I use a statistical process/models to come up with my lines for each game, and then the spread has to be off by "X" for me to call it a "suggested play". In order to determine how accurate my system is, I keep track of how suggested plays do. However, should I only be comparing my spread to opening lines, or use the line movement that makes the game have a suggested play?
If that's not clear, let me give an example. We will say X is 4 (it's not really).
Let's say Team A is @ Team B, and my spread has team B laying 5. The books post -7. So my spread is suggesting A will cover, but it is not a big enough difference for me to call it a suggested play. If I only compare my spread to opening lines, this game will not factor into the w-l of my system. Now let's say team B gets a lot of action and becomes a 9 point favorite (9-5=4) so team A is now a suggested play if I compare to post-opening lines and this game would factor into the system's w-l record.
I currently use the highest line. So if the line ever gets to be X away from my spread (that I see, I don't have time to check all game's spread histories at dozens of books), I count it as a suggested play. However, I wonder if this is opening me up to bad/risky plays, since it is possible that things I am not accounting for in my models (coach firing, etc) move the line to where it should be. Well, maybe I just answered my own question, because if the line is where it should be then the play would be 50/50 and not that risky.
I'm guessing a good number of you come up with your own spreads for games, and then if the real one is off by your own predetermined trigger amount, you bet on it. So I'm wondering if you do that for opening lines only or not. Sorry for rambling and being repetitive, just looking for advice.