It all comes down to volume. 20 bux over a thousand bets is $20,000. -107 versus -110 is about a $254.90 difference in collections at 50%. Betting a flat 20, and not 21.40 or 22 respectively. Still losing money, but losing less.
As far as the value of the half point, of whether +3.5 at -107 is "better" than +4 at -110. It all depends on how often it happens, or is apt to happen. It is borderline at best. it also is relative to the size of a person bets and his expectation. A 20 dollar bettor is literly on going to save/make pennies. At -110 he will collect 18.18. At -107 he will collect 18.69. So 51 cents difference. So in that case getting that extra half point and sacrificing the 50 cents might be worth it.
But for guys betting a grand or more a game, and betting many many games a year, then the better vig is the better way to go in my opinion. Basically it come out to about a 2.5% probabilty for it to make sense to lay the -110. Meaning if you expect to have 3 games in 100 lined at 3.5 end in the fave winning by 4, then laying the extra vig would be worth it. I am not so sure that is the case here. I have not checked the numbers in hoops lately. But I know in football it isn't that high.
So depends on the sport and the expected probabilities I guess.
But same line better vig is a no brainer obviously.
When it does make sense is when the low vig place are offering high vig with the 'light' side, and you can find a flat -110 place offering the -110. This past week is a good example. NE was -3 +119 for awhile. They had -3 +115 for a couple days. So finding +3 -110 is a nickle scalp basically. I pointed out that Intertops had +3 -105. I think the still do actually. But from what I have heard their limits are no feasible to try to scalp with them.
These are all interesting srguments, and really show how different people look at differing aspects.