APBRmetrics

The discussion of the analysis of basketball through objective evidence, especially basketball statistics.
It is currently Sun Dec 21, 2014 4:29 pm

All times are UTC




Post new topic Reply to topic  [ 20 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: Emerging team strategy
PostPosted: Mon Jul 30, 2012 9:55 pm 
Offline

Joined: Thu Apr 14, 2011 11:10 pm
Posts: 2670
Putting Miami, OKC, Boston, San Antonio, LA Lakers and Chicago aside (at least for the moment), what is the basic strategy of the other contenders to contend? Offensive bias, balanced, defensive bias? Perimeter lead offense, balanced or using big men? Big time coach or not? What is the basic strategy of the next 10 teams and what is the frequency & range of those strategies? I will come back to this as I have time but if anyone wants to take one or more contenders to contend and provide this level of analysis and anything further (especially if you follow them closely), feel free. It would be helpful.


Top
 Profile  
 
PostPosted: Mon Jul 30, 2012 11:00 pm 
Offline

Joined: Thu Apr 14, 2011 11:10 pm
Posts: 2670
Nuggets are definitely offensive bias, perimeter lead with reputed or previous big time coach who currently has a near bottom third RAPM coaching impact estimate, which is not "big-time".

Clips are offensive bias, balanced attack without a big time coach (by current rep and neutral RAPM estimate).

Griz are defensive bias, big man lead (if Zac is healthy) with a big time coach (10th best estimated coaching RAPM amongst recent coaches).


Top
 Profile  
 
PostPosted: Tue Jul 31, 2012 3:20 pm 
Offline

Joined: Fri Apr 15, 2011 12:37 am
Posts: 227
I find it a little problematic to describe teams' strategy when it comes to coaching using RAPM when there's no evidence that teams use that in their decision-making. If someone were to hire Mike Fratello tomorrow, would they be hiring a 'big time' coach (2nd in Jeremias' coaching list), or the guy who's coaching in Russia right now and has an NBA career win % of 55?


Top
 Profile  
 
PostPosted: Tue Jul 31, 2012 10:09 pm 
Offline

Joined: Thu Apr 14, 2011 11:10 pm
Posts: 2670
Maybe "big-time" wasn't the best choice of words but I do think that head coaches are an important somewhat independent variable and a distinguishing criteria amongst teams. Some teams appear to not consider the coach to be that important, at least at this point, given the level of experience and performance they get with their hire. Other teams seems to make the Coach a key part of their strategy for contention (such as Adelman, on the court, in practice and a recruiting asset).

I would count Fratello as a big-time coach, at least initially. Maybe I would need to go back and review more than RAPM, but it was a place to start for an objective criteria for coaching quality.


Top
 Profile  
 
PostPosted: Wed Aug 01, 2012 3:03 pm 
Offline

Joined: Fri Apr 15, 2011 12:37 am
Posts: 227
I didn't mean to say that coaches shouldn't be included in evaluating a team so much as I (and I assume others) have no idea how teams decide what coach to hire or if they do so for the right reasons. There are 'players coaches' and there are 'hardliners'; there are guys good at managing vets and there are good teaching coaches; there are Xs and Os guys and there are guys who coach by feel to get the most out of their players. At least those are all things we hear in the media and presumably things that teams take into account when choosing a coach, but are any of these descriptions accurate? How different are coaches on either end of a given spectrum? If I get a 'teaching' coach for my veteran team, does the team suffer? I don't know if we have any of those answers and I don't know if the NBA does either.


Top
 Profile  
 
PostPosted: Wed Aug 01, 2012 6:30 pm 
Offline

Joined: Thu Apr 14, 2011 11:10 pm
Posts: 2670
Warriors are offensive bias (until they prove the defense), balanced attack without a big time coach IMO. I don't have the RAPM estimate but I'd put Jackson in the bottom 10 for sure of current coaches.

Jazz are offensive bias, big man led attack without a big time coach (-2.4 RAPM, one of the 15 lowest amongst recent coaches).

Blazers are offensive bias, probably a balanced attack without any coach yet.

Mavs are probably going to fairly balanced offense / defense (not sure if they will be better than average at either), probably a balanced attack with a coach that Cuban raves about RAPM estimates at -1.6 impact or bottom third.

Rockets will probably be either balanced or offensive biased, perimeter oriented without a big-time coach (McHale rates just above neutral on RAPM).

Suns will be offensive bias, balanced attack without a big-time coach.

Wolves will be offensive bias, big man led attack with a coach with a better rep than his -0.8 RAPM estimate.

There are other teams that could be done but of these 10, it is 6-7 offensive bias, 2-3 balanced, 1 defensive bias; 2 perimeter led, 5 balanced attacked, 3 big man led; 1 big-time coach and a couple of maybe / questionables. There seems like a pattern and worth knowing- offensive bias, balanced attack without a big-time coach.


