• Changing RCF's index page, please click on "Forums" to access the forums.

2014 +/-

Do Not Sell My Personal Information

Pioneer10

Come home Sideshow Bob
Joined
Apr 18, 2005
Messages
16,677
Reaction score
16,703
Points
123
Just to spit in Rich's eye :chuckles: (it's funny but +/- is the one stat you hate but consistently supports your positions better then any other stat) and to throw out some love to AV:

NBA.com net +/-. Raw unadjusted numbers

ESPN Real +/- (this is essentially adjusted +/-: regression analysis to make sure players like Kyrie don't get screwed by having to play with a guy like Bynum): http://espn.go.com/nba/story/_/id/10740818/introducing-real-plus-minus

Gotbuckets Regularized. This is adjusted +/- but with a bayesian a priori adjustement to throw out outliers. This reduces confidence intervals and better predicts future performance. http://www.gotbuckets.com/what-is-apm/

NBA.comESPNGotbuckets RAPM
Varejao1.14.062.1
Deng-12.532.86
Dellavedova0.71.37-0.3
CJ Miles0.91.111.1
Dion-0.6-0.13-0.12
Irving-2.9-0.71-1.12
Hawes-3.2-1.31-2.8
Zeller-0.4-1.55-1.14
Thompson-3.1-2.61-2.19
Jack-3.3-3.38-3.05
Gee-4.1-4.05-3.81
Bennett-1.8-5.33-3.34

<tbody>
</tbody>

For next year, based on my analysis of these stats which fit nicely with my biases lol:

-The ESPN and Gotbuckets stats are confusing with Deng and Hawes as they include a lot of their non-Cavs games

-I'd be wary of trading AV: his trade value is likely not going to get you back anywhere close to equivalent production. Especially if we are trying to improve the loser mentality that infects this team at times

-I would strongly consider resigning Miles

-Cut Gee and if Mike Brown had a brain he would have played Miles at SF when the guy was healthy instead of Clark, Gee, and crap

-I would let Hawes go: his defensive troubles do not show up in the traditional boxscores but watching games and +/-, he is a significant defensive liability and should not be paired with another defensive liability in Irving

-Trade Jack for anything you can get and give all his minutes to Delly/Miles

-I would actively try to trade Tristan as well while he is on his attractive rookie deal: his game like Hawes does not complement Irving and Waiters

-The Irving conundrum comes out again: incredibly talented player who doesn't seen to move the meter as much as he should. I think we're stuck with them and that means trying to adjust the roster to maximize his strengths and hide his weaknesses. In my mind that means finding a PF who can open up the floor for him, a strong defensive center who can protect the rim and cover up his weakness (ala Bogut and Curry), and getting even just a replacement level player at the SF spot
 
I'm curious if you've discovered the actual formula that ESPN is using to calculate RPM? I haven't been able to find it and have gotten the impression they are concealing it. Which to me would automatically disqualify its usefulness. Anything that can't be verified and tested independently is just a black box of meaninglessness. (Of course raw +/- is proven useless so there's no point in even looking at that.)

The gotbuckets RAPM and SWAgR is the most interesting of the three (pending understanding more about what RPM is). That's a nice find. I appreciate that they explain, show, and publish their work. I haven't had time to pour through it yet but at least the info is there so that I (or others) could verify it if we wanted. I also like the way they walk back some of the more questionable "adjustments" in Rosenbaum's APM.

And of course, oh yeah, whatever data you used, I mostly agree with your conclusions. While I'm unsure as to the precision of gotbuckets RAPM -- and quibble here and there with a few things -- most of it seems to be in the general ballpark and cross-checks moderately well with other methods (btw, the box score models like Win Shares and Wins Produced do pick up Hawes defensive failings loud and clear).
 
Agree on Thompson and Hawes.
 
some interesting rankings from hoopstatsdata who measures efficiency and efficiency differential.

Cavs backcourt ranked 21st pre all star. post all star break they rank top 10. Last 10 games the back court is ranked 5th. and so far after 4 games number 1 in the month of april.
 
If Miles was starting at the 3 with Irving and Waiters all year we might not trade for Deng.
 
Very interesting. Thanks for all the work you put into this. I wasn't surprised at Andy's stats. He makes Dion, Kyrie and Jack look good while he's on the floor. Sometimes I have wondered if Kyrie and Dion ask Brown for more minutes to play with him because he does that well for them. Just my opinion, Dion actually had a more productive line up when he played bench with Andy, Delly and CJ if you look at the player line-ups.
 

Rubber Rim Job Podcast Video

Episode 3-15: "Cavs Survive and Advance"

Rubber Rim Job Podcast Spotify

Episode 3:15: Cavs Survive and Advance
Top