One last question. Doesn't the model suffer from Omitted Variable Bias, (as does any basketball regression) since there is no proxy for individual defense. In that case, how can we take this as a reliable metric, if the model is misspecified.

And, how about using team defense as a proxy (albeit a very weak one) for individual defense.

I apologize if it appears I am questioning the validity of the model, I just don't get how you can run a reliable regression if one of the most important variables is missing. I can't imagine you'd get an R-squared above .65 without some sort of a proxy for defense.

And look, I'm actually pretty happy Russell ranks that high (PER ranks him 89th all-time!!). It's sort of remarkable considering that the "intangible" things he did best aren't even picked up in today's boxscore, much less an incomplete, roughly-estimated version from almost 50 years ago (we have to estimate his blocks, for goodness' sake!). Basically, all the stuff Wilt excelled at was tracked back then; the stuff Russell excelled at wasn't.

Besides, I think the gap between the 2 as players has been overstated as the years have piled on (and I say this as a lifelong Celtics fan, btw). It's sort of like the thinking that causes Terry Bradshaw to be considered better than somebody like Fran Tarkenton because of a disparity in championships. I'm not necessarily saying Chamberlain was that much better, or even better at all, but you have to respect both players for their accomplishments and try to avoid the mythmaking that's been built up over the ensuing half-century.

And as a litmus test, any measure that rates Wilt 2nd and Bill Russell 29th has to be viewed suspiciously.

1) You Think You Could Make An All-Time Statistical +/- for Every Player for the Play-Offs

and

2) A All-Time Statistical +/- for a Player comparing "His Teamates Statistical +/-s" and how much Impact that Player had v.s his Teamates.

That would be asoume....

Greetings

The guy who appears to be most overrated is Nate McMillan, or John Salley, or Bobby Jones.
Most underrated : Kevin Willis, then Duckworth.

I think Statistical +/- doesn't consider the defensive strength of the team, but tries to make up for that with high weights on individual D-stats (Blk and Stl)? How else to explain Kirilenko over Duncan and Garnett? Or this sequence:

5.16 Doc Rivers
5.13 Alvin Robertson
5.10 Dominique Wilkins
5.09 Billy Cunningham
4.97 Scottie Pippen
4.87 Elton Brand

Or Mr Salley's neighborhood:
3.28 John Salley
3.22 Chris Mullin
3.20 Clyde Lovellette
3.18 Mark Aguirre
3.16 Maurice Cheeks
3.10 Kevin McHale
3.10 Kevin Johnson
3.04 Jack Sikma

Big Shot Rob among Hall-of-Famers:
2.44 Isiah Thomas
2.44 Robert Horry
2.44 HotRod Williams
2.38 Rashard Lewis
2.36 Paul Arizin
2.35 Lamar Odom
2.33 Lenny Wilkens
...
2.18 Dikembe Mutombo
2.18 Willis Reed
2.16 Doug Christie
2.13 Danny Manning
2.12 Zelmo Beaty
...
1.76 Derrick McKey
1.75 Moses Malone
1.74 Mark Price
...
1.57 Michael Cooper
1.56 Dave Bing
1.55 Sidney Wicks
1.54 Elvin Hayes
...
1.13 Spencer Haywood
1.13 Mario Elie
1.11 Dave DeBusschere
...
All the early-ABA guys are overrated, but that's another can o'worms.