On Assets and Liabilities: Celtics Playoff Bench vs their Peers

Jimbodandy

Member
SoSH Member
Jan 31, 2006
11,405
around the way
That's right. This is going to be on the starters to win or lose, and the bench's objective will be to get in there and compete for the few minutes the starters need to rest. However, they are missing one key starter and this is a series where they cannot really expect much help from one key reserve (Kanter). I think Toronto should be the favorites here and the difference will be whether Tatum rises to a new level. The good news, I guess, is if they can get by Toronto, maybe they get Hayward back at some point in the next series.
I should probably just ask someone to change my tagline to "Defense wins championships", but I like what we can take away from Toronto. It's not about Tatum going full Dame/Harden in this series IMO. It's about executing well on defense and minimizing what Toronto does well (run) and crushing what they already don't do as well (halfcourt). We're gonna score on them regardless.
 

DeJesus Built My Hotrod

Well-Known Member
Lifetime Member
SoSH Member
Dec 24, 2002
48,205
I should have been more precise- PER tends to value volume regardless of efficiency. I'm not sure what the "break even" point is now, where more shooting will increase PER, but a couple years ago it was something like 30% on twos and 20% on threes. So a high-volume, low efficiency scorer, like say Brandon Jennings, has a career PER of 15.7 with a poor career TS% of 49.7%, whereas Marcus Smart stands with a PER of 12.1, with a similar 50.1 TS%. The difference is mainly attributable to Jennings having a much higher FGA rate. A good metric shouldn't reward players for inefficient volume. Brandon Jennings, an "above average player" according to PER, who's still in his prime, was last seen playing in Russia.

PER also doesn't do much to capture defense (it only accounts for blocks and steals), doesn't account for who a player is playing with or against, and, of course, does not capture any non-box score impact- screening, boxing out, good defense that doesn't lead to blocks/steals, ball movement that doesn't lead to an assist, etc... A lot of the stuff that's important to contributing to winning basketball games. It doesn't tell us anything about whether a team is better when a guy is on the court or off. Even a guy like Draymond, who's fantastic at generating blocks and steals, is an on/off beast, and even generally rates well in offensive RPM, has a 15.1 (dead average) career PER, lower than Brandon fickin' Jennings, despite being better at everything on a basketball court except chucking up shots.

There's a good reason nobody in the basketball analytics world uses PER. It's a very blunt instrument with little predictive value, and I just don't think it tells us much about the respective bench depth of play-off teams.

For a comprehensive breakdown and discussion of the pros and cons of PER, this is a Reddit post from the creator of ESPN's Real Plus Minus, Jeremias Engelmann: https://www.reddit.com/r/nba/comments/cmrkv3/basketball_stat_player_efficiency_rating_per/

I think Engelmann's Multi-year RAPM (obviously not particularly useful for rookies) or Jacob Goldstein's PIPM. https://fansided.com/2018/01/11/nylon-calculus-introducing-player-impact-plus-minus/ would be much more useful tools for this kind of analysis.
For example, using PIPM (again, I am sorry, LondonSox), the Celtics have four players who are as bad or significantly worse than the Raptors second lowest ranked player, Matt Thomas -Langford, Wanamaker, Semi and Edwards. The last two are particularly odious using that metric but as you point out, Stevens appears to be effective at optimizing Ojeleye's usage in ways that likely aren't being accurately captured in the available data.
 

radsoxfan

Member
SoSH Member
Aug 9, 2009
13,622
That's right. This is going to be on the starters to win or lose.
I think this is correct though the Hayward injury gives the Celtics a very thin margin for error. One more injury to a top guy or a few games with significant foul trouble and we go from being able to minimize our bad bench with matchups to being in big trouble. Have to hope for some good fortune.

Hayward was obviously signed to be a star/near-star level player, but with the emergence of Tatum/Brown has basically turned into a super-elite role player. His versatility and overall high quality of play is very important for the overall depth of the team given the current construction.

In a weird way, his 35 minutes a game is like having a couple of extra elite role players on the team (plus, someone who can slide into a bigger role with injury/foul trouble).
 

