Several years ago I participated in an office pool that picked against the spread. I spent a lot of time researching each team. What did they score, what did their opponents score, what did they give up, etc. The large pt spreads were often easy to predict.
So, the question is can UDub cover a 17 pt spread against a decent opponent? According to last year, yes they can. So far this season we've seen then struggle against Utah and ASU. Well, last year they barely beat Utah by 3 pts and they lost to ASU, so maybe those are bad match-ups for them.
And if you look at last season, BYU is certainly capable of losing by 17 or more. It's still early this season, but the 17 pt line, after research, isn't out of line.
Back to the original pt, it was easy money/choosing whenever perception created too large of a spread. If a team rarely ever beats any team by 20+ points and they are favored by 20+, and the opponent rarely loses by 20+, take the points. You'll win that bet a lot more often than not. I built a solidly winning record against the spread by finding those games.