RealGM Top 100 List #42

[Clyde Frazier]

Seeing Sheed being mentioned so early, I'd like to ask - how would you compare him to Horace Grant (or even Buck Williams), guys? For Buck, we don't really have impact stats, but we know that Horace looked extremely well in NPI DRAPM in 1997 (+2.52).

This isn't a perfect comparison, but both rasheed and grant had 5 year stretches of deep playoff runs. Rasheed's came a little later in his career while grant was in the middle of his prime:

RASHEED PLAYOFFS 04-08

per game - 13.6 PPG, 7 RPG, 1.6 APG, .9 SPG, 1.7 BPG, 1.6 TOPG
per 100 - 22.2 PPG, 11.5 RPG, 2.6 APG, 1.4 SPG, 2.7 BPG, 2.5 TOPG

42.8% FG, 33.2% 3PT, 71.5% FT, 50.7% TS, 102/97 OFF/DEF RTG, .126 WS/48

GRANT PLAYOFFS 91-95

per game - 12.7 PPG, 8.8 RPG, 2.4 APG, 1 SPG, 1.2 BPG 1.1 TOPG
per 100 - 17.8 PPG, 12.3 RPG, 3.4 APG, 1.5 SPG, 1.7 BPG, 1.5 TOPG

55% FG, 71.5% FT, 58.5% TS, 126/107 OFF/DEF RTG, .174 WS/48

Granted (no pun intended...), Grant was on some very strong bulls teams during his runs, but I can't ignore his +19 net RTG in OFF/DEF RTG to rasheed's +5. While rasheed was certainly reliable for DET and more versatile offensively, I think grant was a more consistent playoff performer across the board. You always knew what you were getting with grant, and i'm not referring to rasheed's attitude. He just seemed to have more of a "sprawling" impact for the lack of a better word. Rasheed's per 100 edge in scoring isn't that significant when you look at the stark difference in efficiency.

[drza]

Spoiler:

Seeing Sheed being mentioned so early, I'd like to ask - how would you compare him to Horace Grant (or even Buck Williams), guys? For Buck, we don't really have impact stats, but we know that Horace looked extremely well in NPI DRAPM in 1997 (+2.52).

This isn't a perfect comparison, but both rasheed and grant had 5 year stretches of deep playoff runs. Rasheed's came a little later in his career while grant was in the middle of his prime:

RASHEED PLAYOFFS 04-08

per game - 13.6 PPG, 7 RPG, 1.6 APG, .9 SPG, 1.7 BPG, 1.6 TOPG
per 100 - 22.2 PPG, 11.5 RPG, 2.6 APG, 1.4 SPG, 2.7 BPG, 2.5 TOPG

42.8% FG, 33.2% 3PT, 71.5% FT, 50.7% TS, 102/97 OFF/DEF RTG, .126 WS/48

GRANT PLAYOFFS 91-95

per game - 12.7 PPG, 8.8 RPG, 2.4 APG, 1 SPG, 1.2 BPG 1.1 TOPG
per 100 - 17.8 PPG, 12.3 RPG, 3.4 APG, 1.5 SPG, 1.7 BPG, 1.5 TOPG

55% FG, 71.5% FT, 58.5% TS, 126/107 OFF/DEF RTG, .174 WS/48

Granted (no pun intended...), Grant was on some very strong bulls teams during his runs, but I can't ignore his +19 net RTG in OFF/DEF RTG to rasheed's +5. While rasheed was certainly reliable for DET and more versatile offensively, I think grant was a more consistent playoff performer across the board. You always knew what you were getting with grant, and i'm not referring to rasheed's attitude. He just seemed to have more of a "sprawling" impact for the lack of a better word. Rasheed's per 100 edge in scoring isn't that significant when you look at the stark difference in efficiency.

To be fair, the part of Sheed's career that would most separate him from a player like Grant is the Portland part. From the numbers you posted you could argue that Grant was better at being a high-efficiency/low volume scoring option when their roles were similar, but it doesn't tell you a lot about their relative defensive impacts.

However, if you used Sheed's 5-year playoff averages from Portland I feel like it would tell a more accurate story (at least offensively) about the difference between Sheed and Grant. Using your format, specifically for the offensive categories:

RASHEED PLAYOFFS 04-08
per game: 17.3 ppg, 1.9 apg, 1.2 TOpg
per 100: 25.4 p100, 2.7 ap100, 1.8 TOp100
46.9% FG, 40% 3PT, 73.8% FT, 54.1% TS, 112 ORtg, .143 WS/48

Grant was still the higher efficiency scorer, but I think this makes it clearer that Sheed was in a much higher volume scoring role and his own efficiency was a lot better than in the numbers you posted. Again, I'm all for Grant getting more credit if it seems to be warranted. But Sheed's Portland years, with him making huge impacts on both offense and defense in a more focal role, are what separates him from players like Grant or Williams.

[trex_8063]

With Isiah Thomas voted in at #39 (now being three spots beyond that), Bob Cousy is a guy I feel should be gaining traction, as there are (imo) quite a few parallels and similarities between them. As consequence, I’ve always had the two ranked fairly close together (never more than 5 spots apart on my personal ATL).

Similarities:
*Similar career length and arc.
**Both are relatively high-scoring/high volume-shooting PG’s who have been criticized for somewhat poor shooting efficiency. Although relative to league standard, Cousy wasn’t quite as inefficient as Thomas: Thomas had just one year on the “+” side of league average (‘86, by +1.35%); Cousy was actually above league average for six consecutive seasons from ‘52-’57 (by as much as +2.5%, and by an average of +1.0%).
Thomas career relative TS%: -2.1%
Cousy career relative TS%: -1.0%
***Both led a #1 offense in their respective careers (Cousy actually led THREE #1 offenses, from ‘53-’55).
****However, both were offensive stars who achieved their greatest team success on squad’s who won with their defense. They were each perceived as the driving force behind the offense on these squads.
*****Both are multi-time champions (Cousy obv more so).
******Both were perennial All-Stars, often All-NBA team, too. Cousy wins out in accolades, although obviously era considerations apply.
*******Both had a dazzling style of play (and handles) that made them very popular with the fan-base of the time.

