[ML4L]

[ QUOTE ]

In fact, I don't remember ever seeing ANY lines as far off as 6.5, except for in cases where injuries were obviously a major factor. That's what freaks me out so much here.

[/ QUOTE ]

I went through a stretch where I compared all lines to the predictor, and it was not uncommon on a weekend with a full slate of games for 5 or more games to be 10+ points off the predictor line. Some of these skews are injury-related as you said. Most others occur with big numbers, where it is unclear to me how well predictor accounts for the marginal decrease in the value of points as a line increases. But, that having been said, there are generally a couple games a week where, for example, predictor says Team A -1 and the line is Team A +10.5.

At a quick glance, here's one of that nature from bowl season:

NC State vs. S. Florida

Predictor: S. Fla -3
Line: NCSU -6.5

Some of that difference is probably "home field" for NCSU since the game is in Charlotte. But, it is still around a TD off...

Also, for whatever it's worth, from my informal research, it didn't appear that blindly playing picks that had value according to predictor was a solid strategy. Didn't really look at it in-depth or have a good sample size, however...

Just food for thought...

ML4L