quote: (post)

Originally posted by The Enigma
The reason why we verbalize, character A > character B, so A gets B's feats, is because B has access to some set of better demonstrations than character A. Often times character A has lots of content, but still somehow has inferior feats despite being more potent from scaling, which suggests a lack of precision with which feats are opportunely mapped onto characters predicated on their internally consistent puissance.

This engenders a perverse incentive to endeavor to pick out the outliers - characters who have eccentrically high exhibitions relative to their low placement on the scaling chain. So if one finds an arbitrary TOR or OCW character with an outlier incredible feat, one can then scale characters massively above them, even if they haven't authentically come proximate to doing the act themselves.

Yet this is conventionally not how one goes about modeling data. Customarily with inconsistently erratic and disparate data, there is an endeavor to take into account the totality of the evidence through some marginally best fit function approximation through the data points. One does not take the absolute highest local maximum on the left cessation of the graph and then draw a straight line that doesn't genuinely fit the evidence.

Yet, just taking the unweighted average of feats doesn't tell the consummate picture - feats are not very congruously calibrated, except for perhaps a general trend towards more intuitively potent characters having better exhibitions. Yet the medium disputably plays as immensely colossal a role, if not more preponderant.

Some scarcely more precise and rigorous system that takes into account all the consequential variables may be indispensable.