DannyDarwinism

Member
SoSH Member
Jul 7, 2007
4,883
Oh, don’t get me wrong, I think the Celtics’ bench is pretty weak, especially compared to the Raptors. I just wanted to bash PER, just in case the 213 previous PER-bashing posts in the Port Cellar haven’t gotten the point across over the years.
 

radsoxfan

Member
SoSH Member
Aug 9, 2009
13,622
PER is always fun to bash, and there are many many instances of it not capturing a player's value well. Certainly a blunt tool that needs to be used with skepticism.

But if you knew nothing about the players involved, and I told you one team is playing 7th-9th rotation guys with a PER in the 12-16 range and the other team was playing guys with a PER in the 6-8 range, that's almost always going to be telling you something about a real quality difference.
 

DannyDarwinism

Member
SoSH Member
Jul 7, 2007
4,883
PER is always fun to bash, and there are many many instances of it not capturing a player's value well. Certainly a blunt tool that needs to be used with skepticism.

But if you knew nothing about the players involved, and I told you one team is playing 7th-9th rotation guys with a PER in the 12-16 range and the other team was playing guys with a PER in the 6-8 range, that's almost always going to be telling you something about a real quality difference.
Yeah, I agree with that, but I’d certainly want to look at luck-adjusted RAPM, PIPM, BPM, RPM or Bowiac’s own DARKO instead. I think any of those will tell you more.
 

wade boggs chicken dinner

Member
SoSH Member
Mar 26, 2005
30,494
I think this is correct though the Hayward injury gives the Celtics a very thin margin for error. One more injury to a top guy or a few games with significant foul trouble and we go from being able to minimize our bad bench with matchups to being in big trouble. Have to hope for some good fortune.

Hayward was obviously signed to be a star/near-star level player, but with the emergence of Tatum/Brown has basically turned into a super-elite role player. His versatility and overall high quality of play is very important for the overall depth of the team given the current construction.

In a weird way, his 35 minutes a game is like having a couple of extra elite role players on the team (plus, someone who can slide into a bigger role with injury/foul trouble).
This is one way to look at it. The other is that if each remaining playoff team had to take its second-highest paid player off the roster and then one other starter, that team would be in pretty bad shape (one other exception is DAL, so long as Doncic remains in the lineup since they've been playing without their highest paid player and have done a great job fitting pieces around Luka).
 

DeJesus Built My Hotrod

Well-Known Member
Lifetime Member
SoSH Member
Dec 24, 2002
48,205
Oh, don’t get me wrong, I think the Celtics’ bench is pretty weak, especially compared to the Raptors. I just wanted to bash PER, just in case the 213 previous PER-bashing posts in the Port Cellar haven’t gotten the point across over the years.
Not at all. The data was already in the Toronto thread so I thought we could show (albeit without the actual numbers) what you were referring to here. In short I agree with you.
 

the moops

Member
SoSH Member
Jan 19, 2016
4,700
Saint Paul, MN
Milwaukee: 6 players with >10 starts, exclude R. Lopez & B. Lopez, then George Hill (backup PG) as #6.

7: SG/SF Pat Connaughton, 18.6 (0.8, 11.5)
8: SF Kyle Korver, 16.6 (0.5, 11.3)
9: PF Ersan Ilyasova, 15.7 (0.6, 14.9)

A very veteran squad for Deep Bench.
Aren't we missing Marvin Williams in this bench analysis?
 

Eddie Jurak

canderson-lite
Lifetime Member
SoSH Member
Dec 12, 2002
44,475
Melrose, MA
I think game 1 gave us a look at the some of the upside of the Celtics bench.

Semi Ojeleye is not what anyone would call a complete, well-rounded NBA player. He's not even a consistent 3&D guy. On defense, he's not a guy who steals anything ot blocks anything. (Kemba Walker had twice as many blocks this year as Semi has had in his whole career). But satrictly in terms of man to man defense he can be a good matchup on certain players (notably Siakam today) and that can make a difference in a way that a generic bench guy cannot.

And Robert Williams gave them some important quality minutes with Theis in early foul trouble.

I think the biggest bench weakness in Wanamaker, who was fine yesterday but may end up being overexposed with all of the injuries.
 

Devizier

Member
SoSH Member
Jul 3, 2000
19,469
Somewhere
Of the first generation all-in-one metrics (efficiency, PER, and oRTG/dRTG), I would probably take PER as the best shorthand option. But that’s like saying I would take batting average over runs and RBI.
 

