My understanding is that eigenfactor rates journals, not individual papers, so if somehow you get low-quality (whatever you want that to mean) papers into nature it has no independent way to realize that your specific paper is low quality. Also eigenfactor is biased towards favoring larger journals, which is not obviously a good thing. It would honestly be really cool if someone did page rank for individual papers. It seems like a much saner metric than anything that is currently used.
Oh good grief you’re right. This is doubly sad because using an ensemble metric for per-author eigenfactor seems like it would be tractable.
Carl Bergstrom is a smart guy so I suppose the practical implementation of the above must have some wrinkles, but with enough brute force it seems tractable. What I despise more than anything is the gaming that takes place for “impact factor”.
I do OK by standard metrics but would very much like to know where I stand by less easily gamed metrics of influence.
