Wouldn't any given approach rapidly lose efficacy as soon as its published? I would even guess that a paper being published means that, at the point the paper

Wouldn't any given approach rapidly lose efficacy as soon as its published? I would even guess that a paper being published means that, at the point the paper started to be written, its alpha had already decreased to zero. Otherwise the writers of the paper would still be using that approach. That's how it can appear to provide no value even if you extrapolate it back in time.