InstaFace

The Ultimate One
SoSH Member
Sep 27, 2016
21,759
Pittsburgh, PA
yeah ok guys I'm a basketball stats newbie. Message received on PER, it was just what was handy. I can re-do the numbers vs remaining contenders - sounds like PIPM is built off of RAPM and improves on BPM (indeed, starts from it), so while I wish I had more of a player-tracking-based metric, we can certainly use that (and its defensive component, D-PIPM) to look at these groupings a bit better.

And I've got some questions posed here that I owe answers to, will get to that. e.g.

Aren't we missing Marvin Williams in this bench analysis?
These are all subjective attempts to look at who would get promoted to a playoff rotation spot if given an injury, on the assumption that most playoff teams only rotate 6 non-bigs. So just to take Milwaukee as a specific example of how I looked at things:

1. First step is, who's a starter? Milwaukee has only 6 players with >10 starts, Giannis Middleton Bledsoe Lopez Matthews was the very-consistent starting lineup, with DiVincenzo getting 24 starts. Ilyasova got 8 starts but those were very intermittent, may have been load-management days, and none since resumption in the bubble.

2. Next, let's pull out the bigs, since those aren't relevant to comparing teams' readiness to sustain an injury like ours to Hayward. Looking at the Play-by-Play table in Bk-Ref, both Lopez-es were 100% C, DJ Wilson was 38% C, and Giannis 24% C. So we're going to only exclude players named Lopez.

3. Then let's look at minutes-played and see if we can get some clarity on the top 5 or 6 non-bigs. Giannis Middleton Bledsoe Matthews have remained healthy and starting all season and into the playoffs, so they're the top 4. SG DiVincenzo (#5) and PG George Hill were right after them in both minutes and MPG, but Connaughton was pretty close to Hill in both regular-season and playoff usage, so I'm giving it to Hill partly on that edge but partly on his better On-Off +/- , and the fact that he's the #2 PG (behind Bledsoe) while Connaughton is the #3 SG (behind Matthews and DiVincenzo). So that puts Hill #6 and Connaughton #7.

4. As we get into the deeper bench, I'm looking partly at the game logs to see if they got consistent (but low) bubble minutes and even some playoff minutes, in judging the #7-9 vs "not really playing at all" edge cases. In Milwaukee's case, there was a pretty easy line to draw between the next two non-bigs and those after, just on total minutes played. Ilyasova (986) and Korver (960) both got ~16 mpg. Korver's gotten 76 playoff minutes and Ilyasova 23. Korver played every single bubble and playoff game, so we're calling him #8. Ilyasova missed only a few with a day-to-day elbow injury, so they both appear to be trusted for the back-of-rotation roles.

5. Complicating this, to your question, is PF Marvin Williams, who was bought out from Charlotte and signed for the vet min at the deadline (so he only got 321 minutes for the Bucks, but did so with 18.9 MPG). Position-wise, he's competing for backup PF (behind Giannis) mostly with Ilyasova. He's gotten more playoff minutes than Ilyasova. But looking at the totality of the post-trade-deadline game log, where their DNP counts are very comparable (5 vs 6), Williams has played 18 MPG to Ilyasova's 13 in the bubble, but a big chunk of that may have been largely attributed to Ilyasova's elbow situation. Looking at the fullness of their seasons, since they appear to be very comparable now, I have to give Ilyasova the edge for that #9 spot mostly due to the season-long role he's had including occasional starting duty. If Williams keeps on getting a much bigger playoff role vs Miami, then maybe he takes that spot. Either way, we're talking about end-of-rotation kind of guys here.

6. Everyone else, e.g. Sterling Brown, got only spot duty in the bubble and got limited minutes, or were depth signings not counted on for anything, such as Frank Mason (two-way contract) or Thanasis Antetokounmpo.

That's kinda the way I went through each team. When in doubt, I mostly focused on minutes totals (both regular season and post), and if they were close I looked at game logs and tried to see who had the bigger role post-deadline and post-restart. Some decisions, such as the above, will be arguable. But I don't think it really changes the analysis.
 
