APBRmetrics

Couple of basketball panels.

As always, "value" largely dependent on whether anyone actually reveals something not already known or has an edgy opinion.


"....Building a Modern NBA Team" panel?

Will the 76ers panelist reveal anything interesting / new about current build design? Reflections on status / coming challenges? Face any tough questions or comments?

Same with Kings.

What do teams think about paying 30+ and 35+ year olds big and relying on them? What does the overall and specific performance data show?

Try to win title with lead(s) on 1st, 2nd or 3rd max? What has worked best over last 5-10 years?

How much to invest in PGs & Centers vs. rest?

Are trends changing? Rely on 3, 5 or 10 yr trends, if any?

The extended thread on topic:

https://x.com/bballstrategy/status/1895542913880502609

Virtually no media coverage of Sloan today showing up on Google news today or before.

Will look tomorrow thru Monday.

Some tidbits on Twitter but didn't see anything big so far.

If you were at The Analytics-Driven Franchise: Building a Modern NBA Team panel, what did you think of it?

Still almost nothing in mainstream media.

Twitter is full of notifications of participation but almost no analysis of presentations. I found one conference writeup but almost nothing mentioned particularly moved me.

Other than the importance of communication from analysts to others.

It was a chance for me to say this again or for the first time (I thought it before but can't be sure at moment whether I said it):

The analytic shop should hire the best lawyer or trial lawyer it has money to buy to do its communication.

Recognize skills you and team have and the skills others might have. Get a hired gun to "communicate" for you. If allowed.


I did find one reference to a player adaptability research paper that was useful to me.

News articles today? I didn't see any.

That research paper on player adaptability? I responded with 20+ comments on how to use / extend the research presented:

https://x.com/bballstrategy/status/1898885036591734958

What was the most newsworthy comment you heard, directly or indirectly?

https://www.youtube.com/@42analytics/videos

I don't if they are all there. Probably. At least most.

The Chemistry presentation about improvement over time of shooters with their teammate passers was good. It was in general but it helps make the case for lineup concentration or at least some main sublineup concentrations.

The Nerds Impact presentation was heated on the 3pt game.

Building a Modern Team was ok but pretty standard. No one addressed my top of thread questions about leading with highly paid vets over 30 or 35.

More on Nerd / analytics Impact panel:

Anything besides 3 point increase a negative or potential negative to discuss?

Panel did not identify anything else but I could say:

Using analytics to jack seat prices to the max is a negative to many.

Using analytics to reduce offensive rebounding effort to better defend fast breaks might be a negative to some on both sides of the court.

More 3s resulting in more offensive rebounds won offsets to some degree.


Suggestion that the NBA will go to 1 free throw in near term. Undiscussed is the impact that shorter ft breaks would have on player wind recovery and ability to stay in game for big stints and long minute totals. And maybe susceptibility to injury. At some marginal level but where more is bad, potentially very bad. If there is only 1 fta, will more players do the Giannis and stretch it out anyway to the max allowed or beyond? Will coaches and entering plays spam out substitutions to negative much of desired time saving? How "necessary" or "useful" fta downtime to players, "the game" and storytelling / video replays / plugs of various kinds, fan breathing, eating, thinking / talking, multi-tasking to scoreboard or betting or family, etc.?

They mentioned that fans may not have always considered secondary and higher level issues associated with a change. Have they on this? Will a 1 fta system make average or below average FT shooters less desirable to draft or play? Who do you lose? Does it make hack a Shaq more common? Will the greater variance in points produced change end game behavior?


Daryl Morey mentioned offensive rebounding twice as something becoming more important. Few on outside pay attention to that. I do. And he said the trend is towards players who can hit the 3, defend thru multiple switches and offensive rebound. 3, D and OR. I sometimes shortened it to 3 & OR as being the basic new type to look for. And saying defend thru multiple switches is important. It might not be just one. Sliding up and down and way up and down is increasingly important in a game and on the same play repeatedly. Likely to see more 6-8s? Probably.

Undiscussed was whether analytics was having overall positive or negative on who gets paid big or the max. Is it better or worse than before? Should the decisions be better or far better than they actually are? Is that the fault of GMs, owners, coaches or the analysts not making a more forceful, effective case? On average which player tier is getting too much or not enough (besides the very top often compensated for difference by outside endoresements)? What can be done better?


Who in "analytics" has taken a major role in CBA revision discussions? I am not aware of any visible. Shouldn't they?

Is anybody in the league using 29pt "pose" video data to try to identity injury hot spots / precise positional behavours and suggest rule changes? They should. Maybe some are.

Disappointed that no one else, to extent I could discern, actively supported Morey's call for 2 ftas on a fouled 3pt attempts, except at end of game. I didn't have that exception in mind previously and not sure I would support now. The G league does it but few fans follow or care about the G league. 3ftas is simply too much by the math for that foul and the variance of 3pts or none is major, even it balances out in long-run.

Several of the panelists in the "Nerd" group said or agreed with statement that "we still have the same amounts of shots at the rim' as 3ptas have grown.

At BRef:

League % of shots within 3 feet

-'2015 season 29%
2020 28.2%
2023 24.5%
2025 22.6%

Does some other dataset show differently?

Same b-r.com does have some conflicting or unexplained numbers. The 0-3' shots don't concur with dunks and layups.
Percentage of FGA by type: Dunks are presumably not a subset of layups. So if we just add them -- final column -- they seem to have grown much larger than the % within 3 ft. And larger than in earlier years.

Looking a bit closer, it seems the accompanying Made columns are not consistent. The %FGA which are Dunks indeed seem to be the % of all FGA; but the % Layups are apparently the % of 2-pt FGA.

A "corrected" version of the table might look like this; now the layup column is the previous one multiplied by the %2's: Dunks + Layups are now less than 0-3' in every season. The difference varies by a lot, though -- final column -- starting and finishing with very few 'at the rim' shots which are neither dunked nor laid up.
Probably just inconsistent observation from team to team, and over the years. Is a Zubac mini-hook within 3'? Is it a layup?

Thanks for the continued exploration.

Messy.

Not sure what to think in the end.

I mentioned this thread to BRef. Will see if / what they respond.

Mentioned to Dean Oliver as well.



Teams are probably using a different database? The "right" answer matters for the impact of 3pt revolution debate.

From [published] BRef:

This season, league 0-3 feet now 22.7%. Celtics 18%. 3pta, league 42.1%, Celtics 53.7%. Celtics more than 25% below league average on 0-3 foot shot frequency.

2020, league 0-3 feet now 28.2%. Celtics 27.6%. 3pta, league 38.4%, Celtics 38.6%. Celtics barely different from average then.

That is not staying the same on this dataset.

So they looked into your basic question and stuck with 0-3 as the definition of 'at the rim'?
And we can wonder about %layups and how this may have evolved over the years. And the source(s) for these designations.

The Shooting page would be a lot more useful if total FGA were listed. Whenever I have tried to analyze something, I have to dig up FGA somewhere else.