Rob Pelinka Discussion: BBall Ops Pres + GM (AKA: Lucky Rob)

I’m seriosly stoked simply having a guy in our front office in a position of need smart enough literally to be a rocket scientist for NASA over a decade … mind blown!







THAT is how Mark Walter rolls.

 

I like that signing

Only thing I'm iffy with the whole thing is the timing of it.
Couldn't or shouldn't we have this extra hands a few months ago? Seems to me that these FO additions come to close to the summer and there's not a lot of time to have an effect. But hey, better late than never

 

meh - big titles mean little. my dad's a "rocket scientist" and he shouldn't be anywhere near an NBA management team despite being a lifelong NBA fan.

I'm more curious what his rep was in NO - I actually think they did a good job in talent evaluation overall during the griffin era but anyone have any info on his particular influence

 

Buzzkill.

 

Lakers Search For A Center, Offseason Q & A

 

He seems to be data oriented so looks like he can lead projects on mathematical models for cap and data analysis to support decision making.

Soccer teams in Europe have teams with dozens of people just for data and performance analysis, it's about time we develop in that regard

 

i say this as the guy on this site probably most associated with analytics interest (which to me is a little funny, as i've played the opposite role in some circles): i'm skeptical about folks claiming analytic advantages. to me, the only analytic advantages require data advantages. ok, that's not a firm rule in all fields or anything, but to me, if everyone's working with the same data, pretty sharp folks will converge on similar methods (i.e., i'm pretty sure in-house metrics aren't as in-house as they seem to those in the house).

i'm actually fascinated into what might be discussed in an interview with someone for this role. could they demonstrate that they were able to locate a talent analytically that others missed on? their former employer obviously has that knowledge now, so...how do you prove you've got more? who on staff is qualified to determine that you're not blowing smoke?

again, i'm just happy we're bringing in outside voices, regardless. rocket scientist or not, it's not just tim harris and the rambii.

 

I have zero inside knowledge, but I would guess that the Lakers have been operating at an analytics DISadvantage. Along with disadvantages in almost every FO aspect. If the Lakers can just get up to league average, that could actually be a huge boost.

 

totally possible. i'm operating on the assumption that everyone has found folks who are part of an extended network of sports analytics people, but it's possible we haven't.

and i know svtzr will be here soon to explain the parallels to why i was wrong to suggest similar about injury treatment, lol. what can i say? i have a worldview of sorts...probably just too many times seeing people claim they have an insight when it's a) either the same as everyone else's who's thought about this problem or b) something worse.

i do find it strange that rob said simply "data, cap, analytics" or something to this effect. first of all, my assumption was that you might have separate folks with specific expertise in cap/finance/rules/etc. and data/analysis (and maybe another set for non-analytic/old school scouting and strategy).

but i really don't know enough about how modern nba front offices are constructed to speak intelligently here.

i do know that our previous way was not the way.

 

Yeah, this is my position in a nutshell. We were operating with disadvantages in nearly every FO aspect, from # of personnel, to treatment facilities, to who knows what else. We were not a modern NBA front office. Becoming a modern NBA front office should boost us, as we go from below average to at least average or above, in those areas.

 

Yeah I also don't feel informed enough on this area, but applying common sense and critical thinking based on what I know about these topics: I think more than "I found player X or Y", it is about having a process that they're able to put in place.

So for example, this summer we have probably like 1 million scenarios possible depending on who we sign, when and at what number. Just one guy asking for 1 million more or less can change completely any other moves. So we really need a powerful model behind it where you can in real time input an occurrence and the model will automatically tell you the next options available. You need someone that understands the cap and data analysis to pull that up, for example.
This is where I think 1 month is too short to do it, but maybe there's already some work done or some sort of template available.

Then regarding other FO stuff outside of this summer, we need a process to scour the league looking for advantages. Build our own data sets so we can play with that. Look at catch and shoot% instead of 3pt %, look at guards who defend big, find guys that are better in fighting through a screen, ability to shoot from a pass from the other side like Luka does often, whatever it is. Data is available to us so I can only imagine how much more data these guys have. I see benefits in having your own in-house expertise because that creates an opportunity to look at things others may not.
Like how you decide what 3&D you go for? I hope it's not on vibes and they have a database they can go through based on what makes sense for JJ offense and defense schemes, best fit to Luka, etc. JJ is very data oriented it seems, so I think we are starting to grow there and this signing is a step towards it

I remember seeing in a podcast a player saying his agent had a whole league analysis on like shooting charts and touches and stuff like that before they made a decision. Even the host got impressed by the level of detail. I even think it was AR when he got undrafted. So even agents are super data driven, I can only imagine the depth you can go in a FO.