Top
 Profile  
 
PostPosted: Thu Aug 02, 2012 8:30 pm 
Offline

Joined: Thu Apr 14, 2011 11:10 pm
Posts: 2670
Will any team have a 15 percentage point improvement in win% next season over last season? I am guessing not. Probably, few will go above a 10 percentage point improvement.


Top
 Profile  
 
PostPosted: Fri Aug 03, 2012 2:55 am 
Offline

Joined: Thu Apr 14, 2011 10:41 pm
Posts: 817
Location: Hotlanta
Warriors should be close to 0.500, which would be a big improvement technically.

_________________
The City: A Golden State Warriors-Centric NBA Blog
NBA WOWY: Find statistics of your favorite team with any arbitrary combination of players on or off the court.


Top
 Profile  
 
PostPosted: Fri Aug 03, 2012 8:41 am 
Offline

Joined: Mon Apr 18, 2011 10:09 am
Posts: 405
Crow, I wouldn't put much stock into the 10yr RAPM values. Without a proper aging/development curve there will be amplification and reduction seen in those values for the players as well as for the coaches. You have to keep in mind that the result of the regression suggests that all players had the same value during all of those 10 years, which is highly unlikely given the fact that the average player career in the NBA is shorter than those 10yrs.

It is a nice try to put a number on coaches, but I don't think it has much value. Just imagine a coach takes over a team which was improving just based on the fact young players are improving with more experience. The regression would now attribute the new coach as a positive while the old coach would probably be seen as negative. Or you have the other way around. You mentioned Carlisle, who had to deal with older players who were declining. Does the numbers just show that? Or Thibodeau, how much of that is just due to Derrick Rose, Joakim Noah, Taj Gibson getting better by default? Would he show the same improvement to a older team like the Mavericks? Is Greg Poppovich responsible for the decline of Richard Jefferson, Michael Finley or Antonio McDyess during their time on the Spurs? RAPM basically assumes that.

Well, without proper aging curves the whole +/- based analysis has a huge problem with longer timespans. There is a reason that prior informed RAPM gives better results than multiyear studies.

xkonk makes a good point about differences between coaches. If I have a young team, a coach who has a big strength in managing the egos of veterans will likely not have more success than a coach specialised in developing young players. For all we know Scott Brooks could have been a horrible coach on the 2008 to 2010 Lakers, while he is doing a fine job on the Thunder. You can't use Scott Skiles on a vet team, because at a certain point he will start to annoy the players. That's happened on the Bulls. The Bucks just taking their team apart and putting a new team together. Avery Johnson gave a great boost to the Mavericks early on, because he was a pretty good motivator combined with someone willing to listen to advanced stats stuff, but as his coaching in the finals 2006 and in the first round 2007 showed, he was not able to adjust properly during a series. On the other hand he was pretty good at drawing up working plays during a timeout. But his personality became also a problem with the older vets on the Mavericks. I suspect him to have a pretty good season with the Nets, but I can see him losing the players in the upcoming years. Well, it is obvious that you can't just put a number on the coaches and assume that the coach is "big time" or not. Let alone that this all is affected by the assistant coaches as well. The Lakers just hired Eddie Jordan for example, a really good X&O coach on offense with his Princeton Offense stuff. Does that make Mike Brown a better offensive coach?

_________________
http://bbmetrics.wordpress.com/


Top
 Profile  
 
PostPosted: Fri Aug 03, 2012 10:48 am 
Offline

Joined: Thu Apr 14, 2011 11:10 pm
Posts: 2670
After spending some time looking at / talking about the coaching RAPM values, I will back off from giving them as much weight and prominence. I didn't take them at face value previously or consider them solely sufficient but I wanted to at least look at them. Using numbers can be a part of evaluating them.

I still believe in treating the coaching decision as a major franchise decision that will play a key role in title contention though.

Will be interesting to see which emerging teams eventually pursue the Van Gundys (over their current lesser coach) and which teams the Van Gundys agree to accept.


Top
 Profile  
 
PostPosted: Wed Aug 08, 2012 3:33 pm 
Offline

Joined: Thu Apr 14, 2011 11:24 pm
Posts: 255
Mystic raises the objection that Jeremias' ratings are suspect inasmuch as they fail to incorporate player-aging effects. True enough. But I think that this effect, taken into account, would serve to compress the range of results. Tom Thibodeau, table-topping rating, must be biased upwards for having inherited a very young team.

The apparent fact of the matter is that "head coaches" (coaching staffs, really) aren't that important in the grand scheme of things. Furthermore, on average, this collective input into the NBA production function appears to subtracts value.

