I find another situation I like equally well. I've made money but it's a tiny sample size. It's not something I recommend but can be fun if you're already inclined toward such things.
Bucks and others hit nicely. Google the below part to see what a rare Tuesday night it was.
A true long shot
One additional perspective on this combination of unlikely events: A $100 three-team parlay on the Magic, Nets and Bucks in Las Vegas would have paid $20,712.50.
Stat of the Day
The Elias Sports Bureau reports that this is the second time in NBA history that the team with the worse record won every game in the NBA in a given day (minimum 15 games into the season, at least six games played). The other instance was Jan. 7, 1969 (six games).