Around the league- some quick comments

[Crow]

Looking at a week ago xRAPM values, 62% of those +2 or better on overall xRAPM played dominantly at that PF or C position. 11% at PG, 27% at the wing spots. That is a pretty strong tilt among the best impact players toward the bigs. (Same thing that also happens with the wins produced metric.)

[DSMok1]

Looking at a week ago xRAPM values, 62% of those +2 or better on overall xRAPM played dominantly at that PF or C position. 11% at PG, 27% at the wing spots. That is a pretty strong tilt among the best impact players toward the bigs. (Same thing that also happens with the wins produced metric.)

Did that happen before adding in the box score prior? Not sure.

[bbstats]

Yes, vanilla RAPM seemed very keen on PFs IIRC.

But now that height is the most important statistical factor for the prior...I don't really know what to think.

[Crow]

For RAPM without priors about 50% of the +2 or better impact players were bigs for 2009, 2010. 2011 (+/- 2 %pts). So the bigs were still overrepresented compared to wings and PGs; but with xRAPM, the bigs have almost a 25% higher representation than in RAPM without priors (62% compared to around 50%).

P.S. On this season's xRAPM 70% of the +2 and better players have 5+ seasons of prior experience.

[Crow]

I looked at the weighted average team stats of the best 12 lineups on team +/- per 100 possessions that were used at least 150 minutes as of about 10 days ago. Here is what they looked like compared to their opponents:

Pace 92
FG% 4.4 %pts better
3P 0.5 more per game
3PA 1.8 less
3P% 5.8 %pts better
eFG% 4.6 %pts better
FT 0.8 more
FTA 1.2 more
FT% tiny bit lower
PTS +11.8 per 100 posessions.

Despite these huge advantages only 6 of these top 12 performing big minute lineups had been used over 200 minutes at time of this data. I would use these lineups far more than 7 minutes per game in most or all cases.

Memphis had the 3rd best lineup and used it the 2nd most. But are now reportedly talking about a trade that would break this powerful weapon up... You have to go back to 2010-11 to find a really good performing lineup without Gay tested over 50 minutes. That had Sam Young. Maybe they can keep a strong heaviest minute lineup without Gay but it would be a leap of faith and research.

[Crow]

If you look at the average regular season offensive and defensive efficiencies for the last 6 titlewinners it is about 110 on offense and 104 on defense (rounded). 5 of the 6 best teams this season meet both critieria or are fairly close. (Memphis fails the offensive criteria by a lot.)

Only one of these teams has an own FT/FGA ratio better than the weakest of the 6 recent titlewinners- OKC. But it is an open question whether that it critical.

[talkingpractice]

I thought this might be fun. Finding the guys who have done the most damage thus far this season (ie produced the most negative value for their teams). This is of course a combo of both the level of atrociousness and the volume of that atrocious play. Here are the 6 guys who have really distanced themselves thus far, based on IPV and their MP through last night.

Norris Cole
Donald Sloan
Will Barton
Thomas Robinson
Austin Rivers
Byron Mullens

The Norris Cole situation is the really interesting one, I think. I can understand why Rivers and Robinson get minutes, but the Norris Cole situation is just baffling. Last year they finally got him out of the rotation by playoff time, but this year he's storming away again on pace for 1400 minutes of -5 basketball, as Chalmers has just 900 mp to Cole's 600. Perhaps the 10 threes that Chalmers hit the other night will be the catalyst to make this change a bit (Cole got only 6 minutes of run last night). I'm barely kidding when I say this, but for purposes of wagering on the Heat, the expected number of minutes that Cole is going to get each night is more important than just about anything. It's about a whole point difference for Chalmers/Cole to go 36mp/12mp as opposed to 24mp/24mp in a given game.

[Crow]

Thanks for noting this. Cole is also estimated as 2nd worst impact player in entire league for last season and this season on xRAPM.

I would guess that there is a good chance that the Heat ignore or dismiss revised public versions of APM. But his winshares per 48 is dreadful too. I wonder what they are doing to try to change his performance or replace him.


Apparently a lot of the guys paid to work on basketball analysis full-time and almost all their bosses and almost all coaches dismiss / ignore the revised public versions of APM or the approach altogether. At least that is what most say when I've seen them comment on the topic. I have never been that impressed or fully satisified with their short answer dismissals of it. Yes there is some amount of error to the estimates and it should be considered carefully in the context of other data (blah, blah, blah) but the impression I get is that most insiders do not give it any meaningfully role in their analysis. At least they won't admit to it.

[Crow]

Cole has the 2nd lowest win% and +/- on the court of any Heat player playing more than tiny minutes.

Lewis is worst on both. 3rd worst on XRAPM at -3.1. Allen only slightly better. 7 Heat players worse than -1.5. Possibly the data is overstretched negative for them to counterbalance the two big stars.

[Crow]

Celtics currently estimated to have 1 chance in 500 to win the title according to ESPN, less than that or none on B-R.

[v-zero]

The betting market on Betfair has them at roughly a 1/50 chance. With the tightened rotations of the playoffs I have them as nearer 1/50 than 1/500...

[talkingpractice]

What's everyone think of this rumored Rudy Gay to Toronto trade?

Allegedly it's Ed Davis as the main piece, they sell high on Calderon and ship him to Team #3 (uncertain as of yet) to shave the contract and play Lowry, and they take back Darrell Arthur along with Gay.

I think that it's a steal from a Memphis perspective if they get Davis in return for Gay and Arthur (given the huge salary savings), unless there are tons of other financial considerations that we're not aware of. That doesn't necessarily imply that it's a bad trade for the other parties, though I don't completely understand Toronto's motives...

[Crow]

On Boston's odds to win a title before the Rondo injury, I would have pegged it at clearly less than 150/1 and maybe worse than 250/1. They were recently listed at 28th on offensive efficency. Three spots worse than last season though 0.7 pts better.

2nd highest team rate of taking 10-15 foot shots. 5th highest from 16-23 feet. 3rd lowest on 3 pt attempts. That much is a very poor strategic distribution. But they had the 10th most at the rim. Terry at by far his lowest rate of taking 3s per minute (33% lower than his previous lowest). Didn't really benefit from that pick and pop game that some said he would exploit well. His 3 pt FG% is the second lowest of career other than his rookie season and well below his career average.

[Crow]

Before his surge of the last 2 weeks Rubio had the second worst career eFG% of any player with 500 or more FGAs since 2000 (Skita had the worst). He has had 1-2 week bursts of good shooting before. Plenty of incredibly awful shooting weeks. Has he or will he establish a better long-term level? Time will slowly add information on that question. If he stays below 40% that will be at least a modest issue.

[Crow]

Memphis had the 3rd best lineup and used it the 2nd most. But are now reportedly talking about a trade that would break this powerful weapon up... You have to go back to 2010-11 to find a really good performing lineup without Gay tested over 50 minutes. That had Sam Young. Maybe they can keep a strong heaviest minute lineup without Gay but it would be a leap of faith and research.

http://www.basketball-reference.com/pla ... y=diff_pts

Starters with Gay were +11 per 100 possessions. Starters with Prince are -4 per 100 so far.
Looks like lineups matter.