Here is the overall snap-shot of Cousy’s career:
Extended Prime Cousy (‘52-’61)--697 rs games
Estimated Per 100 Possessions (rs): 21.9 pts, 6.1 reb, 8.8 ast @ 44.9% TS% (-0.4% to league)
PER 20.1, .139 WS/48 in 37.4 mpg
Estimated Per 100 Possessions (playoffs): 20.3 pts, 5.5 reb, 8.8 ast @ 43.5% TS%
PER 18.0, .121 WS/48 in 40.7 mpg
Career rs WS: 91.1
Career playoff WS: 9.1

6-Time NBA Champion (at least 3 as the 2nd-best/most important player, never less than the 4th-best player)
13-Time All-Star
10-Time All-NBA 1st Team
2-Time All-NBA 2nd Team
1-Time league MVP (‘57, ahead of Bob Pettit)
#36 All-Time in MVP Award Shares
*Seven times he factored into the MVP voting (despite the MVP not even being awarded his first handful of seasons), never lower than 8th (four times in the top 4)
#33 All-Time in RealGM RPoY shares

The guy was still a presence to be reckoned with late in his career. At age 34---in a league that was becoming integrated and had quite a few athletic stars (Russell, Wilt, West, Oscar, Baylor, Pettit, as well as Bellamy, Wilkens, etc)---he'd still averaged 13.2/2.5/6.8 for the best team in the land; and in his final game as a Celtic (game 6 of the finals), despite sprained ankle late in game, came up with 18 pts while helping to hold Frank Selvy and Dick Barnett to a combined 15 pts.

Not saying Cousy is my vote for this spot (though I’m considering him), but it certainly seems he ought to be in the discussion by this time, especially given our rank of Isiah Thomas. For that matter, I feel Allen Iverson belongs in the near vicinity of these two guys, as well.

Otherwise, I’m strongly considering Elvin Hayes for this spot. Can also see arguments for Dolph Schayes, Dwight Howard, Dave Cowens, or Robert Parish. Will try to write more on them later.

[Clyde Frazier]

Spoiler:

Seeing Sheed being mentioned so early, I'd like to ask - how would you compare him to Horace Grant (or even Buck Williams), guys? For Buck, we don't really have impact stats, but we know that Horace looked extremely well in NPI DRAPM in 1997 (+2.52).

This isn't a perfect comparison, but both rasheed and grant had 5 year stretches of deep playoff runs. Rasheed's came a little later in his career while grant was in the middle of his prime:

RASHEED PLAYOFFS 04-08

per game - 13.6 PPG, 7 RPG, 1.6 APG, .9 SPG, 1.7 BPG, 1.6 TOPG
per 100 - 22.2 PPG, 11.5 RPG, 2.6 APG, 1.4 SPG, 2.7 BPG, 2.5 TOPG

42.8% FG, 33.2% 3PT, 71.5% FT, 50.7% TS, 102/97 OFF/DEF RTG, .126 WS/48

GRANT PLAYOFFS 91-95

per game - 12.7 PPG, 8.8 RPG, 2.4 APG, 1 SPG, 1.2 BPG 1.1 TOPG
per 100 - 17.8 PPG, 12.3 RPG, 3.4 APG, 1.5 SPG, 1.7 BPG, 1.5 TOPG

55% FG, 71.5% FT, 58.5% TS, 126/107 OFF/DEF RTG, .174 WS/48

Granted (no pun intended...), Grant was on some very strong bulls teams during his runs, but I can't ignore his +19 net RTG in OFF/DEF RTG to rasheed's +5. While rasheed was certainly reliable for DET and more versatile offensively, I think grant was a more consistent playoff performer across the board. You always knew what you were getting with grant, and i'm not referring to rasheed's attitude. He just seemed to have more of a "sprawling" impact for the lack of a better word. Rasheed's per 100 edge in scoring isn't that significant when you look at the stark difference in efficiency.

To be fair, the part of Sheed's career that would most separate him from a player like Grant is the Portland part. From the numbers you posted you could argue that Grant was better at being a high-efficiency/low volume scoring option when their roles were similar, but it doesn't tell you a lot about their relative defensive impacts.

However, if you used Sheed's 5-year playoff averages from Portland I feel like it would tell a more accurate story (at least offensively) about the difference between Sheed and Grant. Using your format, specifically for the offensive categories:

RASHEED PLAYOFFS 04-08
per game: 17.3 ppg, 1.9 apg, 1.2 TOpg
per 100: 25.4 p100, 2.7 ap100, 1.8 TOp100
46.9% FG, 40% 3PT, 73.8% FT, 54.1% TS, 112 ORtg, .143 WS/48

Grant was still the higher efficiency scorer, but I think this makes it clearer that Sheed was in a much higher volume scoring role and his own efficiency was a lot better than in the numbers you posted. Again, I'm all for Grant getting more credit if it seems to be warranted. But Sheed's Portland years, with him making huge impacts on both offense and defense in a more focal role, are what separates him from players like Grant or Williams.

Right, that's why I said it wasn't perfect. I liked that the sample size and the team's accomplishments were similar, so I went there first. Rasheed's role was more prominent offensively in POR, but I think his utter lack of rebounding was also a concern. For an all time ranking, there's a good chance i'd still side with rasheed, but I don't believe he belongs in this range. I'm all for the discussion that RAPM data brings to the table, but in this case it isn't enough for me to seriously consider rasheed here.

[D Nice]

So when the RAPM results from his prime DID come out and a) support that he was having a consistently positive offensive impact despite the common criticisms that he shot too many jumpers and b) supported that he was repeatedly measuring out as a monster defensive presence consistently in both Portland and Detroit...for me, as I mentioned before, it kind of puts his career in a different perspective than I would have had without the data.

