Comments on: Predicting Round 3 With SPM

By: Neil Paine

Neil Paine — Wed, 20 May 2009 01:34:07 +0000

Anon: I won our little predict-off at APBRmetrics using it, but I haven't studied it extensively -- I've definitely maximized SPM's predictive power over the past few months, but we'll have to do an extensive retrodiction test this summer with various metrics (SPM, Win Shares, PER, Wins Produced, etc.). Schtevie: Going just by stats after the Billups trade is kinda what this exercise is trying to avoid -- the "true SPM skill" is a weighted average of the past 3 years, and it's supposed to represent impact independent of whether you're playing with Chauncey Billups or Allen Iverson (obviously you can't get completely away from context, but in theory that's what you want to do).

By: schtevie

schtevie — Tue, 19 May 2009 18:13:41 +0000

Neil, any interest in seeing what the Lakers vs. Nuggets series would look using both teams' stats post-A.I./Billups trade? It seems relevant.

By: Anon

Anon — Tue, 19 May 2009 15:43:12 +0000

I really like these SPM predictions Neil. Although some conspiracy theorists out there would suggest that the chances of the Cavs and Lakers each getting to the Finals is actually 100%, thanks to the "evil" commissioner Mr. Stern. :)

In your experience how do you think SPM compares to other box-score evaluation methods like ORtg/DRtg in making predictions?