New DeanO Pod

Home for all your discussion of basketball statistical analysis.
rlee
Posts: 2011
Joined: Tue Sep 10, 2013 3:58 pm

New DeanO Pod

Post by rlee » Sat Sep 09, 2017 11:42 pm


schtevie
Posts: 357
Joined: Thu Apr 14, 2011 11:24 pm

Re: New DeanO Pod

Post by schtevie » Sun Sep 10, 2017 2:22 am

Thanks for the link, and it was good to hear old friend Dean O hold forth in what was generally a good interview.

A couple of points.

Why in the world didn't the interviewers ask who the 25 point, difference-making player was and for what season? Though it is not uncommon for Dean to guard his secrets, it seemed like he would have been happy to divulge this interesting piece of information - what for perspective would be half again as large as the largest in the xRAPM/RPM record.

Second, I remain chagrined, or make that highly unpersuaded, about his continued assertion as to the value of coaching. Gregg Popovich has proven himself again and again to be a very fine human being, but that his contribution on the court is 12 wins simply strains credulity. If that were to be the case, it is equivalent to saying (attributing this effect to the entirety of his career) that with an average coach (i.e. one imparting zero wins) that his rosters would only have averaged 46.4 wins (what is approximately only 2 points per 100 possession margin of superiority). Really? Robinson, Duncan, Ginobili, Leonard, and besting their opponents by such a slim margin, and not 47 wins per year?

Anyways, good stuff.

Crow
Posts: 5954
Joined: Thu Apr 14, 2011 11:10 pm

Re: New DeanO Pod

Post by Crow » Mon Sep 11, 2017 9:08 pm

Nice enough START. Didn't say much NEW though imo. The estimates of coach / superstar value don't mean much to me without a thorough discussion of the estimate methodology which was not provided.

roland_beech
Posts: 9
Joined: Tue Jun 14, 2011 4:06 pm

Re: New DeanO Pod

Post by roland_beech » Thu Sep 14, 2017 4:58 pm

Ha, fun with Dean!

...as far as the value of coaching goes, it's a shame there's still believers in Berri's "coaches are deck chairs" line of thought...the challenge is it's difficult to put actual numbers on coaching value since they are so integrated in all that happens (and while I have great admiration for much of what Jeremias has created, Coach Rapm, as I've said elsewhere, is garbage...)

Dean is very familiar with these two bookend examples of the 'value of coaching' --

The '04-05 Nuggets were 17-25 when George Karl took over...and went 32-8 the rest of the way
...is it possible George and his staff might have added 12 wins in HALF a season? sure...

Then after a 57 win season in '12-13 the Nuggets surprisingly fired George, and new GM Tim Connelly hired Brian Shaw as his man...36 wins followed, a massive drop off of 21 wins...there was some mitigating circumstances such as Gallinari missing the season with injury...but I'm personally of the belief George could have manufactured many more wins than that had he stayed...

It's easy to say Popovich had 'great players' but of course he has much to do with them becoming great, no? Kawhi might not have become the player he is today had he been drafted somewhere else and coached by different coaches (we can all probably think of places it wouldn't have gone so well for him...)

Crow
Posts: 5954
Joined: Thu Apr 14, 2011 11:10 pm

Re: New DeanO Pod

Post by Crow » Thu Sep 14, 2017 10:14 pm

The measure could have several meanings or none, in principle or to any person, but comparing actual wins to pythagorean expected wins (and judging split seasons as a whole) Karl's actual win total exceeded 10 times, fell short 7 times and tied 10 times. Mildly positive but pretty scattered.

schtevie
Posts: 357
Joined: Thu Apr 14, 2011 11:24 pm

Re: New DeanO Pod

Post by schtevie » Fri Sep 15, 2017 1:34 am

roland_beech wrote:...as far as the value of coaching goes, it's a shame there's still believers in Berri's "coaches are deck chairs" line of thought...the challenge is it's difficult to put actual numbers on coaching value since they are so integrated in all that happens (and while I have great admiration for much of what Jeremias has created, Coach Rapm, as I've said elsewhere, is garbage...)
Hmmm. I don't know (or particularly care) what Berri's views are on the matter, but as with any analytical issue in basketball (or sports generally) you need to start with an accounting framework, and in the instance it is straightforward: basketball is a zero-sum game, there are points for and points against, netting to zero. So, if you want to ascribe value (net points) to coaches/coaching, you are simultaneously and necessarily subtracting it from the obvious others, the players. You can disagree with Jeremias' estimates, but not the framework. Unless you are prepared to present another?
roland_beech wrote:Dean is very familiar with these two bookend examples of the 'value of coaching' --

The '04-05 Nuggets were 17-25 when George Karl took over...and went 32-8 the rest of the way
...is it possible George and his staff might have added 12 wins in HALF a season? sure...
Now I think I'm being trolled...Really? George Karl? George Bleeping Karl? George "I coached a team of NBA greats to 6th place at the 2002 World Championships played in the US and tried to place the blame for the result on Paul Pierce" Karl? That guy? 12 wins in HALF a season?