Once again I urge you to take his case against players of this caliber (as it stands solely on the basis of his D-RAPM) at extreme caution. He joined a team that was #4 in league DRTG the season prior and #2 (basically tied with Indy) that year prior to his arrival WHILE replacing a rather poor defender in Corliss Williamson. If you find yourself thinking he was in the same galaxy as [4x DPOY] Ben Wallace defensively you need to re-evaluate.

In Portland, with fairly strong defensive casts overalll between '98 & '03 (particularly the teams where Pippen was still useful and you had Bonzi/Grant/Sabs/Augmon) his teams ranked 8th, 6th, 5th, 9th, 13th, 13th. The Wallace-led Pistons ranked 8th, 8th, and 4th while playing awful defenders like Chucky Atkins, Jery Stackhouse, and Corliss Williamson MAJOR minutes. Even adjusting for conference I'd say this is a huge gap in “impact" or "lifting."

Antoine Walker led the league in D-RAPM (2.9) in 2002 on the #5 league defensive seemingly because he played power forward and for no other reason. In 2003 he was top 10 (1.9) and they were #8. This is Antoine friggin Walker we are talking about and it’s not like Walter McCarty (his backup) is Amare or Boozer.

Now, like Doc MJ, I'm not ready to just shoot him up the rankings to where he'd be if I took RAPM results as gospel (his RAPM scores were in the same vicinity as guys that got voted in long ago, for instance, so by RAPM alone he should have been supported back in the 20s (using DocMJ's 5-year averages method, Sheed slotted in just behind Nash and ahead of both Kidd and Paul). The other career considerations (e.g. attitude, line between focal point and excellent role player, etc.) do exist.

I have no idea what the scaling methodology is but unless you are talking about years we absolutely don’t have actual data for I would suggest not giving it much weight. I don’t know why you would want to add more bias and noise to data by collating information from different seasons, particularly when seasons fall 3-5 years apart. WAY too much changes from season to seasons. Guys get better every year, then level off for 4-5 years, then get worse every year. Teammates change. Coaching changes. An entire 82-game season does not need to be treated like an “outlier” that requires "smoothing” based on data from different environments (seasons). It’s missing the point of what we’re actually after, which is what was a guy’s impact on season-by-season basis period.

Bringing it back to the question, I don't think that Grant or Williams showed the ability to be a featured offensive player like Sheed and I don't have as much evidence that they were as good on defense. But I'm also open to that line of thought, and if it can be supported that Grant and Williams WERE dominant defenders that were huge different makers I wouldn't have any problem with seeing them in a better vein than they are currently mentioned.

As for Horace Grant, he pretty much is Rasheed minus the first option talent except that Sheed didn’t want to be a first option half of the time. Sheed has more range but Horace was much more efficient/effective from mid range (making Horace a more useful 3rd-option type IMO). I don’t see much meaningful separation on D. Sheed was better but not extremely so. Horace was either the best defensive PF of the 90s or 2nd to Dennis Rodman. I don’t feel that Sheed was a better defender than Worm so he’d fall around the same range (defensively) if he were slotted into the 90s (interestingly enough, Worm at 4 is probably the best stylistic comparison for Wallace on D).

To be clear I think Sheed is clearly a better talent than Horace, but I also think he’s much closer to Grant than he is to Dwight/Zo/McHale. FWIW Horace was the best/highest impact defender on the 2001 Lakers in the playoffs, a team that boasted an all-time-great defensive rating. And this is old-man Horace at the time (age 35) so if the argument is that Sheed’s D scales better…well…it doesn’t. Horace did as much or more covering for Shaq than Sheed ever did for Sabonis, and I believe there have been some studies that show Grant having a comparable impact to Jordan/Pippen on the early 90s Bulls defenses.

Spoiler:

This isn't a perfect comparison, but both rasheed and grant had 5 year stretches of deep playoff runs. Rasheed's came a little later in his career while grant was in the middle of his prime:

RASHEED PLAYOFFS 04-08

per game - 13.6 PPG, 7 RPG, 1.6 APG, .9 SPG, 1.7 BPG, 1.6 TOPG
per 100 - 22.2 PPG, 11.5 RPG, 2.6 APG, 1.4 SPG, 2.7 BPG, 2.5 TOPG

42.8% FG, 33.2% 3PT, 71.5% FT, 50.7% TS, 102/97 OFF/DEF RTG, .126 WS/48

GRANT PLAYOFFS 91-95

per game - 12.7 PPG, 8.8 RPG, 2.4 APG, 1 SPG, 1.2 BPG 1.1 TOPG
per 100 - 17.8 PPG, 12.3 RPG, 3.4 APG, 1.5 SPG, 1.7 BPG, 1.5 TOPG

55% FG, 71.5% FT, 58.5% TS, 126/107 OFF/DEF RTG, .174 WS/48

Granted (no pun intended...), Grant was on some very strong bulls teams during his runs, but I can't ignore his +19 net RTG in OFF/DEF RTG to rasheed's +5. While rasheed was certainly reliable for DET and more versatile offensively, I think grant was a more consistent playoff performer across the board. You always knew what you were getting with grant, and i'm not referring to rasheed's attitude. He just seemed to have more of a "sprawling" impact for the lack of a better word. Rasheed's per 100 edge in scoring isn't that significant when you look at the stark difference in efficiency.

Why are you using +/- ORTG/DRTG? ORTG for players is an individual efficiency estimation stat. DRTG is what the team actually allowed when said player was on the floor (aka most of the time) with a simple (useful) pace adjustment. They aren’t related in any capacity. Not singling you out, I’ve seen it elsewhere, but it doesn’t mean anything, at least not what I think you think it means. ORTG/DRTG "splits” or “differentials" are pointless unless you are doing them for a team.

[Basketballefan]

I can't believe Carmelo Anthony hasn't been voted in. I know the melo hate is strong in this forum, but come on, cp3 and durant high than melo on the all time list!? This is to the point of stupidity.

Vote: Carmelo Anthony

Cp3 and Kd are flat out better players than Melo so it's only right that they went higher.

