Basketball Analytics Discussion Thread
Goto page Previous  1, 2, 3

 
Post new topic    LakersGround.net Forum Index -> General Basketball Discussion Reply to topic
View previous topic :: View next topic  
Author Message
Aeneas Hunter
Franchise Player
Franchise Player


Joined: 12 Jul 2005
Posts: 23994

PostPosted: Wed Oct 24, 2018 3:48 pm    Post subject:

I promised that I would look at it with an open mind once Cranjis finished his work. Is there some place where this is collected? I couldn’t find it.
Back to top
View user's profile Send private message Reply with quote
Aeneas Hunter
Franchise Player
Franchise Player


Joined: 12 Jul 2005
Posts: 23994

PostPosted: Thu Sep 05, 2019 10:29 am    Post subject:

For those of you with access to The Athletic, this is a good article. He hits the nail on the head -- the typical criticism of analytics is that they aren't perfect. Analytics tell us that shooting threes is better than shooting long twos. Someone like Morey builds a team that focuses on shooting threes and layups. The team does not win a championship. The naysayers proclaim that analytics has failed. In other words, analytics were not perfect, therefore they most be junk.

Quote:
Yet despite this well-worn path, there is a temptation among traditionalist skeptics but also from within the analytics community itself to expect perfection. There is something about a number that implies precision. Accuracy. Certainty. Immutability. The implied message is a problem solved in totality. While exacting standards are generally a good thing, the notion of perfection can be a mortal enemy of improvement in light of how progress actually happens.

From the standpoint of the traditionalist, the quest for perfection can be a cudgel. The intricacies of whichever game is being assailed by quantitative analysis (and worse, analysts) escape categorization. The flow of play creates situations so unique as to make systematic measurement either incomplete or incomprehensible. While there is a point to be made, cautioning against dogmatic applications of misunderstood findings — simply dribbling down the floor and shooting a pull-up 3 because 3s are good likely wouldn’t produce a great offense — the appeal to complexity is rarely that note of judiciousness. Rather, the inability to account for everything becomes a lack of trustworthiness in accounting for anything and thus argues for a rejection of data-driven understanding entirely.


https://theathletic.com/1184593/2019/09/05/the-art-of-being-less-wrong-how-analytics-actually-works/
Back to top
View user's profile Send private message Reply with quote
Mike@LG
Moderator
Moderator


Joined: 10 Apr 2001
Posts: 62783
Location: Orange County, CA

PostPosted: Wed Nov 20, 2019 8:06 am    Post subject:

Aeneas Hunter wrote:
For those of you with access to The Athletic, this is a good article. He hits the nail on the head -- the typical criticism of analytics is that they aren't perfect. Analytics tell us that shooting threes is better than shooting long twos. Someone like Morey builds a team that focuses on shooting threes and layups. The team does not win a championship. The naysayers proclaim that analytics has failed. In other words, analytics were not perfect, therefore they most be junk.

Quote:
Yet despite this well-worn path, there is a temptation among traditionalist skeptics but also from within the analytics community itself to expect perfection. There is something about a number that implies precision. Accuracy. Certainty. Immutability. The implied message is a problem solved in totality. While exacting standards are generally a good thing, the notion of perfection can be a mortal enemy of improvement in light of how progress actually happens.

From the standpoint of the traditionalist, the quest for perfection can be a cudgel. The intricacies of whichever game is being assailed by quantitative analysis (and worse, analysts) escape categorization. The flow of play creates situations so unique as to make systematic measurement either incomplete or incomprehensible. While there is a point to be made, cautioning against dogmatic applications of misunderstood findings — simply dribbling down the floor and shooting a pull-up 3 because 3s are good likely wouldn’t produce a great offense — the appeal to complexity is rarely that note of judiciousness. Rather, the inability to account for everything becomes a lack of trustworthiness in accounting for anything and thus argues for a rejection of data-driven understanding entirely.


https://theathletic.com/1184593/2019/09/05/the-art-of-being-less-wrong-how-analytics-actually-works/


There's a huge gray area for Houston. They avoid high percentage, open midrange looks, the shots the team is easily capable of hitting, for semi-contested 3s for the sake of PPP.

Sometimes you just need a bucket, and I think they leave points on the table for avoiding those shots.
_________________
Resident Car Nut.

Pipedream:
CP3
Caruso
LeBron
Jerami
Anthony Davis

http://bball-index.com/team/lal/
Back to top
View user's profile Send private message Reply with quote
Aeneas Hunter
Franchise Player
Franchise Player


Joined: 12 Jul 2005
Posts: 23994

PostPosted: Wed Nov 20, 2019 10:06 am    Post subject:

Mike@LG wrote:
There's a huge gray area for Houston. They avoid high percentage, open midrange looks, the shots the team is easily capable of hitting, for semi-contested 3s for the sake of PPP.

