Kings/Beech

Home for all your discussion of basketball statistical analysis.
Crow
Posts: 6250
Joined: Thu Apr 14, 2011 11:10 pm

Re: Kings/Beech

Post by Crow » Sun Apr 23, 2017 11:03 am

Insiders frequently dismiss / laugh at web draft sites... but then it comes out in discussion that they look at them (regularly it appears). Why bother if they are unworthy? And then it comes out that DX projections perform better than GMs and DX guest outside analytic models do better than GMs or DX on average.

Outsiders critique the insiders and find plenty to disagree with & highlight immediately and down the road.

Both sides have cases... and reason to consider the other's.

tarrazu
Posts: 72
Joined: Mon Aug 04, 2014 5:02 pm

Re: Kings/Beech

Post by tarrazu » Sat May 06, 2017 4:59 pm

Short Q/A with King's Luke Bornn
http://www.nba.com/kings/blog/kings-qa- ... luke-bornn#
“I don’t think there’s any one magic number for player evaluation or strategy or scouting -- it really is context-specific.

So, what you look at for an off-the-bench shooting guard is very different than what you might look at for a starting center. It’s about knowing the right suite of metrics that help you understand a player’s full game. There are metrics which try to capture a players’ overall performance in a single number, but if you really want to understand how a player is going to fit within your team, you must go beyond these one-size-fits-all metrics and dig in to the characteristics of each player’s playing style and how they’re going to contribute on the court.”

---

"I’m currently working towards setting up an analytics team and infrastructure so that there are processes and systematic decision-making tools in place that will be used for years and years down the road."

rlee
Posts: 2137
Joined: Tue Sep 10, 2013 3:58 pm

Re: Kings/Beech

Post by rlee » Sat May 06, 2017 5:13 pm

Thanks for posting that quote, Tarrazu. It is astonishing how many folks still don't get the import of what he's saying and still mess around in a fruitless search for a "holy grail."

Crow
Posts: 6250
Joined: Thu Apr 14, 2011 11:10 pm

Re: Kings/Beech

Post by Crow » Sat May 06, 2017 7:58 pm

So come up with role specific metric roll-ups, for general roles or for your really specific one. Unless you roll-up objectively you do it subjectively and then likely do it less consistently and maybe less accurately.

Look at every component of a metric. Which ones do you object to? Change them. Which ones are you uncertain of? Well, then you are going to be uncertain of them with or without the objectively rolled up metric result.

Which is a better STARTING POINT, discrete stats, metrics results and subjective thoughts and roll-ups or just 2 of the 3? If your brain is so smart that it doesn't need metrics to roll lots of inputs on different scales up accurately & consistently ir use one or more of them as a guide or check, then go ahead with that mental magic over explicit math.


Bornn: “Ultimately my role, and that of analytics more broadly, is to help the team make better, more objective decisions."
Want to make better, more objective decisions? Then probably use objective processes as much as possible... including perhaps objective metrics in addition to subjective assessments.

rlee
Posts: 2137
Joined: Tue Sep 10, 2013 3:58 pm

Re: Kings/Beech

Post by rlee » Sat May 06, 2017 9:37 pm

Crow, I read his statement totally differently than you did. My read is that he is not obsessed with "ranking" using a single metric; but rather favors using metrics (of the sorts many of the people here and elsewhere develop) to inform strategic, tactical and personnel decisions. I happen to agree with him that the search for a holy grail single metric does not serve those purposes very well. See Chapter 15 of Basketball On Paper. Dean has said it a lot better than I could ever hope to.

Crow
Posts: 6250
Joined: Thu Apr 14, 2011 11:10 pm

Re: Kings/Beech

Post by Crow » Sat May 06, 2017 10:03 pm

He might use customized metrics, he did say "right suite of metrics". But he said that in heavy, continual contrast to a one metric. He might have meant discrete stats or what I consider discrete stats, which is everything that isn't a roll-up. eFG% isn't a metric to me, it is a discrete stat. But if you consider lots of stuff generically a metric that isn't all in one or a lot in one then you could say that and mean something different than I do.

