two more factors not in performance ratings
- i should have added that i havent tried to capture defense
intangibles. reputation is subjective and frequently dated or
wrong. how and what to incorporate it in a linear weight system?
the promise of development of better data on defense would perhaps
allow for inclusion of one or two weights for defensive stops/FG%
allowed. that might be enough to supplement the existing weights for
defensive acts (def reb, steal, block, etc.). or if 82 games broke
out Opponent Production by player not just Position for all stats
categories you could give every player a truly fair half offense
rating, half defense rating by using the exact same performance
formula you use on offense for the opponent rather than try to add
those one or two extra weights to the offensive focused rating of our
also, i explain the failure of my and most systems to pass the shaq
at #1 or laugh test because these performance ratings are based on
regular season data and give a regular season rating. shaq's
undisputed crown as #1 is now only really #1 in playoff basketball.
in this regular season he isnt statistically the best. but if you
turned to last year's playoff data, the formula probably would kick
out #1 ranking for him.
actually i decided to run that data and kevin garnet just barely nips
shag in 02-03 playff performance rating. 29.04 for garnett, 28.64
for shaq. here is my top 20 from that dataset: