Not the entire history in the NFL, but from what I can remember its been above 75% since like 02, when I first started looking into this, but dont quote me on that number. All I can say is that it is a very very high percentage. I didnt track the spread, such at -3 or below, I just kept track of all games.
For example, I got 41 games that had lines on them this weekend in college football. Here was the breakdown on them...
There was 1 push, so we will make that to 40 total games then
Out of 40 games, 35 of those games the spread did not matter. If the fav won, they also covered the spread, and if the dog covered, they also won SU.
So out of 40 games, the spread came into effect in only 5 games, thats only 12.5% of games where that mattered. (hope i did my math correctly)
Not saying this is an easy way of winning games, but picking the SU winner and hoping for them to cover, you have a good chance they will.
They put spreads out there to balance out action. The books will make money no matter what the outcome is in the longrun.