I'm not a Melo stan nor a Melo hater but i can say he doesnt deserve traction for at least another 10-15 spots.

[ronnymac2]

One thing to add about Alonzo Mourning: He lead the best defensive team in the league in 1997, a Miami Heat squad which eventually would slow down prime Michael Jordan and the Bulls in the ECF more than any other team Chicago faced in the 90's.

Jordan put up 30.2 points and 2.6 assists on 38.7% shooting (47.5%TS) with an individual ORTG of 108 (You can look through the series logs on bball-ref...I don't think MJ in the 90's ever had his playoff production made so inefficient in a playoff series...correct me if I'm wrong). Chicago, which lead the league with an incredible 114.4 ORTG in the REG SEA, was held to a 104 ORTG in the ECF vs. Miami. Alonzo's defense was a huge part of the reason why CHI's elite offense wasn't operating full capacity.

Take a look at 2000, Zo's peak year. After he crushed Detroit in the first round (18, 8.7, 4.7, 60.4%TS, 32 MPG), the classic NYK vs. MIA battle ensued. NY was the 6th-best defensive team in the league that year with a great defensive frontcourt anchored by Patrick Ewing, Kurt Thomas, and Marcus Camby.

Zo dropped 23.1 points, 10.6 rebounds, 2.7 blocks, 10.6 FTAs, 52.4%TS, 31.4 USG% with a 101 individual ORTG (higher than anybody on the MIA team who was within the same stratosphere as far as volume/USG) in 40 MPG over 7 slug-it-out, slow-paced games. He had the best game of his career, a 29 point, 13 rebound, 5 block, 59.4%TS monster line in a super slow Game 7.

[E-Balla]

Do you rate Sheed over Billups and Ben Wallace? If so, why? I tend to think the other two are the main reason for the Pistons' success and Sheed was another very good complementary piece, a step up from Rip or Tay but more like them than leading the squad like Chauncey or the Fro.

Rebel without a Cause did a nice job with this question; do you have anything to add?

Billups is close but he's easily better than Ben. Big Ben lost most of his defensive impact with his horrible offense IMO. Now Detroit Shees is worse than Ben but Sheed in 1999-02 is better than Ben at his best.

I do think people that were very impact minded should be voting Sheer but his techs were an issue. His hot headed antics got him kicked out of playoff games (game 1 against LA in 2000) and I'm sure refs were less likely to give his teammates a favorable call because they didn't like him.

My early vote would go to Paul Pierce. His longevity as a serviceable player (2001-13) compared to everyone else I'm considering is very impressive. Dwight and Zo have too many major problems (injuries and immaturity).

Pierce actually got voted in last thread. He's only been mentioned here as a continuation of conversations that began last thread, and as a foil for Rasheed Wallace.

FWIW, I was pretty vocal about having Sheed over Pierce in the last two threads, but I was pretty clearly in the minority on that front.

Well I guess changing it to Alonzo Mourning. I love Dwight's playoff performance over Zo's and it think Zo was only slightly better in the regular season while in the post season the gap is sizeable. Mourning just didn't hurt his team off the court like Dwight did and he never gave up on them. That's enough to make up for Dwight's one on one post scoring advantage.

[Doctor MJ]

Re: Drafting for a future franchise
I agree that this is a scenario where I would prefer Pierce. However in an All-Time list this is not what I am looking at. I would also prefer to draft Pierce over someone like Shaq, because well, you know Shaq is a troubled player to build around as welll (and will take the high road)

So when I look at a player, the first thing I try to do is to figure out the quality of a player.
Drza had a great post comparing Pierce and Rasheed in thread number 40 or 41 and I tried to lay out the quality of Sheed as well.
Then I try to apply certain things like personality, character and portability.

That's fine. For me it's not simply about draft status, but it's a anchor point that I look at.

From a perspective of Pierce vs Sheed, put everything in his Portland years in a basket and compare it to the comparable amount of time with Pierce in Boston, which franchise feels like they got the better deal? I think that's obviously Pierce.

From there, Sheed could surpass Pierce if the remainder of his career was clearly better than Pierce's, but to me it's at best only comparable.

With Shaq, as you may know, I really do hold his stupidity against him, but his career isn't one where he found his new stability as some kind of complimentary guy. Shaq was Shaq wherever he went and the only changes came with what his body could give.

Re: Dismissing Portland Rasheed because it is not sustainable
The fact that Sheed messed up in Portland after some time does not change the fact that he had 6 incredible prime seasons there. At least for me.
Seasons that eclipse Pierce by far.
(Shaq again, always burned bridges and imloded on teams after some time. Do we fault him for that too?)

Another question popping up,
Was Sheed really that problematic outside of the technicals?

His tenure in Detroit ,NY and Boston prove otherwise. His recent coaching gig in Detroit as well.
Sheed comes off as a good teammate and in the right circumstances totally under control.
So maybe it was Portland, the coach and the staff, that simply handled Rasheed the wrong way?

I'm not saying it should be ignored by any means, but to me it warrants looking at a bit more closely.

So for example, Sheed's really big RAPM numbers end in '99-00, which was also the Blazers big moment in the sun. That was only his 2nd season playing star-like minutes. So quite literally by his 3rd big minute season, the Blazers were going downhill and his crazy was becoming the face of what the franchise was.

Re: How problematic? Sheed didn't just get technicals. He spoke out publicly in problematic ways, he got into conflict with teammates, he got suspended for threatening people while not even on the court, and most fundamentally: He was by far the biggest presence in a locker room that quickly became legendarily bad.

The latter part I think people need to take seriously. If you're the best player, and you're vocal, you're leading whether your realize it or not. People are following you wherever you're heading. So when we have a star with a problematic attitude that very clearly cannot be explained by unique circumstances (Sheed is crazy, this is not debatable) and we then see major off court issues with a team, to me it's naive to assume they aren't related. And by "naive" what I mean is that our best estimate for what happened there is not "total coincidence".