 

You know it's the same age old adage of: nothing is new under the sun, the only caveat to that is, when it actually is something new under the sun haha.

So I tend to agree with you, I think when the systems are closed and known - like the cap or CBA or analytics derived from box scores and data tracking... everyone sort of has the same inputs and gets similar outcomes.

The body is quite complex and there are alot of unknowns with many variables - so I tend to see a lot of interesting treatments, the various health advancements are a testament to that.

 

I think part of the reason we signed Smart is because Reddick and the coaching staff have a strong analytics team. They deemed that smart's hustle and defensive metrics were still through the roof even though it wasn't being reflected in time/play for losing teams.

 

prime example of where analytics fails. we don't really have good ways to quantify (or maybe even qualify) some contextual factors that really matter. two big areas are the impact of injury/recovery and coaching/strategy. like, some coaches don't use guys well, and some find incredible use for players who would not otherwise be very useful.

this is part of why i'm really not certain how valuable keon ellis is. i think mike brown found good uses for him that doug christie had more trouble finding. does that mean brown's good and christie's bad (i actually believe both of these things)? even if that's true, what does it mean for ellis in his next stop? did atkinson just not have the right role, or did he agree with christie that ellis just wasn't good (as slick suggests)? if we're considering signing ellis, redick's input here would be pivotal, i'd assume. probably moreso than anyone in the front office or advance scouting, etc.

some context here is i'm reading a book right now by some computer scientists about the limits of AI, and they reiterate truths from standard research of the sort that i have experience with: essentially, predicting individual outcomes in complex systems is really difficult, and even insane computing power does essentially nothing for you in a lot of these cases. because the data they're based on are in and of themselves fuzzy measurements at best.

like we can model shot quality in certain ways, but can you quantify the way that players talk about how receiving a pass from doncic or kidd differs from receiving passes from other players. i pick this example because we've found various ways to quantify, say, open shots (e.g., tracking data has no player within 6ft of shooter on catch, etc.), but none of those also calibrate whether the velocity is perfect, the ball is delivered right to the shot pocket, the player is in rhythm with respect to shooting motion, etc. those are things that arise from system, coaching, a specific talent of a teammate, etc.

so, interestingly, in the kind of closed system that svtzr describes, it almost takes someone seeing all the analytics (and understanding them) and then having an intuitive sense as to why one case is going to break a different way. it's more art than science at that point. we got into this re: the orthopedic stuff, but my clinical psych friend essentially says this is the state of psychotherapy. yes, there's a science to the underlying treatments, but individual decisions end up being based on that and something less scientific.

anyway, all that said, it could be as panko said, and we were actually operating previously without even the first part. which would be bad. as just plain ol' horse sense ruling all is kind of the hallmark of a bad gm in the modern nba (e.g., magic, maybe dumars in NO).

 

I think this is the key. Our front office has been thin for so long that just getting to a normal size staff will be big for us.

 

Yes, we managed to hire a person who can manipulate Claude into giving him data ... I am amazed.

 

yeah, gotta admit that at this point when i hear ai models being touted...i get a little antsy. i'm not to the point where if anyone says it i'm mad or anything, but if you work in any organization right now, there's a lot of dumb people shoving things down your throat that show a lack of understanding of both the thing and what it can and can't do.

i'm all for outsourcing repetitive mundane stuff that can be a source of human error, but any time someone brings in AI to assist with "thinking" i kind of hate it, immediately. it's basically a scourge in academia, and i imagine it's similar in other industries.

hopefully here, AI models are trained on tracking data to extract things we want to know or as applied to the cap to maybe model potential FA outcomes that will impact player valuation (i could totally see value in this, btw).

 

Exactly. In the past month, we've posted openings for 3 other junior to senior roles in this one area, a data scientist, a senior data engineer, and a software developer. So that's 4 new hires in data with potentially more to come. We were understaffed before, and still may be.

If we were still under the cheapo Jeanie regime, then yes, even 1 new hire doing anything, would indeed be amazing!

Either there is useful work being done with data or not. If there is, then regardless of fluffy AI language being tossed around or not, there is a benefit to not being understaffed. New ownership seems to think we were understaffed by a minimum of 4 people in this one area.

 

This is exactly my thoughts on the whole AI thing.
I work in IT and it's sad how companies are jumping to stuff with no foundation whatsoever.

A friend of mine mentioned this is the new bubble and basically many companies will go all in into AI, fire a bunch people, it will start failing in multiple directions and costing money to the company where they then return to some sort of middle ground and start hiring people again and it stabilizes there.