Re: Estimating Replacement Level: Sabermetrics perspective
Posted: Wed Aug 06, 2014 10:50 pm
My long-stated belief is that there is no particular utility in attempting to establish a replacement player value benchmark, but I think that this discussion has changed my mind. Still not as it relates to the value of VORPiness; rather, establishing a benchmark of the value of those under minimum contract offers seems to offer insight into the insightfulness of NBA franchises.
Am I the only person struck by the apparent implication of Daniel's "preseason" estimate of -1.95 and talkingpractice's -1.78? Isn't the suggestion that players, freely available from the "discard pile" at minimum salary and taken in the preseason, have an average productivity that is roughly equivalent to the 9th best player on the average roster?
I looked a bit closer at data for 2012-13 (87 players) and linked it up to Jeremias' xRAPM estimates, and got a total (pre- and in-season) average of -2.2. (Chris Andersen was not included, as he was an extreme outlier.) So, this is in line with Daniel's estimates from the 2009 to 2013 sample.
But interestingly, in the year chosen, the "ins" beat the "pres", -1.9 to -2.3. And if this is borne out in the larger sample, it offers possible separate comment on the perceptivity of the average NBA front office.
We talk about the draft being largely an exercise in dart throwing, but is it possible that the same is true for veterans available on minimum salary after multiple years of toil in the Association? With perceptions of value available in the preseason no more acute than what's in the midseason leftover pile?
Then finally, another perhaps interesting "split" is that in 2013, the "pres" that were centers and forwards were superior to their counterpart "ins", but the opposite was true with guards. To be checked...
Interesting stuff.
Am I the only person struck by the apparent implication of Daniel's "preseason" estimate of -1.95 and talkingpractice's -1.78? Isn't the suggestion that players, freely available from the "discard pile" at minimum salary and taken in the preseason, have an average productivity that is roughly equivalent to the 9th best player on the average roster?
I looked a bit closer at data for 2012-13 (87 players) and linked it up to Jeremias' xRAPM estimates, and got a total (pre- and in-season) average of -2.2. (Chris Andersen was not included, as he was an extreme outlier.) So, this is in line with Daniel's estimates from the 2009 to 2013 sample.
But interestingly, in the year chosen, the "ins" beat the "pres", -1.9 to -2.3. And if this is borne out in the larger sample, it offers possible separate comment on the perceptivity of the average NBA front office.
We talk about the draft being largely an exercise in dart throwing, but is it possible that the same is true for veterans available on minimum salary after multiple years of toil in the Association? With perceptions of value available in the preseason no more acute than what's in the midseason leftover pile?
Then finally, another perhaps interesting "split" is that in 2013, the "pres" that were centers and forwards were superior to their counterpart "ins", but the opposite was true with guards. To be checked...
Interesting stuff.