It is just very common for NBA analytics guys to publically disclaim one metrics to prove they aren't robotically under their programming and to hopefully get others to accept them as a more well-rounded basketball guy or one of the guys, not just a numbers nerd. It bugs me somewhat. He is a numbers guy comparatively. He was hired primarily because of his analytically training & production. He probably should be a metric aware. Pick n chose and know their power sure; but broad-brush disclaim of them is kinda simplistic, imo.

Crow
Posts: 6250
Joined: Thu Apr 14, 2011 11:10 pm

Re: Kings/Beech

Post by Crow » Sun May 07, 2017 4:10 am

Same 4 lineup analysis for different 5th players is one way to do what he seeks without relying on one metrics. I am in agreement in doing that with real data for different players - if coaches test adequately. But they usually don't for most permutations. So in addition to this good but coach challenged approach, one could look at how similar player type 4s do with one "player type" added to their types vs. others. Like David Sparks did years ago but using different, additional and more modern inputs, perhaps including ORPM and DRPM or RPM factors. (Can tinker with same 2s and 3s as well.)

Mike G
Posts: 4429
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: Kings/Beech

Post by Mike G » Sun May 07, 2017 2:33 pm

RPM sez a starting center is good for about +2 points per game relative to league averages. A backup SG is something like -2.
If you need a backup guard but can get a starting center for the same price, you should go for the big guy; then trade him for 2 or 3 backup players.
As long as there are GM's who don't believe summary metrics mean anything, the smarter people will win.

rlee
Posts: 2137
Joined: Tue Sep 10, 2013 3:58 pm

Re: Kings/Beech

Post by rlee » Sun May 07, 2017 3:43 pm

I suggest re-reading Basketball On Paper, Chapter 15: "The Holy Grail Of Player Ratings". Dean says it much better than I can.

JoshEngleman
Posts: 127
Joined: Thu Oct 23, 2014 12:13 pm
Location: Wilmington, NC
Contact:

Re: Kings/Beech

Post by JoshEngleman » Mon May 22, 2017 4:28 pm

Honestly, this could be a lot worse. I'm making a lot of assumptions in the graphic below, so we can tweak if need be. Obviously, Rudy Gay is gone. Afflalo only has a $1.5M guarantee, so he should be waived. Anthony Tolliver has a $2M guarantee, but is only slotted to make $8M next season. They need bodies. He could be a candidate for trade at the deadline next year. From there, there have a few free agents. I expect Evans to be gone. It wouldn't shock if they re-signed Collison, but I can't imagine him wanting to come back. Lawson should be able to get some money elsewhere. The final two guys are interesting. I assume they'll cut bait on McLemore, but really, it depends on the offers that come in. Bogdanovic is interesting. They'll have to sign him using cap space. I have a rough estimate of $8M per year as a starting salary, but that's just a guess.

From there, they'll have the three picks. Using all of my assumptions, they'll have ~$38M in space. They honestly have a ton of flexibility and options moving forward. It's all about how they use their cap space. They need to just play the young guys and see how it all shakes out. Move some of the vets at the deadline next year and just continue to rebuild.

Ignoring the mystique of the Kings right now, I think I would rather be them than the Blazers.

Image

jgoldstein34
Posts: 240
Joined: Fri Sep 30, 2016 6:38 pm

Re: Kings/Beech

Post by jgoldstein34 » Mon May 22, 2017 4:39 pm

You're absolutely right in what they SHOULD do, but it's the Kings... so they'll probably do the exact opposite.

JoshEngleman
Posts: 127
Joined: Thu Oct 23, 2014 12:13 pm
Location: Wilmington, NC
Contact:

Re: Kings/Beech

Post by JoshEngleman » Mon May 22, 2017 5:00 pm

Not that I would expect him to end up here, but this would be an excellent spot for Otto Porter. I would also think about OG Anunoby. The Kings are in a perfect position for him if he has to miss the season. It's OK to be bad in '17-18.

Post Reply