I'll update these around 6pm tonight since most of these games are late but here is current situation:
LAL predicted by 17 and 230.2 total
ATL predicted by 2.2 and 193 total
Cleveland predicted by 9.5 and 181.7 total
Accuscore runs 10,000 simulations of every play during the game and comes up with averages. While the system is not flawless it has it's days both good and bad. I've been tracking it over the past month via paid subscription for NBA/NCAA. Interesting trends do develop when you break the plays out by scoring differentials from vegas lines.
I take the accuscore and measure against the vegas spread to get a differential # and use it to narrow my focus on capping those games. Using the accuscore as a whole metric is about 50%, but breaking it down amongst the differential ranges of numbers like < 1.0, 1.1-2.0, 2.1-2.9, more accurate #'s come about, not only in terms of strong plays but also fading the system.
For example currently: Dogs with a 4.0 differential from vegas line are 7-1. Dogs with a 2.1-2.9 differential are 8-16-1.
I used it for conference play in the NCAA's and I'm using it again for post season neutral court play as well, so far it is 9-7 on NIT Sides, 12-4 on NIT O/U, 4-3 CIT Sides, 4-3 CIT O/U, 3-5 CIB Sides, and 2-6 CIB O/U's.