Last edited:

Jimbodandy

Member
SoSH Member
Jan 31, 2006
11,405
around the way
That's kinda the way I went through each team. When in doubt, I mostly focused on minutes totals (both regular season and post), and if they were close I looked at game logs and tried to see who had the bigger role post-deadline and post-restart. Some decisions, such as the above, will be arguable. But I don't think it really changes the analysis.
Your analysis is good and makes sense. It really draws out how unbelievably weak our scoring is off the bench, remarkable even to dorks who have watched every game for years.
 

lovegtm

Member
SoSH Member
Apr 30, 2013
11,997
Your analysis is good and makes sense. It really draws out how unbelievably weak our scoring is off the bench, remarkable even to dorks who have watched every game for years.
I don’t think the Celtics have the bench they want (as a clear championship contender), but they’re also a weird team. They have way more playmakers in their starters+Smart than anyone else does, so the bench doesn’t need that sparkplug bucket guy in the same way.
 

lovegtm

Member
SoSH Member
Apr 30, 2013
11,997
Moving past VORP and PER and into real-world revealed preferences in the playoffs:

In a nearly must-win game, Nurse played Siakam, Van Vleet, Lowry and OG 43, 42, 40, and 38 minutes respectively. He went to Ibaka at C because of Gasol's foul trouble.

The other guys who played: Powell, Davis, and Boucher, for 13, 5, and 9 minutes. They combined for 4 points.

This isn't new, either. Lowry (at age 34) and VanVleet played 36 mins/game this year.

The Celtics bench is very, very meh, but Nick Nurse's actions strongly disagree that Toronto's is much better.
 

Jimbodandy

Member
SoSH Member
Jan 31, 2006
11,405
around the way
I don’t think the Celtics have the bench they want (as a clear championship contender), but they’re also a weird team. They have way more playmakers in their starters+Smart than anyone else does, so the bench doesn’t need that sparkplug bucket guy in the same way.
I think that we're in full agreement. Even so, I was surprised at how bad the offensive numbers were.
 

Saints Rest

Well-Known Member
Lifetime Member
SoSH Member
I don’t think the Celtics have the bench they want (as a clear championship contender), but they’re also a weird team. They have way more playmakers in their starters+Smart than anyone else does, so the bench doesn’t need that sparkplug bucket guy in the same way.
I think this is very well said.

For me the ultimate "spark plug bucket guy" was the original "Microwave," Vinnie Johnson. But the Pistons desperately needed him to be that kind of player because they had multiple other starters who weren't playmakers in Salley, Mahorn, and Rodman.

Same with Lou Williams last year. The top 5 MPG Clips averaged almost 68PPG. LouWill added 20.

By comparison, the top 5 MPG Celts this year averaged 94.6PPG. (That number drops to 90.9 if you count Theis and not Marcus even though MS averaged about 8 MPG more than DT).
 

chilidawg

Member
SoSH Member
Jan 22, 2015
5,934
Cultural hub of the universe
Good piece in the Athletic this morning from John Hollinger. Highlights the play of the two Williams and the lowlights of Semi. He thinks Toronto ought to go small off the bench to try to spark the offense and take advantage of the C's lack of depth there. Not often we hear about teams needing to go small against us to counter our strength up front, but game 2 definitely saw Brad play a little bigger bench with GW at the 4 and Wanamaker only getting 9 minutes.

https://theathletic.com/2040785/2020/09/03/how-boston-is-winning-the-bench-battle-and-what-toronto-can-do-about-it/?source=dailyemail
 

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
31,191
Yeah, I agree with that, but I’d certainly want to look at luck-adjusted RAPM, PIPM, BPM, RPM or Bowiac’s own DARKO instead. I think any of those will tell you more.
That’s appropriate. The thing that remember about the above metrics (and I’ve been trying to get Bowiac to understand the nuance of this for literally years) is that they have sample size and matchup/use problems too. You can only measure what occurs, and so when you get into bench guys you have a smallish sample, with limited ability to measure correlations with how players are used (e.g. is a bench player often out there with 3-4 starters or with 3-4 bench players?) and role. That doesn’t invalidate what they measure but requires us to be honest about the limitations.

The easiest case to get ones head around is often backup centers. Many all in one metrics show a couple backups as being among best in league; it’s understandable as they play a very specific role time-wise, Gather useful stats while out there, and are often are highly efficient offensively. But they are limited in what they can do—if you played them 35 minutes a game you’d get killed, and I’d you asked them to guard wings most of them would get killed. So someone having a great on/off or whatever other metric tells you something, but can’t be compared to a different player in a different role because the big can’t play that role.

