I checked on the 1 unit bets (bet on the team if it lost last week by 10 points or more ATS---I did not exclude games where both teams were 1 unit bets so I under state the results a very little bit
weeks 2 thru and including 16, no week 17 and no playoffs
1996 thru 2009
raw numbers
won 697
lost 659
51.4% win rate
sorry to break the bubble
didn't check the 2 or 3 unit bets
Some of those could be cancelled out because of a team coming of a win and a buy. Maybe it only works when the circumstances are only without buy teams.
I did notice that there more wins from teams off buys weather they won or not, which is not all that earthshocking to hear, but the "raw" data is only useful similar to an encyclopedia. You've got to read about the circumstances before comparing and contrasting anything, sorry to break your bubble.