I'm a bit confused. The basic assumption here seems to be that the big
sportsbooks (e.g., Pinnacle) are right more frequently than the small ones. So whenever you see a line change in a big sports book, spend say 1% of your bankroll to pound any small sportsbook that disagree with Pinnacle. If Pinnacle is more accurate, then you'll make money on average doing this.
But if all of this is true, couldn't you automate the entire process? Write some software to automatically detect big Pinnacle line changes, and then automatically place bets in small books? You could even do this for sports that you know absolutely nothing about, and still win big (again, assuming that Pinnacle is more accurate.) Since you've automated things, you could bet on literally thousands of events a year, thus reducing your variance to almost nothing and making tons of money.
Somehow, this sounds too easy and good to be true. What is the catch?