Sometimes you just need a bucket, and I think they leave points on the table for avoiding those shots.


That argument sort of assumes its conclusion. Are open midrange looks "high percentage" shots that the team "is easily capable of hitting"? I haven't done the research, but the stats are readily available these days. What's the leaguewide average for, say, an open 12 foot jump shot? I don't have the time to run it down, but I'd guesstimate that the number is probably in the 50% range. This season, the league average is .411 from 10-16 feet, but if we could break out open looks, the percentage would go up. From the perspective of analytics, that is not an easy, high percentage shot.

There is a separate tactical issue about Moreyball. Until someone wins a title with that sort of style, people are going to say that you can't win that way. The same people would have said this about the Warriors' style, which was close to Moreyball before they got Durant. We shall see how Moreyball pans out.
Back to top
View user's profile Send private message Reply with quote
Aeneas Hunter
Franchise Player
Franchise Player


Joined: 12 Jul 2005
Posts: 23994

PostPosted: Wed Dec 18, 2019 10:08 pm    Post subject:

Here’s another interesting point from Seth Partnow in The Athletic:

Quote:
If one were making a short list of the things that statistical analysis of basketball are better at picking up than traditional scouting, I’d put fouling propensity near the top. We intuitively know that committing fouls and sending players to the line is bad: the average field goal attempt is around 1.05 points per shot in most seasons, while committing a foul on a shot turns the same hypothetically average shooter from just over 50 percent to almost 80 percent accuracy.

But even knowing how damaging sending the opposition to the line is, it’s hard to differentiate between players who are good at avoiding fouls and those who aren’t. Montrezl Harrell sent opponents to the line in “non-intentional” foul situations the most, committing 161 combined shooting fouls and personal or loose ball fouls while in the penalty, according to NBA play-by-play data. Jeremy Lamb committed 81, roughly half as many, and was 147th in the league. With that narrow a spread of outcomes, judging by observation alone is next to impossible. Plus, to accurately credit a player for not being foul prone, one is searching of evidence of absence, always a tricky task.

All of this is a prelude to say that the reason some defenders who are well thought of might not perform as well as their reputations on plus/minus based models is those models can pick up on the damage of foul-proneness much more easily. For example, Patrick Beverley is clearly a dogged defender, and routinely takes on the challenge of guarding scorers much bigger than he is. We see the activity and intensity and naturally want to credit him for effectiveness. But, the side effect of that physicality and ferocity is that he can be … handsy. Last year Beverley committed 49 non-intentional fouls while opponents were shooting the bonus, tied for the fourth-highest single-season total since 2005-06. In terms of converting point differential to expected wins, that’s almost a full game in the standings’ worth from all the extra half points he’s given away.

I only use Beverley as an example, as he still grades out as an above-average defender in systems such as Ryan Davis’ version of RAPM, but he has not rated as a top 100 defender in that system since 2016-17. Food for thought next time you’re watching a defender on your favorite team really “get into” someone: they might be doing as much harm as good.
Back to top
View user's profile Send private message Reply with quote
L4L
Starting Rotation
Starting Rotation


Joined: 28 Nov 2007
Posts: 291

PostPosted: Fri Jan 24, 2020 1:51 pm    Post subject:

Does anyone know where I can pull conditional on/off stats? For example, 4th quarter when James is on the floor but Davis is specifically off?
Back to top
View user's profile Send private message Reply with quote
Mike@LG
Moderator
Moderator


Joined: 10 Apr 2001
Posts: 62783
Location: Orange County, CA

PostPosted: Fri Jan 24, 2020 8:06 pm    Post subject:

Aeneas Hunter wrote:
Mike@LG wrote:
There's a huge gray area for Houston. They avoid high percentage, open midrange looks, the shots the team is easily capable of hitting, for semi-contested 3s for the sake of PPP.

Sometimes you just need a bucket, and I think they leave points on the table for avoiding those shots.


That argument sort of assumes its conclusion. Are open midrange looks "high percentage" shots that the team "is easily capable of hitting"? I haven't done the research, but the stats are readily available these days. What's the leaguewide average for, say, an open 12 foot jump shot? I don't have the time to run it down, but I'd guesstimate that the number is probably in the 50% range. This season, the league average is .411 from 10-16 feet, but if we could break out open looks, the percentage would go up. From the perspective of analytics, that is not an easy, high percentage shot.

There is a separate tactical issue about Moreyball. Until someone wins a title with that sort of style, people are going to say that you can't win that way. The same people would have said this about the Warriors' style, which was close to Moreyball before they got Durant. We shall see how Moreyball pans out.


That's a league average shot. But when it's specifically Houston, and they're best shooters are open to those shots, those shots should be 45%+, not 41%.

