5Dimes understands the value of a run. Going from a moneyline to +2.5 runline, there are simple conversions you can perform to come up with the correct +2.5
odds, just like you can convert a moneyline to a +1.5 runline. If you perform this yourself, you will see that the runlines are correctly priced. I do this routinely with +1.5 and -1.5. You have not gained a magical edge by playing short odds. I have not ran an exercise for +2.5 but something tells me that it will be priced correctly - actually it will be priced worse, as alternate lines typically have higher juice than standard lines.
So, if you work through that and come to terms with the fact that +2.5 runlines give you between -3 and -10% EV on average, parlaying the +2.5 runlines subtracts another 5% or so of expected value, and thus you will lose long term. You could win thousands short term, but it does not matter, the game you are playing is mathematically configured to hold money from players over the long term.
That is my quick analysis of the situation. Without a massive game database and lots of backtesting, you are just gambling like everyone else. Even if you were to run this through every home team last year and you had records of the +2.5 odds for every game, and you come up with something that wins --- you still just found something in the data without having an original hypothesis and novel thought process as to why you are seeing what you are seeing. You must follow scientific method. Otherwise you just look for something that worked once.
My suggestion is to start a database and start logging +2.5 runline odds (and the total and other game prices, of course). Along with the odds history, begin logging box scores. This should be easy to do. The more sources of +2.5 lines the better.
You've got to perform an exercise like this before putting money on it.
And one other point - you are not the first to parlay +2.5 (many others on SBR have done it publicly (Rake?), and I'm surprised they haven't chirped in here yet telling you how badly it failed)