Going back to the crazy, and your assessments of his times elsewhere, what I would say is that no one would claim he ever became normal. He was in the best of times a goofball, and in the worst of times much more problematic than that.

Re: Sheed dropped into a team with good organization and talent in Detroit
That is right where you wanna put any good player. A contender. I actually think the opposite. He made himself fit into that already strong team and lifted them over the top. "Portability",
Why would you wanna have an inept organization or franchise any way?
Maybe Pierce lifts worse or disorganized teams a bit more, but is that the goal and more important?
Those jailblazers are an examplfe for poor organization and toxic environment, yet how was Sheed still that great there?!

My point, if I'm remembering correctly, is not one about his ability but about the vibe. Much like with Rodman or Artest, if you can slot these guys into places where things are already going well and they need someone to fill the niche that player is most known for, most of the issues go away.

And this is a key thing for me: When people talk about primary and secondary guys with the notion that primary guys are more important, that doesn't always fly to me. Yes the secondary guy may play off the primary, but finding guys who can jack shots isn't actually that hard. But if by "secondary" guy we mean someone who needs others to build a structure in order for him to maintain his professionalism, that's a bigger deal to me.

Re: comparison with Ben Wallace
Sheed as a player, in my opinion, needs to be ranked over Ben Wallace and rather clearly. There is also no boxscore advantage for Ben, one that Pierce for example, had.
Ben does not even have half of the longevity and misses 2 way impact, even if he was a better defender.


Essentially it comes down to the quality of a player. Personality can drag someone down or make him look better, but by how much?

I really don't have a problem with the Sheed > Ben argument. There is some debate in my mind, but I think it's based on me getting emotional when I think of things like the Hall and legacy. Because of Ben's place as the face of that Piston team, along with the rarity of 4 DPOYs, to me he's much more a lock for the Hall than Sheed and frankly rightly so. But that's in part because it's the "Hall of Fame", not the "Hall of GOATs".

So yeah, while I'm still pretty set on Pierce and the other players being mentioned for the most part ranking ahead of Sheed, I am inclined to vote Sheed before Ben at this point.

[Doctor MJ]

Now, like Doc MJ, I'm not ready to just shoot him up the rankings to where he'd be if I took RAPM results as gospel (his RAPM scores were in the same vicinity as guys that got voted in long ago, for instance, so by RAPM alone he should have been supported back in the 20s (using DocMJ's 5-year averages method, Sheed slotted in just behind Nash and ahead of both Kidd and Paul). The other career considerations (e.g. attitude, line between focal point and excellent role player, etc.) do exist.

I have no idea what the scaling methodology is but unless you are talking about years we absolutely don’t have actual data for I would suggest not giving it much weight. I don’t know why you would want to add more bias and noise to data by collating information from different seasons, particularly when seasons fall 3-5 years apart. WAY too much changes from season to seasons. Guys get better every year, then level off for 4-5 years, then get worse every year. Teammates change. Coaching changes. An entire 82-game season does not need to be treated like an “outlier” that requires "smoothing” based on data from different environments (seasons). It’s missing the point of what we’re actually after, which is what was a guy’s impact on season-by-season basis period.

Can't tell if you're actually asking for an explanation of my scaled RAPM or you just used the same word. Let me know if you want me to elaborate.

As far more seasons adding noise, I actually don't understand what makes you think that. It's like you think I'm only showing 5-year averages and not the individual years. The individual years are there to and it basically goes without saying that it's wise to look at them, but if a person is looking to make a quick reference to the data, it's much better to give the 5-year averages than it is to just list the 1 season that was the peak for any player for whom we have plenty of data for and who had a normal career arc. Doing so makes for a less noisy shorthand, not a more noisy one.

[Doctor MJ]

Seems to me that if you were a big KG supporter, you should also be a big Sheed supporter. Obviously KG is the superior player in just about every way(3-pt shooting the clearest exception), but both are great big man defenders, but not traditional anchors. Both were clearly at their best as part of ensemble casts as opposed to being the clear "man". Both were the ultimate in team players despite always being the most talented guy on their team.

I agree that way too much gets made of the techs with Sheed. This wasn't really a bad guy as much as it was a bit of a hot head. Who cares really? In no way should he be be blamed for what happened in Portland. That's what happens when you just try and pile up talent without looking at character and leadership. He wasn't a leader. I think its appropriate to knock him a bit for that, but not everyone has to be the alpha.

I'm not personally ready to vote for him yet, but I do think he was a really good player for a long time. And in Portland he seemingly could do anything he wanted---he scored inside and outside, he defended, he rebounded. If anything he should have been more selfish on that team. But his team-first attitude proved to be the perfect fit on those criminally underrated Pistons squads that went to 7 straight ECF(and yeah I know the East sucked, tsherkin, but winning at least 2 playoff series for 7 straight years is really hard to do).

I think it would be more appropriate to say:

If you were a KG and Shaq supporter, you better think very hard before you pass on Sheed?

The thing for me is that I've always been very critical of Shaq, and been up front that I do hold him being a petty moron against him, so it's completely in character for me to do the same with Sheed.

You may ask how Shaq could possibly still be in my Top 10 if Sheed's not in my Top 50? Well:

1) Shaq's the most talented player ever, so I consider him only being in my Top 10 to be a fairly massive smackdown.

2) Sheed quite literally peaked in impact in his 2nd big minute season, while Shaq peaked in his 8th. Sheed's got a considerably bigger issue on this front in terms of his brain keeping him from reaching his potential.

3) Along those lines, but to make clear: Sheed was dangerous in a way that Shaq wasn't. Sheed had anger issues, broke the law, and threatened people. By comparison Shaq's issues were always just kid stuff.

I namechecked Artest before, I see Sheed much more like Artest than like Shaq, and it's my personal belief that Artest should have been banned from the league long ago - Stern was a fool not doing more on this front, Artest could have killed someone, and then where would the NBA be? I say this as someone who quite likes Artest actually, I consider his personality far more defensible than even Shaq, it's just that he's got big boy problems and poor impulse control.

