Page 5 of 6
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Fri May 13, 2011 5:05 pm
by EvanZ
Just as a fun thought, what if Wade had ended up in Chicago? A lineup of Rose, Wade, Deng, Gibson(?), and Noah. Man, that would have been a good team.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Fri May 13, 2011 7:00 pm
by Crow
There are obvious challenges / potential problems with using average performance of lineups and individual players to make match-up choices and perhaps predict performance against specific counterparts and opposing lineups.
Coaches claim both a need and an ability to pick these match-ups better than the stats alone, but it would be possible with enough time and effort to compare a coach's actual match-up win%, the percent of times his choice of match-ups beat the average performance of the league against a certain player or lineup, for an entire season and compare that against a computer program written by a person which made selections entirely based on whatever model of past season average performance or past performance against "similar" players and lineups or against a person who used various past performance data and additional judgment. That would be interesting to see. It would add valuable detail beyond what can be seen in player and lineup stats over time, team point differential, win performance vs point differential expectations and Coaching Adjusted +/-.
Have any teams and their consultants gone this far? I would hope so. If you scored every match-up at player level and lineup level of every stint of every game for your coach for a full season or a few and at least a few other coaches for comparison I'd think you'd have a very useful database for understanding your coach, his style and strengths & weaknesses and seeing places for him to try to jump on the biggest opportunities and hit the most right button more often and places for him to avoid certain choices and make better ones.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Fri May 13, 2011 7:24 pm
by greyberger
This is something that Wayne Winston is talking about all the time. I'd be surprised if he hasn't taken a crack at it.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Fri May 13, 2011 8:34 pm
by Crow
Yeah, it would strengthen Winston's case if he could point to a superior match-up win% for his Adjusted +/- or Impact Rating over the coach's actual selections. With enough time & effort you could run a variety of metric based decision rules and hybrids and with the right program & skill maybe even reverse engineer the best performing decision rules and decision metric (perhaps tailored to lineup or even player) at least for that team and set of data. Using several seasons would probably help in finding the biggest good and bad moves with some more confidence. Focusing on the biggest with the highest significance would reduce chasing quite as much noise. But with enough testing you could get a better handle on how much "noise" to chase and when it becomes worthwhile on average (even for a fraction or tiny fraction of a point gain) over following a coach's conventional thinking.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Fri May 13, 2011 8:56 pm
by Crow
EvanZ wrote:Just as a fun thought, what if Wade had ended up in Chicago? A lineup of Rose, Wade, Deng, Gibson(?), and Noah. Man, that would have been a good team.
Hypotheticals like this are indeed interesting and important. It is hard to "know" with confidence ahead of time what that specific lineup would do... but you could study what the closest versions of that did on the Bulls with Rose (SG closest to Wade), maybe even the closest versions of that from the University of Memphis with Rose when with the highest scoring SG option (just to see if you could pick up any clues about interactions patterns to check further on or watch for) and the most similar lineups from your coach's history and system too and other NBA PGs like Rose with lineups with similar player types and qualities at the other positions and especially those from similar coaches / systems.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Fri May 13, 2011 10:07 pm
by Crow
Rose in 08-09 with Gordon was -2 per 48 minutes as was Rose / Salmons while the team as a whole was almost even on point differential. At first quick look their individual stats seemed pretty similar with and without each other. Rose-Hinrich was +5 per 48 and both shot a bit better than their overall marks and moreso the without split. Maybe more assisted shots back and forth between them.
Rose in 09-10 with Salmons was neutral while Hinrich-Rose must have slipped into negative territory and perhaps or probably because of it went away last summer. Team as a whole was negative.
In 10-11 Rose had positive team +/- with all the other main guard options.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sat May 14, 2011 9:54 am
by huevonkiller
It was a poor MVP choice, and Neil basically nailed it at basketball-reference. He's another Allen Iverson, and the stats support it once you adjust for league average.
The Bulls are a great team though and they are dangerous. Hopefully Erik ditches Z, he is such a terrible defender at this stage. Bibby is doing nothing for Miami either.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sat May 14, 2011 3:02 pm
by Mike G
It's a shame Neil doesn't post and reply over here. Oh, well.