If you take J.E.'s numbers and weight by games played, the average coaching staff's RAPM is -0.5. And were one to average by possessions, my suspicion is that the rating would be a bit lower still (as slower pace is probably correlated with better coaching performance). And this (non-positive) result should be robust, given that it is based on 10 years' worth of data, no?

So, Crow, courage! I would basically take J.E.'s results at face value in such discussions. They are at least a good starting point for conversation and provide very useful perspective.

Consider the most recent case in point, where Portland after what was surely much due diligence hires Stotts (-1.9) to replace McMillan (-1.1). Stotts' previous head coaching experience: with Atlanta, where he replaced Lon Kruger (-1.4) and Milwaukee, where he replaced Porter (-0.7). Good stuff!

And what do we learn from the experts, commenting on the hire (http://espn.go.com/nba/story/_/id/82454 ... head-coach)?

"Terry is one of the elite offensive minds in the NBA, has extensive experience with multiple organizations and was instrumental in the Dallas Mavericks winning the 2011 NBA championship," general manager Neil Olshey said.

Stotts' Offensive RAPM -3.1. 10 year average Offensive RAPM -1.3.

Maybe he's improved of late.

Then:

"He has coached under Rick Carlisle the past four seasons. In a recent interview with The Oregonian, Carlisle said: "Of all the coaches presently in play, Terry is by far the best available. He's a professional and understands the process of NBA coaching and how to communicate and teach young players."

What is perhaps most interesting about this remark is not that it likely to be completely false, but that at the same time it might not be far from the truth.


Top
 Profile  
 
PostPosted: Wed Aug 08, 2012 4:12 pm 
Offline

Joined: Thu Apr 14, 2011 10:41 pm
Posts: 817
Location: Hotlanta
I'm just glad nobody hired Mike Malone off the Warriors. I have no idea what Mark Jackson's RAPM is, but I'm guessing it's a lot higher with Malone on the staff.

_________________
The City: A Golden State Warriors-Centric NBA Blog
NBA WOWY: Find statistics of your favorite team with any arbitrary combination of players on or off the court.


Top
 Profile  
 
PostPosted: Wed Aug 08, 2012 5:33 pm 
Offline

Joined: Fri Apr 15, 2011 12:02 am
Posts: 2425
Location: Asheville, NC
Quote:
If you take J.E.'s numbers and weight by games played, the average coaching staff's RAPM is -0.5.
I don't see how it could be other than zero.
Maybe the -0.5 is actually the "aging effect" : On average, a returning player is 0.1 pts/100 poss weaker than he was the previous season; and a lineup of such players is -0.5 relative to the previous year?

... weight by games coached?


Top
 Profile  
 
PostPosted: Wed Aug 08, 2012 6:13 pm 
Offline

Joined: Thu Apr 14, 2011 11:24 pm
Posts: 255
Mike, it is a reasonable prior that the average "head coach" RAPM should be approximately zero, but it would only be so by coincidence, and there is no reason that the actual result cannot diverge. The only (approximate) zero adding up constraint is for the contributions of all inputs, players and coaches both.

All the -0.5 (or whatever the possession-weighted value is) means is that a corresponding +0.5 is spread out across players (proportionately weighted, of course). Or at least this is my understanding.

It is also a reasonable prior, by the way, that average "head coach" Offensive and Defensive RAPMs would be zero, but reality is at even greater variance on this account. Average Offensive RAPM, as noted, was -1.3 and average Defensive RAPM was +0.8.

The question then is, why is it that the "average" NBA coach decreases his players' offensive efficiency more than he improves their defensive efficiency?

This (and many related questions) would be the basis of a very interesting econometric paper. And were one to present it at the MIT Conference, the author could be sure to never get a job in the NBA!


Top
 Profile  
 
PostPosted: Wed Aug 08, 2012 10:28 pm 
Offline

Joined: Fri Apr 15, 2011 8:28 am
Posts: 665
To clear up some confusion:
the regression moves player/coach values around to get the best fit around 0, while keeping the difference between the values intact.
Say you have 10 players in all, 9 players are -5 (compared to average), one is +5. The regression will give 9 players a 0, and one a +10.

This is because 9*0^2+1*10^2 < 10*5^2 (simplified, gets more complicated when there are different lambdas involved)

So basically, there's some shifting to find the best fit.

I realized this a couple of months ago and starting shifting the numbers, after doing the regression but before putting them on my site, so that
sum_over rating(i)*possessions(i)
equals 0, but it does not matter really. The coaching analysis was done before I did the shifting

_________________
http://stats-for-the-nba.appspot.com/


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 20 posts ]  Go to page 1, 2  Next

All times are UTC


Who is online

Users browsing this forum: Bing [Bot] and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group