[D Nice]

Now, like Doc MJ, I'm not ready to just shoot him up the rankings to where he'd be if I took RAPM results as gospel (his RAPM scores were in the same vicinity as guys that got voted in long ago, for instance, so by RAPM alone he should have been supported back in the 20s (using DocMJ's 5-year averages method, Sheed slotted in just behind Nash and ahead of both Kidd and Paul). The other career considerations (e.g. attitude, line between focal point and excellent role player, etc.) do exist.

I have no idea what the scaling methodology is but unless you are talking about years we absolutely don’t have actual data for I would suggest not giving it much weight. I don’t know why you would want to add more bias and noise to data by collating information from different seasons, particularly when seasons fall 3-5 years apart. WAY too much changes from season to seasons. Guys get better every year, then level off for 4-5 years, then get worse every year. Teammates change. Coaching changes. An entire 82-game season does not need to be treated like an “outlier” that requires "smoothing” based on data from different environments (seasons). It’s missing the point of what we’re actually after, which is what was a guy’s impact on season-by-season basis period.

Can't tell if you're actually asking for an explanation of my scaled RAPM or you just used the same word. Let me know if you want me to elaborate.

As far more seasons adding noise, I actually don't understand what makes you think that. It's like you think I'm only showing 5-year averages and not the individual years. The individual years are there to and it basically goes without saying that it's wise to look at them, but if a person is looking to make a quick reference to the data, it's much better to give the 5-year averages than it is to just list the 1 season that was the peak for any player for whom we have plenty of data for and who had a normal career arc. Doing so makes for a less noisy shorthand, not a more noisy one.

Ah, I gotcha, it's my mistake then. What I had gathered from the way fpiii DRZA and others were talking about it I thought there was a "smoothing" factor being applied via some sort of regression of the 5 seasons prior to whatever season was being calculated, not just an average of the 5 1-year data points. My mistake. That makes way more sense.

But yeah, while I've got you would love an explanation of what your sheet is doing with "scaled chronology" that is different from the standard NPI calculation. Although if you've already explained this before and it's easier you can just link me to a past breakdown.

[Basketballefan]

Ill Vote Allen Iverson. I suspect he probably doesn't get serious traction for another 3-5 spots but i think he's clearly deserving at this point.

He led the Sixers to 5 consecutive playoff appearances and 6 appearances overall. He carried that franchise for many years without a lot of offensive help, he's constantly criticized for his low efficiency, high volume shots which is fair, but somewhat unwarranted, he didn't have anyone to take defensive attention away from him.

He won MVP, Scoring title, and led his team to the finals all in one season. Overall his resume is pretty impressive:

11 time all star
MVP
4 scoring titles
3 time All NBA first team
3 time All NBA second team
Rookie of the year
etc

He is 46th in career PER for NBA and 48th for ABA/NBA combined.
22nd all time in career points in NBA and 27th for ABA/NBA combined.

Several years as a top 10 player and arguably even top 5 in 2001.

7th all time in career ppg average.
2nd all time in career playoff ppg average.

His peak is impressive as well : 33 3 7 54 ts% 26 PER..his scoring efficiency is nothing special here but there's nothing wrong with high volume on average efficiency.

So aside from maybe his later years, Iverson had a positive impact on his teams, even though some will deem him as a cancer.

I also disagree that he isn't capable of leading his team to a title, he was 3 games away from doing so and he happened to be up against a dynasty that had 2 top 10 players ever.

[ElGee]

One thing to add about Alonzo Mourning: He lead the best defensive team in the league in 1997, a Miami Heat squad which eventually would slow down prime Michael Jordan and the Bulls in the ECF more than any other team Chicago faced in the 90's.

Jordan put up 30.2 points and 2.6 assists on 38.7% shooting (47.5%TS) with an individual ORTG of 108 (You can look through the series logs on bball-ref...I don't think MJ in the 90's ever had his playoff production made so inefficient in a playoff series...correct me if I'm wrong). Chicago, which lead the league with an incredible 114.4 ORTG in the REG SEA, was held to a 104 ORTG in the ECF vs. Miami. Alonzo's defense was a huge part of the reason why CHI's elite offense wasn't operating full capacity.

Jordan's lowest box EV score was 5.4 (92 NYK), second lowest vs. 97 Heat (5.7)
Jordan's had 3 PS series with a TS% of -3% or worse: 97 Hawks (-3.4%), -4% (98 CHH) and 97 Heat (-4.5%)
Jordan's lowest GameScore was the 96 Finals (18.5), second worse was 97 Miami (18.7)

[Doctor MJ]

Ah, I gotcha, it's my mistake then. What I had gathered from the way fpiii DRZA and others were talking about it I thought there was a "smoothing" factor being applied via some sort of regression of the 5 seasons prior to whatever season was being calculated, not just an average of the 5 1-year data points. My mistake. That makes way more sense.

But yeah, while I've got you would love an explanation of what your sheet is doing with "scaled chronology" that is different from the standard NPI calculation. Although if you've already explained this before and it's easier you can just link me to a past breakdown.

Goes like this:

First off, I"m using PI RAPM here. I could do the same thing with NPI, but when trying to get a grasp on a player throughout his career, I think PI gives the better snapshot.

Now, with RAPM in general, the machine learning aspect of the algorithm disconnects the 1-to-1 relationship between actual basketball points and points in the stats' results. APM doesn't have this issue, only RAPM does. Within a given study the relative values of the players in RAPM are the most reliable you can get, but it means it's incorrect to compare one study to another without some kind of adjustment being made.

I made the adjustment with what I call 'normalized' or 'SD' RAPM, which just takes the standard deviation of a given study, and then divides the player numbers by that standard deviation. It's not a perfect method by any means, but to me it's a logical thing to do, and I've yet to hear anyone say it's not a good idea.

Then, the less important step: All things being equal, I'd like my data to not simply be consistent in its variance from year to year, I'd like to try to get that 1-to-1 relationship with actual basketball points back. So I found the standard deviation of Ilardi's 6-year APM study, and multiplied it with my normalized numbers to create the 'scaled RAPM'. Again, this is not a perfect approach, but it seems like a clear improvement.