I did this earlier, but it's melted down. Adjusted for league and for team, updated Derrick Rose similar players:
Code: Select all
diff yr per-36 rates : Sco Reb Ast PF Stl TO Blk
.00 2011 Derrick Rose 28.6 4.6 8.2 1.7 1.0 3.4 .6
.37 1996 Anfernee Hardaway 25.0 4.6 7.1 2.0 2.1 2.8 .5
.41 2006 Dwyane Wade 29.5 6.0 6.9 2.7 1.8 3.4 .7
.42 2005 Kobe Bryant 27.5 5.7 5.6 2.5 1.2 3.8 .8
.42 2006 Tony Parker 25.7 3.8 7.6 2.2 1.1 3.4 .1
.44 1970 Jerry West 27.2 3.6 6.4 2.0 2.1 3.4 .3
.45 2002 Jerry Stackhouse 23.8 4.5 5.9 2.2 1.1 3.7 .5
.46 2006 Lebron James 30.8 6.6 6.1 2.0 1.3 2.8 .7
.47 1996 Terrell Brandon 24.9 4.3 7.7 2.1 1.9 2.0 .5
.48 2004 Tracy Mcgrady 26.4 5.7 5.4 1.8 1.3 2.5 .6
.50 2005 Allen Iverson 28.3 3.7 7.3 1.7 2.2 4.2 .1
.53 1967 Oscar Robertson 24.9 4.1 8.1 2.5 2.4 3.5 .4
Until I've updated my file, this only covers player-seasons thru 2006.
Rose was a bigger scorer and passer than prime Penny, etc.
Most of his numbers are about midway between Wade and Iverson. But more Ast and fewer Stl than either.
EDIT: Found the b-r.com blog post, "D-Rose and Iverson"
http://www.basketball-reference.com/blog/?p=9101
compare his campaign to that of Allen Iverson in 2001...Statistically, the two players are incredibly comparable. If you translate Iverson from the 103.0 league-ORtg environment of 2001 to the league ORtg of 107.1 in 2011...
Also found 100 player seasons thru 2006 that look to me more closely comparable to Rose of 2011. Including 3 by Billups, 5 by Kobe, 4 by KJ, 4 by McGrady, 5 by Payton, 6 by Oscar, and 11 by West. Also 7
other years by Iverson.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sat May 14, 2011 5:18 pm
by huevonkiller
Iverson also has impressive volume stats, that's not really the point. Most of those years are weak MVP seasons.
From Neil's analysis:
"Statistically, the two players are incredibly comparable. If you translate Iverson from the 103.0 league-ORtg environment of 2001 to the league ORtg of 107.1 in 2011, his ORtg/%Poss/DRtg becomes 110.5/33.8/103.0, production that is basically equivalent to Rose's after adjusting for usage."
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sat May 14, 2011 5:23 pm
by Mike G
Can you please edit your post to exclude the repetition of my entire post?
Including, if you'll notice, the double repetition of what Neil wrote?
Of course, I've adjusted for league and for team. Oh yeah, I already wrote that.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sat May 14, 2011 5:28 pm
by huevonkiller
Can you please edit your post to exclude the repetition of my entire post?
Including, if you'll notice, the double repetition of what Neil wrote?
My Neil quote is intentionally different.
Mike G wrote:
Of course, I've adjusted for league and for team. Oh yeah, I already wrote that.
Yeah I edited that but you didn't see it. I think you're putting too much of an emphasis on Big Name players (similar to Kobe?), and volume stats. Yeah he's similar to their unworthy MVP seasons, it is a simple analysis. The Bulls are still a great team.
Rose is not an efficient player, and neither were Kobe or Jerry West in their off-peaks.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sat May 14, 2011 11:04 pm
by Assani Fisher
greyberger wrote:At the risk of reviving arguments that were tedious months ago...
Rose is not a crazy choice for MVP. What's crazy is how there was precious little conversation to begin with. Howard was simply not an option, as Phil points out, because the Magic weren't as good this year and weren't a one or two seed. Lebron's disqualified several times over, for various reasons, and doesn't get a fair shake from the logic MVP voters employ anyway.