Ultimately, coaching makes a huge difference in bench play. This is why San Antonio has good numbers for years from bench players—they know how to deploy guys for roles. I suspect Celts bench is in practice better than metrics suggest for this reason...and I agree it is a weakness too.
 

lovegtm

Member
SoSH Member
Apr 30, 2013
11,997
Good piece in the Athletic this morning from John Hollinger. Highlights the play of the two Williams and the lowlights of Semi. He thinks Toronto ought to go small off the bench to try to spark the offense and take advantage of the C's lack of depth there. Not often we hear about teams needing to go small against us to counter our strength up front, but game 2 definitely saw Brad play a little bigger bench with GW at the 4 and Wanamaker only getting 9 minutes.

https://theathletic.com/2040785/2020/09/03/how-boston-is-winning-the-bench-battle-and-what-toronto-can-do-about-it/?source=dailyemail
I don’t see how going small helps them a ton. The Celtics would just play Wanamaker more, who is good enough catch-and-shoot and driving closeouts that the offense stays humming.

The real issue is that Gasol and Ibaka have been fairly effective at the rim, and small negates that.
 

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
31,191
Good piece in the Athletic this morning from John Hollinger. Highlights the play of the two Williams and the lowlights of Semi. He thinks Toronto ought to go small off the bench to try to spark the offense and take advantage of the C's lack of depth there. Not often we hear about teams needing to go small against us to counter our strength up front, but game 2 definitely saw Brad play a little bigger bench with GW at the 4 and Wanamaker only getting 9 minutes.

https://theathletic.com/2040785/2020/09/03/how-boston-is-winning-the-bench-battle-and-what-toronto-can-do-about-it/?source=dailyemail
Interesting idea from Hollinger—but I gotta say only Houston might have the guys to like the net of going small against Celtics. Raps are giving up talent all over the court if they follow that plan. Maybe they can pace their way out of that (which I think is Hollinhger’s hope) but if I were a Raps fan I’d rather try my regular rotation another game and hope for better shooting.

He doesn’t hit on what I observed in game 1 and which to me is the real reason to go small: Raps have yet to show they can stop Boston penetration and it’s a must for them. Going small might help here, but it has to work as the back-line help will be gone.
 

DannyDarwinism

Member
SoSH Member
Jul 7, 2007
4,883
That’s appropriate. The thing that remember about the above metrics (and I’ve been trying to get Bowiac to understand the nuance of this for literally years) is that they have sample size and matchup/use problems too. You can only measure what occurs, and so when you get into bench guys you have a smallish sample, with limited ability to measure correlations with how players are used (e.g. is a bench player often out there with 3-4 starters or with 3-4 bench players?) and role. That doesn’t invalidate what they measure but requires us to be honest about the limitations.

The easiest case to get ones head around is often backup centers. Many all in one metrics show a couple backups as being among best in league; it’s understandable as they play a very specific role time-wise, Gather useful stats while out there, and are often are highly efficient offensively. But they are limited in what they can do—if you played them 35 minutes a game you’d get killed, and I’d you asked them to guard wings most of them would get killed. So someone having a great on/off or whatever other metric tells you something, but can’t be compared to a different player in a different role because the big can’t play that role.

Ultimately, coaching makes a huge difference in bench play. This is why San Antonio has good numbers for years from bench players—they know how to deploy guys for roles. I suspect Celts bench is in practice better than metrics suggest for this reason...and I agree it is a weakness too.

I agree with all this (I referenced Amir Johnson as a kind of shorthand for the back-up big problem you lay out- Nick Collison is the more commonly cited one), but a larger sample size should, at least in theory, reduce some of the noise and collinearity issues, which is why I'd prefer the multi-year PIPMs and RAPMs. Obviously the multi-year stuff isn't much use for rookies or guys without a lot of NBA minutes under their belts so it's not particularly helpful to our current discussion of the Celtics bench, but it's useful to look at for guys with a large enough sample of minutes. For years, for example, I was told that Demar Derozan's terrible (and basically unprecedented for a guy with an All Star rep) on/off numbers were a product of the Raptors bench and rotations as an excuse for why his teams so consistently were better when he wasn't playing, but that bench changed, he got traded, and his numbers remained terrible. His career PIPM at least reflects that he's awful on defense and overrated on offense (-1.29, 0.54 respectively, for overall -0.74). His PER, on the other hand, was higher than Jayson Tatum's this year (21.6 vs 20.4), despite the fact that the Spurs got waxed when he was on the floor and the Celtics were much, much better when Tatum was playing. And of course, PER has the same problem of over-inflating back up bigs- Trez Harrell and Enes Kanter in the top 20 with 0.3 and -1.9 PIPMs, whereas Tatum was 7th overall in PIPM (4.66) and tied for 44th with Nerlens Noel in PER.

