re: Different way to look at returning experience (LSU #1)

Very cool article and method. And even though he acknowledges that his OL metric is flawed, it seems like that should've been dealt with already (I'm greedy ). Returning three of last year's starters seems magnitudes more valuable than returning a three year starter. I would think his simple proposed solution of using only the previous year's starts would be sufficient, but you could consider weighting by recency (2x for previous year and 1x for 2+ years prior).

I think that considering the consensus that OL play is largely influenced by experience and highly correlated with offensive success, it's imperative to devise a metric for OL experience that properly reflects this correlation. With other stuff you don't wanna fish and fiddle, but there's no way OL experience doesn't matter.