So while I was waiting for Gray to respond, I thought I'd check to see if there was any reliability to the SOS at the beginning of the season. I looked at the SOS for week 17, both at week 17 and at week 1, and ran correlations for each position. I looked at the SOS for 2009 to 2011. The results were interesting. Almost all were positive, some much more than others. The results are below.
Code:
2009 2010 2011 AverageQB .28 .17 .24 .23RB .76 .52 .62 .63WR .21 .23 .38 .24TE -.03 -.06 .15 .02K .37 .44 .07 .29TD .81 .40 .35 .52
That suggests rushing defenses and poor offenses are somewhat predictable at the beginning of the season. Passing and kicking defenses are harder to predict and trying to predict defenses against TEs is nearly impossible.