That's what I mean.
_________________
Resident Car Nut.

Pipedream:
CP3
Caruso
LeBron
Jerami
Anthony Davis

http://bball-index.com/team/lal/
Back to top
View user's profile Send private message Reply with quote
Aeneas Hunter
Franchise Player
Franchise Player


Joined: 12 Jul 2005
Posts: 23994

PostPosted: Sat Jan 25, 2020 11:49 am    Post subject:

L4L wrote:
Does anyone know where I can pull conditional on/off stats? For example, 4th quarter when James is on the floor but Davis is specifically off?


You can do something like this.
Back to top
View user's profile Send private message Reply with quote
L4L
Starting Rotation
Starting Rotation


Joined: 28 Nov 2007
Posts: 291

PostPosted: Mon Jan 27, 2020 5:30 pm    Post subject:

Aeneas Hunter wrote:
L4L wrote:
Does anyone know where I can pull conditional on/off stats? For example, 4th quarter when James is on the floor but Davis is specifically off?


You can do something like this.


For anyone’s future reference, this is extremely useful and has more options than the NBA.com tool:

https://www.pbpstats.com/wowy/nba
Back to top
View user's profile Send private message Reply with quote
Aeneas Hunter
Franchise Player
Franchise Player


Joined: 12 Jul 2005
Posts: 23994

PostPosted: Thu Mar 05, 2020 4:34 pm    Post subject:

https://theathletic.com/1656087/2020/03/05/searching-for-the-heart-of-sloan-on-the-state-of-the-analytics-union-in-2020/

This is a fascinating discussion between Hollinger and Partnow about what is really going on with analytics at the team level. Here are a few interesting or amusing tidbits:

Quote:
I joke, but on the team side the research process often degenerates into what I call “drunk-lamppost analytics,” where the analytics research “question” is really a hunt for data to prove somebody’s preconceived point.

Bigger picture, I think understanding uncertainty is still a big fail point. Take the draft, for instance, which is probably the place people are most likely to get it but still often don’t. People intuitively understand that a future draft pick will have some kind of variance to it in terms of where it falls, and another kind of variance to it in terms of the quality of player it will yield. Yet for some people, that lack of certainty becomes a massive impediment that they can’t get around. I’ve actually had team officials tell me they want “pick certainty” in a trade. Fine, call me back in June then. Or better yet, try me 10 years from now and we can go back and figure out what the pick was worth.


Quote:
Before turning to something more positive, I want to agree with you regarding understanding and accepting uncertainty into our lives. In a way, I think well-trained analytics and analytics-adjacent professionals are ill-suited to be effective in an environment that can reward the appearance of assuredness bordering on bombast. One of the outside criticisms often leveled at the purveyors of statistical findings that I find frankly hilarious is the notion that we’re so sure of ourselves despite [INSERT FAILING X USUALLY NEVER PLAYED THE GAME].

People, trust me when I say that most often nobody knows more about how wrong models are and why they miss than the people who build them in the first place. Which is sort of the heating vent in the Death Star in group discussion settings, where the more “traditional” might not have explored their own biases in the same way so as to understand why they are also frequently wrong.


Quote:
I think we’ve made a ton of progress on health and wellness analytics in particular, and I think the cross-sport cooperating in this area has been a huge catalyst — nobody in English soccer gives a crap if American basketball teams copy what they’re doing, so the sharing has been relatively free and honest.


Quote:
But lastly and most interestingly, I think there are going to be advances in better understanding what is currently lumped under the “intangibles” nameplate. I’m not an expert enough in the field to be certain, but I think there are a lot of findings coming out of the industrial psychology discipline that could shed some light on personality fit, culture building and things of that nature. This could affect not just roster decisions, but also hiring decisions up and down basketball operations as getting a better handle on who will work well together brings some fairly obvious benefits.

The thing that I think is less likely to manifest is a similar “eureka” moment to seemingly sudden recognition that 3>2. Much like baseball made the fundamental discovery that not making outs was the most important thing, that was basketball’s apple-on-the-ground moment and we’re unlikely to find another One Big Thing that fundamentally changes the way the game is played in the near future.
Back to top
View user's profile Send private message Reply with quote
Display posts from previous:   
Post new topic    LakersGround.net Forum Index -> General Basketball Discussion All times are GMT - 8 Hours
Goto page Previous  1, 2, 3
Page 3 of 3
Jump to:  

 
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
You cannot vote in polls in this forum






Graphics by uberzev
© 1995-2018 LakersGround.net. All Rights Reserved. Privacy Policy. Terms of Use.
LakersGround is an unofficial news source serving the fan community since 1995.
We are in no way associated with the Los Angeles Lakers or the National Basketball Association.


Powered by phpBB