The Stat Lab

Discussion in 'Pistons and NBA' started by TaS, Jan 6, 2009.

  1. Mogilny

    Mogilny All-Star Forum Donor

    Joined:
    Jan 22, 2010
    Messages:
    3,008
    Likes Received:
    1,235
    I found this interesting. This is the offensive and defensive top 3 and bottom 3 2012 Advanced Statistical Plus/Minus Value over Replacement Player.

    Offense

    Top 3:

    8.21
    6.19
    5.78

    Bottom 3:

    -8.14
    -7.01
    -6.23

    Defense

    Top 3:

    -3.85
    -3.44
    -3.37

    Bottom 3:

    3.55
    3.22
    2.83

    The difference in impact between the best and the worst offensive player is a staggering 16.35 points while the difference between the best and the worst defensive player is 7.4. I have always believed that offensive superstars have a bigger impact than defensive but I didn't think the difference was this large. Looks like defense isn't half of the game, it's only a third!

    The stats:

    http://godismyjudgeok.com/DStats/aspm-and-vorp/2012-aspm/

    Advanced Statistical Plus/Minus explained (a good read for the nerdiest of us)

    http://godismyjudgeok.com/DStats/2011/nba-stats/a-review-of-adjusted-plusminus-and-stabilization/
  2. Nemo

    Nemo Fantasy-Football Snub Forum Donor

    Joined:
    Dec 1, 2006
    Messages:
    12,230
    Likes Received:
    1,475
    Gender:
    Male
    Location:
    Lake Elmo
    Flip Saunders knew that.........
  3. TaS

    TaS All-Star Forum Donor

    Joined:
    Sep 1, 2005
    Messages:
    17,566
    Likes Received:
    2,895
    Gender:
    Male
    Location:
    Chicaaago
    Defenders can be avoided if there are weak links on the team. Offensive players like Lebron and Kobe cannot be avoided because there is only one basketball and they have it.

    Maybe that has something to do with the lack of symmetry. I'm not sure that I can fully grasp that stat though. Would have to really think about it.
  4. Mogilny

    Mogilny All-Star Forum Donor

    Joined:
    Jan 22, 2010
    Messages:
    3,008
    Likes Received:
    1,235
    The stat is just regular +/- with a tweak. They have used regression to adjust for who the players play with and against to remove the context dependence of raw +/-. The VORP part doesn't matter since I'm looking at the difference between the worst and the best players.
  5. TaS

    TaS All-Star Forum Donor

    Joined:
    Sep 1, 2005
    Messages:
    17,566
    Likes Received:
    2,895
    Gender:
    Male
    Location:
    Chicaaago
    I get the adj +/-, but I'm trying to think of all the potential biases of the VORP difference. There are a lot of things imbedded in there and it is difficult to know exactly what it is telling you.

    For instance, replacement level- "replacement level estimates numerically the performance of players who can be acquired at little or no cost to the team. This generally means either players signed as free agents during the season or those who make teams as training-camp invitees."

    Maybe defensive abilities are not as highly recognized in the NBA and those replacement level guys are actually pretty good defenders and tend to be more offensively challenged? For instance, maybe there are a ton of Damien Wilkins, Darvin Ham, young Ben Wallace types of guys who flow into that replacement level player pool. I'm not saying this is the case, just that it's the kind of thing you'd want to figure out before believing in the VORP difference. I'm open to the possibility that the opposite is true and it's a bunch of Alex Acker, Will Bynum types who can fill it up, but at a huge cost to the defense. It's a complex stat.
  6. Mogilny

    Mogilny All-Star Forum Donor

    Joined:
    Jan 22, 2010
    Messages:
    3,008
    Likes Received:
    1,235
    I see your point but I don't think it matters when you look at a differential. If the VORP-value would be 3 points higher the list would look like this:

    Top 3:

    -0.85
    -0.44
    -0.37

    Bottom 3:

    6.55
    6.22
    6.83

    The difference is still the same.
  7. TaS

    TaS All-Star Forum Donor

    Joined:
    Sep 1, 2005
    Messages:
    17,566
    Likes Received:
    2,895
    Gender:
    Male
    Location:
    Chicaaago
    Those links were really interesting to read through, but there are a ton of variables going on. I just need to read through it all thoroughly before I'll feel like I understand what is imbedded.

    What is "offensive adjusted +/-" exactly? Is it essentially offensive rating on court minute offensive rating of the replacement player on court? If that is the case, then it doesn't seem like it is a +/-. Doesn't adj +/- take into account offense and defense combined while on and off the court?

    It seems like they are able to do it- when I filtered for Detroit only, it seemed to position our players pretty accurately on the spectrum, but I just don't understand how the calculation worked.
  8. Mogilny

    Mogilny All-Star Forum Donor

    Joined:
    Jan 22, 2010
    Messages:
    3,008
    Likes Received:
    1,235
    I read quite a lot of APBRmetrics last year and noticed the guy (DSMok/DanielM) writing the blog is one of the most respected guys in the "APBRmetrics scene" and his speciality is plus/minus. I really enjoy his stuff too since I'm more of a +/- kind of guy than a box score kind of guy. I think he got hired by some NBA team because of his writings.