A quick look at the preceding season results, and the circumstances of the one in question, tells a bit of a different story, or should I say suggests a different context. The Nuggets were 43-39 in 2003-04 under Jeff Bzdelik. Then they began a bit slowly with a relatively similar roster in 2004-05 (13-15) what saw him lose his job, and he was immediately followed by Michael Cooper who went 4-10. And it was into this situation of organizational "chaos" that Karl came in and performed very well (32-8). An 80% win percentage, formidable!

But now we scratch a bit below the surface. This record was composed of a superlative home performance 19-1 (where Denver is known for its special home advantage, as I recall, owing to altitude) and a more ordinary 13-7 on the road. But the quality of its competition was below average. Looking at the Predicted Wins of each subset, on average, their home opponents were 37.5 wins and the away opponents were 39.1 wins. And then the year following, again with core roster stability, they were a 44-38 team, essentially identical to two year's prior.

So, what would a prudent observer make of this? That George Karl was a super-fantastic great coach in partial year 2004-05, followed by a terrible horrible coach the next year, causing his stellar team to deteriorate from an 80% win clip to 54%?

Or would it be that he inherited a 54% team that was mired in a bit of a topsy-turvy situation and had a pretty good run of luck (and an advantageous schedule)?

Hmmmm.....
roland_beech wrote:Then after a 57 win season in '12-13 the Nuggets surprisingly fired George, and new GM Tim Connelly hired Brian Shaw as his man...36 wins followed, a massive drop off of 21 wins...there was some mitigating circumstances such as Gallinari missing the season with injury...but I'm personally of the belief George could have manufactured many more wins than that had he stayed...

And here we go again. Not mentioned is a rather significant mitigating circumstance, specifically Andre Iguodala leaving the team. I will leave it to the interested observer to do the plus-minus math (independent of the ordinary factor of an initial expected debit associated with coaching turnover) to determine how unexpected the result of Karl exiting was.
roland_beech wrote:It's easy to say Popovich had 'great players' but of course he has much to do with them becoming great, no? Kawhi might not have become the player he is today had he been drafted somewhere else and coached by different coaches (we can all probably think of places it wouldn't have gone so well for him...)
I've been meaning to write a little piece about the effect of coaching (and general managing) in general, with a special focus on Gregg Popovich, but let me anticipate my eventual remarks by answering the question posed in the negative (with a possible asterisk). It is entirely unreasonable to attribute the inherited skill of David Robinson to Gregg Popovich; indeed I think no one would do so. Similarly, Tim Duncan, entered the NBA with the nickname "Mr. Fundamental" and who was consensually, make that unanimously, regarded as a sure-fire, great, great player, with a well-recognized, well-rounded game. To ascribe his NBA performance to "coaching", well, the burden of proof is on those who would take that point of view. Then Manu Ginobili, here is a talent that was not anticipated (as he was apparently only worth the derisory cost of a bottom 2nd round pick) who turned out to blossom in Europe, becoming a well-recognized superstar, who then entered the NBA in the prime of his career. Gregg Popovich had nothing to do with any of that.

Add up the contributions of these three players, and there is really nothing exceptional about the residual, in terms of team performance for many years of the Popovich dynasty. That is the simple reality. Whatever value one might wish to impute to Gregg Popovich at this point, the adding up constraint then obliges one to subtract the same amount from the performances of other players. But then one quickly careens into the land of the absurd. How bad could the supporting cast plausibly be?

But there is an exception to this story, and that is Kawhi Leonard. Here is a guy, whose contributions are instrumental to maintaining the same story into the more recent years, and here is one instance where one might choose to impute critical Popovichian influence in his developement.

But really? Is this the thin reed upon which the hagiography rests? One most exceedingly unlikely (given his low draft location) superstar, owing essentially all his superness to a coach? As opposed to all the other, rather conventional stories that well explain the team's success?

