This is cool. But, why don't we all just agree to scrap this composite bullshit? ESPN is terrible and distorts the rankings too much. Just list the 247 scores, and while not perfect, it is better than including the faggots at ESPN who fuck it all up completely.
I would, it's just that the composite is the basis for Bud Elliott's foundational research on the whole "need blue chips on at least half your roster to win the title".
If we're tracking how we rate on his metric then we have to use the composite. If we're using the metric of our choosing then we're basically acknowledging the ratio is BS because it's based on a flawed premise anyway and we can stop worrying about it.
I'm good either way. Personally I much prefer the 247 ratings but obviously they like our? guys better so duh.
I can justify why you can use 247 rankings while still using Bud Elliott's metric that was founded on using composite by writing a huge @Tequilla special of "I was told there would be no maff." or you can just believe me.
TLDR version is that there's "noise" in a specific subset of the data which applies to certain regions/teams(us?) which are themselves outliers and therefore don't have much impact on the correlation value of the overall population/metric. This is not an unusual adjustment for a subset of data that has a measurement error for a population with an established correlation.
TTLDR Statisticians dump outliers and adjust measurements all the tim. It's cul.