[D Nice]

Ah, I gotcha, it's my mistake then. What I had gathered from the way fpiii DRZA and others were talking about it I thought there was a "smoothing" factor being applied via some sort of regression of the 5 seasons prior to whatever season was being calculated, not just an average of the 5 1-year data points. My mistake. That makes way more sense.

But yeah, while I've got you would love an explanation of what your sheet is doing with "scaled chronology" that is different from the standard NPI calculation. Although if you've already explained this before and it's easier you can just link me to a past breakdown.

Goes like this:

First off, I"m using PI RAPM here. I could do the same thing with NPI, but when trying to get a grasp on a player throughout his career, I think PI gives the better snapshot.

Now, with RAPM in general, the machine learning aspect of the algorithm disconnects the 1-to-1 relationship between actual basketball points and points in the stats' results. APM doesn't have this issue, only RAPM does. Within a given study the relative values of the players in RAPM are the most reliable you can get, but it means it's incorrect to compare one study to another without some kind of adjustment being made.

I made the adjustment with what I call 'normalized' or 'SD' RAPM, which just takes the standard deviation of a given study, and then divides the player numbers by that standard deviation. It's not a perfect method by any means, but to me it's a logical thing to do, and I've yet to hear anyone say it's not a good idea.

Maybe I’m missing something here, but it seems like you’re penalizing guys who had fantastic years (or help out the guys with awful ones) just so you can lessen the variance from year to year (I’m assuming when you use the word study, you are referring to 1-year data sets aside from the Ilardi reference, if not, I need you to clarify what vernacular means what when). Why do you want to get rid of variance when, invariably (pun intended) not all years are going to follow the same “impact curve” when regarding individual players? Sometimes it will, sometimes it won’t, I don’t really think it’s necessarily a “bad” thing when this occurs, it’s just reflecting what actually happened that particular season. I might be wrong, but it seems like you’re trading away some potential insight for the sake of added “comfort and consistency" when you do this.

I’m also not really getting when you refer to the standard deviation of a given study. Are to talking about player-to-player variance relative to past years? Or just the standard deviation of players in that particular year? Or are doing the calculations year by year, finding the standard deviation of the entire 15+ year data-set, and then going back and re-calculating the data season by season?

This really shouldn’t be hard for me to understand, and yet…I’m having trouble figuring out exactly what you did.

Then, the less important step: All things being equal, I'd like my data to not simply be consistent in its variance from year to year, I'd like to try to get that 1-to-1 relationship with actual basketball points back. So I found the standard deviation of Ilardi's 6-year APM study, and multiplied it with my normalized numbers to create the 'scaled RAPM'. Again, this is not a perfect approach, but it seems like a clear improvement.

This part makes sense to me.

[trex_8063]

Wanna talk about Elvin Hayes a little.......

Extended Prime Elvin Hayes ('69-'80)--long prime: 12 years, 978 rs games
Estimated Per 100 (rs): 24.5 pts, 14.7 reb, 2.1 ast, 1.3 stl, 2.7 blk with 3.2 tov @ 49.1% TS% (-1.8% to league)
PER 18.5, .127 WS/48 in huge 41.7 mpg. 102 ORtg/100 DRtg (+2)---data for '78-'80.
Better in playoffs.....
Estimated Per 100 (playoffs): 25.1 pts, 14.2 reb, 2.1 ast, 1.3 stl, 2.9 blk with 3.2 tov @ 50.4% TS%
PER 19.3, .140 WS/48 in 43.4 mpg. 103 ORtg/98 DRtg (+5)---data for only '78-'80.

His shot-selection and shooting efficiency obviously leave a fair bit to be desired, but he nonetheless appears to have produced a net positive result for the team offense on the teams he joined (more on that below). I do suspect that his ability to exert positive effect on offenses might be limited to bad offenses (the offenses he was joining were almost exclusively below average ones). But anyway....

He was a very good rebounder, and solid low-post defender who was also a fair rim-protector (and seems to consistently have exerted a positive effect on team defenses).
And then he's got near iron-man longevity. An extended prime that lasts like 12 years (during which he missed only 6 games); and only 9 missed games in 16 seasons . He was at worst a solid role-player in 15 of 16 seasons.


Some stuff on apparent impact......

In '68, the San Diego Rockets were 15-67 (dead last by a full 8 games to the next worst team) and -7.94 SRS (dead-last). They were the 12th-rated (of 12) team offensively and 10th of 12 defensively.

In '69 they lose Dave Gambee, John Barnhill, Jon McGlocklin, and an aging Johnny Green. Only noteworthy new acquisitions are rookies Rick Adelman and Elvin Hayes (same coach and everything).......they improve by 22 games (to 37-45) and 7.64 SRS pts (to -0.30, 7th of 14). Their ORtg improves by 2.2 relative to the league (now 12th of 14, as apposed to dead-last). DRtg improves by 4.7 relative to league (now 3rd of 14).
Admittedly, they never would quite get over mediocrity during Hayes's four seasons there; but that's a heck of jump from the extreme basement of the league (which I think they can mostly thank Hayes for).


In '72, the Baltimore Bullets---who had Wes Unseld, Archie Clark, Phil Chenier, and Jack Marin (all basically healthy and in their primes), along with at least a couple decent role players in Dave Stallworth and Mike Riordan----went 38-44, -1.26 SRS (10th of 17). They were 10th of 17 offensively, 9th of 17 defensively.

In '73, they lost Jack Marin, but otherwise still have all of the above characters (basically all healthy except Archie Clark who misses 43 games), same coach, too; only real noteworthy new acquisitions are rookie Kevin Porter (would only play 17.1 mpg his rookie season), and Elvin Hayes........they improve by 14 games to 52-30 and by 4.1 SRS pts to +2.84 (7th of 17). In ORtg, although their league rank fell from 10th to 12th of 17, they actually did improve by 0.9 relative to the league average. In DRtg, they improved by 3.2 relative to league (finishing 5th of 17).