    It was a while ago I read about his plus/minus so I don't remember in detail but his adjusted +/- takes into account offense and defense combined and then he just divides it by how much the offensive and defensive rating is affected when the player is on and off the floor.
  9. TaS

    TaS All-Star Forum Donor

    Joined:
    Sep 1, 2005
    Messages:
    17,566
    Likes Received:
    2,895
    Gender:
    Male
    Location:
    Chicaaago
    It shows pretty clearly how much more valuable Greg Monroe is than anyone else on the team as an overall player.
    MotownPride likes this.
  10. Walter

    Walter All-Star Forum Donor

    Joined:
    Jan 30, 2009
    Messages:
    4,773
    Likes Received:
    1,480
    Gender:
    Male
    Interesting stats:
    Pistons Advanced Stats Snapshot: Wins Produced Game Splits Visualized - Detroit Bad Boys

  11. TaS

    TaS All-Star Forum Donor

    Joined:
    Sep 1, 2005
    Messages:
    17,566
    Likes Received:
    2,895
    Gender:
    Male
    Location:
    Chicaaago
    Great article. I have no doubt that Knight is doing more harm than good out there this year, but I'm not overly concerned about that. His flashes of brilliance warrant giving him a full 2+ year development stint on the court. You don't want to give up on a PG or a C too soon, that is for sure.

    Also very cool how they show the trends for the per minute efficiency. It's hard to argue with win shares.

    A guy like Wilkins might get shafted a little if he is in more often on defense and is asked to guard better players, but that effect is probably very tiny.

    Ben Wallace has been effective for sure. It's just really evident that he defends with his brain when you watch him this year. No falling for pump fakes, great hand action, perfect help defense, good positioning, and still a rebounding force.
  12. Walter

    Walter All-Star Forum Donor

    Joined:
    Jan 30, 2009
    Messages:
    4,773
    Likes Received:
    1,480
    Gender:
    Male
    It just goes to show you how special of a player Big Ben really is. At his age and given his offensive limitations, he has no business being one of the top players in win shares, or showing improvement as the season wears on. But man, oh man, stats don't lie. I feel fortunate that I was there to watch his career.

    Well, I'm assuming he got plenty of practice during scrimmages with Tay over the years.
    brofmfa likes this.
  13. Laimbrane

    Laimbrane All-Star Forum Donor

    Joined:
    Jun 2, 2008
    Messages:
    1,190
    Likes Received:
    832
    My favorite quote from that article: "What the Wins Produced numbers won't tell you is that Damien leads the league in shot fakes per minute." I don't know if it's true (I'm notoriously bad at picking up on subtle sarcasm), but I love the idea of it being true.​

    And for all the bashing we do about the Pistons' frontcourt, those four guys (Moose, Viking, MaxSlam, and O.B.W.) are four of the top 5 Pistons in win shares, and the one that was sent down to the D-league (Mack-Baby) actually leads the team in WP48.

    What the Pistons appear to really need is a point guard. Here's hoping Knight actually turns into someone halfway decent.
  14. pass99

    pass99 All-Star Forum Donor

    Joined:
    Aug 6, 2006
    Messages:
    2,721
    Likes Received:
    388
  15. The Panda

    The Panda Garbage Post All Star Forum Donor

    Joined:
    Jan 29, 2012
    Messages:
    2,001
    Likes Received:
    1,088
    Gender:
    Male
    Can I just point out that Greg Munroe is sitting pretty at 14th in the league in PER with only Faried and Ginobili in the top 15 with less minutes.

    Unbelievable.
  16. Nemo

    Nemo Fantasy-Football Snub Forum Donor

    Joined:
    Dec 1, 2006
    Messages:
    12,230
    Likes Received:
    1,475
    Gender:
    Male
    Location:
    Lake Elmo
    How's Stuckey doing these days with PER ?
  17. The Panda

    The Panda Garbage Post All Star Forum Donor

    Joined:
    Jan 29, 2012
    Messages:
    2,001
    Likes Received:
    1,088
    Gender:
    Male
    18.54

    Above average
  18. Nemo

    Nemo Fantasy-Football Snub Forum Donor

    Joined:
    Dec 1, 2006
    Messages:
    12,230
    Likes Received:
    1,475
    Gender:
    Male
    Location:
    Lake Elmo
    thanks..........
  19. TaS

    TaS All-Star Forum Donor

    Joined:
    Sep 1, 2005
    Messages:
    17,566
    Likes Received:
    2,895
    Gender:
    Male
    Location:
    Chicaaago
    I've come to the conclusion that Stuckey was the best player on the team this year. I could also make the argument for Monroe, but I just think he was close 2nd.

    Data points:
    - We were almost breakeven with Stuckey on the floor this year (nobody else on the team, except for the non qualifying Macklin was anywhere close).
    - Without Stuckey on the floor, we felt a huge negative shock of -10 points per 100 possessions. That is about twice the shock as any other player removed.
    - Stuckey had by far the higher production differential of any player on the team and held his opponent to the lowest.
    - Stuckey's "simple rating" was 6.8 to Monroe's 2.3
    - Stuckey's PER was 2nd only to Monroe's
    - Stuckey's WS and WS/min was 2nd to Monroe
    - His points per minute were #1
    - His assist/ tov ratio was #1 (2nd in assist rate behind Walker Russell)
    - Most free throws per game by a wide margin (11 games with double digit attempts)
    - Shot .337 from deep... his highest percentage of his career by a mile
    - #2 in total shooting percentage (a hair behind Monroe)
    - Had the highest offensive rating on the team
    - 4 games of 30+ points, including against Chicago and the Lakers
    - 26 games with 1 turnover or fewer
    - 50 out of 52 games with < 4 turnovers. The other 2 games were 5. Very secure with the ball... considering a lot of these per probably charges.


    On a side note, both PER, Winshares, and total shooting % are all pointing to Jerebko as the 3rd best player on the team. There is likely a bias since he came off the bench mostly and was facing easier competition, but something to think about a little more.
    brofmfa likes this.
  20. Mogilny

    Mogilny All-Star Forum Donor

    Joined:
    Jan 22, 2010
    Messages:
    3,008
    Likes Received:
    1,235
    Stuckey isn't a great defender but Monroes Villanueva like defense settles the case.

Share This Page