They always start with team wins versus expectations and proceed by making a little mental exercise where each player is responsible for so and so many wins. Only the best players are considered in this cowboy calculus - you're not allowed to consider coaches or any player who isn't known nationally as contributing wins. For example your mental MVP list for the Thunder should split 40 or so of their wins between Durant and Westbrook in some fashion, leaving 15 for the rest of the team. Do this for every team (remember, you're only allowed to name a player and give him specific credit if he's a established, marketable piece - two to four names per team, please!) and it's obvious that Rose is your choice.
After all, Boozer and Noah were injured and never played that well in the regular season! Boom, credit this kid with 30 wins, start the coronation in early March because this team is going to finish high and we need to jump out in front of it!
this post is spot on imo
greyberger wrote:Unless you're sure that Howard's supporting cast was more than 10 wins worse than Rose's this year...
How can you compare the supporting cast in Chicago and Orlando and not conclude that Chicago has the much better team?
The Magic's second best player this regular season was Jameer Nelson. Who's their third best player, and how does he compare to Deng or Noah?
Agree with your conclusion(that Rose's supporting cast is way better), but I think "compare 2nd best players, then compare 3rd best players" is a bad way to demonstrate it because Ryan Anderson, Brandon Bass, Jameer Nelson were all decent this year. Rose's cast is great because of how deep it is and how every single player has seemingly bought into playing great D. Guys like Omer Asik weren't on anyone's radar coming into the year, but they played very very well. Orlando, on the other hand, had to play a ton of negative players for major minutes.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sun May 15, 2011 12:25 pm
by Mike G
If you translate Iverson from the 103.0 league-ORtg environment of 2001 to the league ORtg of 107.1 in 2011, his ORtg/%Poss/DRtg becomes 110.5/33.8/103.0, production that is basically equivalent to Rose's after adjusting for usage."
I don't know if ORtg, %Poss, and DRtg are equivalent to 'production', but I do believe a league's averages are less significant than a team's averages, and especially their opponents' averages
vs that team. That is, after all, the league's averages in the games these players were in.
Playing for an offensively challenged Sixers team in '01, Iverson shot .518; Philly as a whole shot .518, and their opponents shot .499 (TS%).
Rose shot .550 this year, the rest of the Bulls shot .538, and their opponents shot .509 .
Rose shot .012 better than the rest of his team and .041 better than the opposition.
Iverson'01 shot no better than his teammates and .019 better than their opponents.
Iverson scored 28.4% of his team's points and 29.8% of what opponents scored.
Rose had 25.1% of Bulls' points and 27.1% of what opponents got.
Ast% : Rose 38.7 - 23.0 Iverson; Ast/36 Rose 7.4 - 3.9 Iverson
TReb% : Rose 6.4 - 5.2 Iverson
Stl+Blk% : Iverson 3.7 - 2.8 Rose
TO% : Iverson 10.0 - 13.1 Rose; TO/36 Iverson 2.9 - 3.3 Rose
Some of these are close, and some are not at all close. Rose "not an efficient player"? In what way?
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sun May 15, 2011 1:44 pm
by EvanZ
I think Iverson is a good case study. When he moved to Denver, his USG dropped about 10 points and he posted the best TS% numbers of his career. Hoopdata doesn't go back before 2007, so I can't look at all his Philly years, but if you compare his shots from Denver and Philly in 2007, you see that all of a sudden when he got to Denver he took 3 fewer 16-23 ft shots per game the rest of that season and in 2008. That alone seems to account for most of the improvement in his efficiency. It's pretty clear to me that the reason he didn't have to take as many of those "bad" shots is that Melo already served that purpose for the Nuggets. He didn't have that kind of help in Philly.
As I wrote about in my blog before, that 16-23 ft shot is what kills a player's efficiency (especially if they are unassisted). But somebody has to take them. When a player has teammates who can share that burden, his efficiency will *always* improve. That's just how the math works out.
Re: Rose Wins MVP And There's Apparently Little Doubt About
Posted: Sun May 15, 2011 4:14 pm
by Mike G
How many players, in at least 2000 minutes, have averaged at least 24 points and 7.3 assists, shooting .550 or better TS% ?
Rose, LeBron last year, Nate Archibald in '73, and Oscar in 6 of his first 8 seasons.