More on this at a later date, perhaps.

Crow
Posts: 5954
Joined: Thu Apr 14, 2011 11:10 pm

Re: New DeanO Pod

Post by Crow » Fri Sep 15, 2017 2:05 am

Will Karl try to get one more coaching job? Is there a real chance? My first thought is no. Perhaps the Suns if the term was short, money low. There are people who might want to go with him.

Rd11490
Posts: 111
Joined: Mon Sep 29, 2014 4:54 am

Re: New DeanO Pod

Post by Rd11490 » Fri Sep 15, 2017 2:10 am

Will Karl try to get one more coaching job? Is there a real chance? My first thought is no. Perhaps the Suns if the term was short, money low. There are people who would want to go with him.
If he had any intentions of returning, they went out the window when he wrote his book.

Crow
Posts: 5954
Joined: Thu Apr 14, 2011 11:10 pm

Re: New DeanO Pod

Post by Crow » Fri Sep 15, 2017 2:14 am

Probably but as an "interim coach" with further apology at or before the announcement, it might pass. Even if it is impossible, doesn't mean he won't try. I didn't think he'd get the Kings job but he did.

hoopstudies
Posts: 15
Joined: Thu Apr 21, 2011 3:03 pm

Re: New DeanO Pod

Post by hoopstudies » Fri Sep 29, 2017 6:43 pm

What's remarkable to me is that David Berri's method for evaluating coaches was pretty good. But his conclusions from that were weak.

What he did was to make a projection of player performance based on age, experience, etc. plus whether a certain coach was with them or not. That basic model is a good idea. He was looking over how many players got better with a coach present/absent. Looking over a lot of players, this should tell you something. It did show Pop and Phil being really good coaches, maximizing a lot of what they had as players. Players played above their base expectations when those coaches were there.

But Dave truncated his list. George Karl was pretty significant, maybe not at 95%, but he did get a lot out of most of his players. I studied it when I was there with him. If you have a good coach, he will get more out of the players you bring in. That makes it easier to make trades. Were there players that he wouldn't maximize? Yes, and we figured out a bit of what kinds of players that would be. The not-so-smart ones. I didn't know that when I first got there.

Dave also got criticized for his player method. Not wanting to open that debate, but you can develop the same methodology with other player metrics and see what it says...

There is a flaw to this method, though. It doesn't tell you why a coach or his staff had success. What did they do or say to make the change? You could probably narrow it down to offense or defense and maybe a little more detailed than that, but it's hard.

Further, you can't really know whether it's the head coach or assistants that consistently go along with the coach. There is no division of credit there. This makes it really hard to hire an assistant coach from a certain organization.

So, whereas I think we can identify good/bad coaching staffs decently, I think it's hard to make a projection for coaches who haven't had a chance... which is probably where big advantages can be gained.


schtevie
Posts: 357
Joined: Thu Apr 14, 2011 11:24 pm

Re: New DeanO Pod

Post by schtevie » Mon Oct 02, 2017 4:47 pm

Hey, Dean, if you're still around, how about an answer to the above question: who was the 25 point, difference-making player you referred to in the podcast and for what season?

hoopstudies
Posts: 15
Joined: Thu Apr 21, 2011 3:03 pm

Re: New DeanO Pod

Post by hoopstudies » Tue Oct 03, 2017 1:33 am

Where was that? I don't remember that.

Rd11490
Posts: 111
Joined: Mon Sep 29, 2014 4:54 am

Re: New DeanO Pod

Post by Rd11490 » Tue Oct 03, 2017 2:40 am

At 25:09 you say "But i've seen things as high as 25 that's hard to discredit" in a discussion about how many wins a super star is worth.

hoopstudies
Posts: 15
Joined: Thu Apr 21, 2011 3:03 pm

Re: New DeanO Pod

Post by hoopstudies » Tue Oct 03, 2017 7:39 pm

I was dubious of it, as I said.

Some of the estimates for Westbrook this year were in that 25+ wins added range. I toned it back to say that some methods were estimating 25 because I think it was actually higher, which was really wrong. BPM/VORP were the main ones too high and probably flawed by cross terms. Not terrible methods in general but flawed for some of the extremes. Shows up for other years like Lebron's great ones, I believe.

I don't usually put my work in terms of wins as much as points. But I tend to find best seasons in the 15-20 win range.

Post Reply