Two years later they would be in the NBA finals. Three years after that they would win the title. Hayes would lead the league in playoff WS during that title run: 20.3 PER and .169 WS/48 in playoffs that year (20.7 ppg/11.9 rpg/1.6 spg/2.0 bpg @ .509 TS% in the finals).

Accolades/Ranks:
12-Time All-Star
3-Time All-NBA 1st Team
3-Time All-NBA 2nd Team
2-Time All-Def 2nd Team
#38 all-time in career rs WS (#58 all-time in career playoff WS)
#49 all-time in MVP Award Shares (twice finished 3rd in MVP voting)


Thoughts? I'm really thinking of casting in with him any time now.

[Doctor MJ]

Maybe I’m missing something here, but it seems like you’re penalizing guys who had fantastic years (or help out the guys with awful ones) just so you can lessen the variance from year to year (I’m assuming when you use the word study, you are referring to 1-year data sets aside from the Ilardi reference, if not, I need you to clarify what vernacular means what when). Why do you want to get rid of variance when, invariably (pun intended) not all years are going to follow the same “impact curve” when regarding individual players? Sometimes it will, sometimes it won’t, I don’t really think it’s necessarily a “bad” thing when this occurs, it’s just reflecting what actually happened that particular season. I might be wrong, but it seems like you’re trading away some potential insight for the sake of added “comfort and consistency" when you do this.

I’m also not really getting when you refer to the standard deviation of a given study. Are to talking about player-to-player variance relative to past years? Or just the standard deviation of players in that particular year? Or are doing the calculations year by year, finding the standard deviation of the entire 15+ year data-set, and then going back and re-calculating the data season by season?

This really shouldn’t be hard for me to understand, and yet…I’m having trouble figuring out exactly what you did.

If I'm understanding you correctly you're just wondering why I used PI instead of NPI. Is that right?

In general it's a reliability vs validity trade off, which is something I take seriously. I don't like sacrificing validity, but:

1) The entire reason that RAPM exists relative to APM, and PI exists relative to RPI, is that it's a major reliability gain and only a minor validity loss. On average you're going to get better data this way.

2) You're quite right that a player with an erratic career may get underrated by PI, but I know this going in. As you clearly see here, I'm certainly not rating guys solely based on how they did on the spreadsheet.

3) Beyond the typical benefit of reliability in a 1 year study, being able to view PI results back-to-back-to-back-to-back means that we're really able to see "the level" of any player with a consistent prime.

4) This is a subtle one:

APM typically prefers Garnett over Duncan
NPI RAPM typically prefers Duncan over Garnett
PI RAPM typically prefers Garnett over Duncan

Each method, from APM to NPI to PI is more reliable than the previous method, yet there's a systematic zig zag here that shows a wrinkle in the process. To attempt to put it into words:

APM: "Wow, it's amazing what Garnett is doing with such problematic teammates!"
NPI: "Don't be naive, that's noise that won't repeat. Regress him to the mean a bit."
PI: "Same thing happened last year though. Doesn't look like noise to me."

And of course, repeat this process year after year. I wouldn't bring this up if it only happened once. We're talking about a half decade's worth of independent yearly studies where this happens.

Re: What do I mean by 'study'. Everything in my study is from single-season PI RAPM studies done by either Engelmann or acrossthecourt.

Re: standard deviation. I'm literally taking the standard deviation of a given study. So take the data from all the players in the league within that study, and use a standard deviation function on it.

[lorak]

...

Could you answer to that: viewtopic.php?p=41315824#p41315824
?

[Moonbeam]

Should Dolph Schayes be getting some traction soon? He had a long career (for the time period) and consistently was among the top players in the league, with 6 All-NBA 1st team selections and a further 6 All-NBA 2nd team selections and 6 top 10-MVP finishes, 3 times in the top 5.

His field goal shooting was generally around league average, but his free throw shooting was amazing for a power forward, with a career mark of .849, including three league-leading seasons, topping out at .904 twice... as a power forward/center! He drew lots of free throws too with a career free throw rate of .512. Add it up, and he was quite an efficient scorer:

Code: Select all

Year    TS     Rel    Z
1950  .4799  +.0699  1.46
1951  .4681  +.0404  0.78
1952  .4684  +.0300  0.58
1953  .4952  +.0500  1.18
1954  .4981  +.0558  1.37
1955  .4896  +.0343  0.79
1956  .4973  +.0394  0.97
1957  .5015  +.0521  1.42
1958  .5076  +.0584  1.37
1959  .4879  +.0304  0.87
1960  .4960  +.0326  0.85
1961  .4816  +.0122  0.43
1962  .4611  -.0176 -0.33
1963  .4710  -.0216 -0.42
1964  .3986  -.0862 -1.99
Total .4877  +.0350  0.86


That efficiency came on decent volume, too, with 11 seasons among the top 10 in scoring. Add on a further 11 seasons in the top 10 in rebounds (including 9 in the top 5) and a reputation for solid defense, and he was clearly a premier player for a very long time.

Adding to the picture is that he increased his scoring (+1.2 per36), efficiency (+.015) rebounding (+0.4 per36) in the postseason, and he emerges as an elite player of the early NBA. His teams sported a .569 winning percentage overall in an era where variance in wins was lower than it is today.

For what they are worth, his advanced metrics look outstanding at this stage, with a truncated career WS/48 of 0.192 (missing his first two seasons due to minutes not being recorded, which were seemingly better than 1952, where he posted a WS/48 of 0.203), a truncated career PER of 22.0, and playoff averages of .189 WS/48 and 23.3 PER. Those lead to career rankings of 26th on WS/48 (including 29th on total WS) and 30th on PER. He's got a sizable metric edge on Bob Cousy and a notable longevity edge on Paul Arizin.

I'm not sure if I'll vote for him yet (leaning to Dwight or Zo here), but he's definitely on my radar.