That's actually pretty good cofaga, I like the logic. You can tweak it a hundred ways but at the end of the day you're playing off a base revenge scenario where the opponent comes in hot off a big win (and is thus potentially overvalued even as an away dog).
With that -5 to -10 line parameter you set the dog isn't getting much beyond the standard away court bonus, which is great considering the home team is a superior performing team on the season who likely just lost the prior meeting on a fluke. That line filter tends to fit with the core logic, which is what you want.
The sample size is low at 4-6 plays per season but that can be an advantage to some extent because if you're on the fence about the system in terms of it being valid or not it's not going to bankrupt you or lose you a disproportionately high amount of units as you play it to find out since it's grouped in with all your other system plays.
If say you make 100 or more total plays in an NBA season on average from the systems you've created or follow, then adding in a borderline low-volume scenario or two like this one won't hurt you badly if it winds up being a 40-50% winner...and of course if it wins more than 52.4% then it just adds to the others as well.
Nice work.