I don’t want to give these tools too much credence, and I’d rather hear/read sensible analysis from knowledgeable people based on what they’ve seen recently, but if we are going to try to quantify stuff, I think they’re the best we have access to.
 

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
31,191
I agree with all this (I referenced Amir Johnson as a kind of shorthand for the back-up big problem you lay out- Nick Collison is the more commonly cited one), but a larger sample size should, at least in theory, reduce some of the noise and collinearity issues, which is why I'd prefer the multi-year PIPMs and RAPMs. Obviously the multi-year stuff isn't much use for rookies or guys without a lot of NBA minutes under their belts so it's not particularly helpful to our current discussion of the Celtics bench, but it's useful to look at for guys with a large enough sample of minutes. For years, for example, I was told that Demar Derozan's terrible (and basically unprecedented for a guy with an All Star rep) on/off numbers were a product of the Raptors bench and rotations as an excuse for why his teams so consistently were better when he wasn't playing, but that bench changed, he got traded, and his numbers remained terrible. His career PIPM at least reflects that he's awful on defense and overrated on offense (-1.29, 0.54 respectively, for overall -0.74). His PER, on the other hand, was higher than Jayson Tatum's this year (21.6 vs 20.4), despite the fact that the Spurs got waxed when he was on the floor and the Celtics were much, much better when Tatum was playing. And of course, PER has the same problem of over-inflating back up bigs- Trez Harrell and Enes Kanter in the top 20 with 0.3 and -1.9 PIPMs, whereas Tatum was 7th overall in PIPM (4.66) and tied for 44th with Nerlens Noel in PER.

I don’t want to give these tools too much credence, and I’d rather hear/read sensible analysis from knowledgeable people based on what they’ve seen recently, but if we are going to try to quantify stuff, I think they’re the best we have access to.
I also like Tiago Splitter to make the point---some argued he was a hugely undervalued asset based on all-in-one metrics on Spurs, then he went to Hawks and proved he was, in fact, a creation of the system.

I fully agree those metrics are all better than PER, and should have been clearer that my point was about how accurately we can use RAPM, etc. to evaluate Boston's bench vs Toronto's bench. It is a good starting place, and a helpful shorthand, and it is also quite limited in application to a particular series. I do think Celtics bench is a weakness, but it is likely not as big a weakness as it appears because they have different credible players who can play different roles based on matchups...which is precisely what RAPM, etc. do not have sufficient ability to predict/measure. They are not using Kanter because he'd be terrible in this matchup, and they will use Semi/Granite differently based on matchup and who else is out there. Those kinds of adjustments get washed out in the regression analysis of any of the metrics because the sample size is just too small and it's just too hard to fully adjust to who else is out there. No criticism of the metrics---there is amazing work going on to get where we are---but we also should recognize the limitations.
 

DeJesus Built My Hotrod

Well-Known Member
Lifetime Member
SoSH Member
Dec 24, 2002
48,205
I really want to look at the Celtics internal metrics. Again, I suspect they would show you some fascinating things about players like Semi etc. When you watch the adjustments Stevens makes, its plain as day that they are reacting to not just data but data that has been refined to isolate the variables they are looking to control.
 

PedroKsBambino

Well-Known Member
Lifetime Member
SoSH Member
Apr 17, 2003
31,191
Fully agree. There is zero chance Brad is playing Semi because he doesn't realize his limitations---he has a specific belief about what he's good at/how he impacts the game. And I agree it is almost surely analytically-derived or scouting derived and analytically-validated.

I think Kanter's use has similar underpinnings, those are just easier for us to discern because they roughly line up with our amateur scouting assessment of the speed/bulk tradeoff of the opposing center.
 

