mostly agree with your basic point about weaknesses in reliance on computers, except for generalization about "using a computer, of teams almost 20 years apart, is pretty inaccurate". I think only a computer is capable of doing something like this 20 years apart. It's the exact same thing human minds attempt to do when comparing teams .. whether 20 years or 1 year apart .... or even same season, who's conferences don't usually play each other ... but humans do it with a very limited set of data points, and lot more subjectivity and bias. Whereas computers at least are able to do so based completely on a set of facts consisting of virtually and unlimited # of relative data points (i.e. combinations) to support it. So I assume computers are equally as likely to over-rate a team as they are to under-rate one, especially when comparing teams this many years apart.
I think the main flaw in computer rankings is the tendency for strong conferences to cluster at top and weak conferences to cluster at bottom during a single season. So by end of season, I would agree, computers tend to self-perpetuate. Not necessarily wrong, and certainly not entirely wrong, just can get skewed at extremes. Example being, it just becomes hard for a truly great team in an average conference who may have started a little slow to move up against teams in the best conference who play only themselves the last 8 games of season. This is further why I tend to think computer rankings tend to be at their most credible around weeks 6 to 9 .. when prior season (including playoffs) and the current season's 3 non-conf games still carry fair amount of weight still.
Back to my main point ... I rely on Massey mainly for SOS, not for predictive value in any one game. I just wish some polls would give more objective consideration to SOS, and particularly the FCS selection committee. It has improved considerably since 2013 .. still have some work to do.
Not trying to be argumentative ... (it just comes naturally to me

) ... but do hope that clarifies my support and use of Massey.
Bookmarks