JakeRae

Member
SoSH Member
Jul 21, 2005
8,125
New York, NY
Fully agree. There is zero chance Brad is playing Semi because he doesn't realize his limitations---he has a specific belief about what he's good at/how he impacts the game. And I agree it is almost surely analytically-derived or scouting derived and analytically-validated.

I think Kanter's use has similar underpinnings, those are just easier for us to discern because they roughly line up with our amateur scouting assessment of the speed/bulk tradeoff of the opposing center.
I doubt Brad has a very different assessment of Semi than we do. The bottom line is that Brad needs wing minutes and has limited options for those minutes while playing down a man. That he understands Semi’s limitations is evidenced by the fact that as Grant Williams has looked more and more trustworthy, Semi has played less and less.
 

DeJesus Built My Hotrod

Well-Known Member
Lifetime Member
SoSH Member
Dec 24, 2002
48,205
I doubt Brad has a very different assessment of Semi than we do. The bottom line is that Brad needs wing minutes and has limited options for those minutes while playing down a man. That he understands Semi’s limitations is evidenced by the fact that as Grant Williams has looked more and more trustworthy, Semi has played less and less.
To be clear, I agree with this - Semi's limitations are his limitations regardless of how you measure them. My point is that they have cuts of information that show very specific player attributes and match-up results that we can only dream of. Its not that Ojeleye is better than we think - its that he has situational value and I would be very interested to look at the supporting data.
 

Jimbodandy

Member
SoSH Member
Jan 31, 2006
11,405
around the way
I really want to look at the Celtics internal metrics. Again, I suspect they would show you some fascinating things about players like Semi etc. When you watch the adjustments Stevens makes, its plain as day that they are reacting to not just data but data that has been refined to isolate the variables they are looking to control.
Yeah, I suspect that we could construct a lot of what they're looking at from their comments over the years (CBS in particular), but not the formulas per se.

Deflections, hockey assists and other counting stats related to passing, contesting 3s, hitting 3s by category, defensive FG%, average distance from shooter, etc. We'd guess some of them anyway.
 

Eddie Jurak

canderson-lite
Lifetime Member
SoSH Member
Dec 12, 2002
44,475
Melrose, MA
Stevens just gave the definitive answer of what he thinks of his bench. In the double OT win, he gave the bench a mere 39 minutes, exactly none of it coming during the 4th quarter or OTs. Daniel Theis played 47 minutes. Coming into this game, Theis had played in 209 NBA games, including playoffs, and never played 40+ minutes in any of them - in fact he only even broke the 30 minute mark in 14 games, all of those coming in this series against Toronto.

If you measure the bench by the level of trust the coach has in it, the Celtics bench can only be described as awful.
 

128

Member
SoSH Member
May 4, 2019
10,016
Stevens just gave the definitive answer of what he thinks of his bench. In the double OT win, he gave the bench a mere 39 minutes, exactly none of it coming during the 4th quarter or OTs. Daniel Theis played 47 minutes. Coming into this game, Theis had played in 209 NBA games, including playoffs, and never played 40+ minutes in any of them - in fact he only even broke the 30 minute mark in 14 games, all of those coming in this series against Toronto.

If you measure the bench by the level of trust the coach has in it, the Celtics bench can only be described as awful.
Wanamaker, Grant Williams and Time Lord actually were solid last nite, but I wouldn't have trusted them in the OTs, either, though perhaps Wanamaker could have hit a shot.
 

Eddie Jurak

canderson-lite
Lifetime Member
SoSH Member
Dec 12, 2002
44,475
Melrose, MA
Wanamaker, Grant Williams and Time Lord actually were solid last nite, but I wouldn't have trusted them in the OTs, either, though perhaps Wanamaker could have hit a shot.
There were virtually no bench minutes in the entire second half.
 

Eddie Jurak

canderson-lite
Lifetime Member
SoSH Member
Dec 12, 2002
44,475
Melrose, MA
True, but my point was that the bench was not, in reality, awful in Game 6. In general, though, it's clearly not up to par.
Sure. But whether Stevens' fault or the bench's fault, they do not have his confidence, and that is a problem.
 

128

Member
SoSH Member
May 4, 2019
10,016
Apparently the reigning champs have the same problem.
Yeah, Toronto really goes only seven deep: starters plus Ibaka and Powell. In Game 6, though, they played 59 minutes and scored 36 points between them.