Let's Play A Game With On/Off Court Stats

Discussion in 'Butler Basketball' started by Jared Grubbs, Nov 6, 2018.

  1. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    I've been having discussions with multiple people today about On/Off Court efficiency stats from last season. I challenged one person to try and identify who specific players were based on these stats, and I thought it was very interesting to see the response so I want to extend this challenge to all of you.

    Below I have listed the on/off court team efficiency numbers for all 8 rotation players from last years team (Martin, Baldwin, Jorgensen, Wideman, McDermott, Fowler, Thompson, Baddley). The numbers listed are team points per possession with that player on the floor/team points per possession with that player on the bench and then the total net points per possession.

    So for example, Butler scored 1.16 points per possession when player A was in the game and only 1.06 ppp while he was on the bench. Defensively they allowed 1.03 points per possession while he was in and only 1.01 ppp while he was out. So the offense functioned much more efficiently with him in the game but the defense was slightly worse than it's average with him in.

    Your goal is to correctly identify all 8 players. I will provide some clues as the guessing progresses if (and probably when) no one gets it spot on.

    ***Disclaimer***: These stats are not adjusted for opponent, and the net rating is NOT meant to rank players. It is meant to be interesting only, not to be a single tool to judge players.

    Player A - Offense (1.16/1.06), Defense (1.03/1.01) - Net +0.08
    Player B - Offense (1.13/1.09), Defense (1.02/1.02) - Net +0.04
    Player C - Offense (1.12/1.07), Defense (1.02/1.01) - Net +0.04
    Player D - Offense (1.10/1.15), Defense (1.00/1.07) - Net +0.02
    Player E - Offense (1.07/1.15), Defense (1.00/1.05) - Net -0.03
    Player F - Offense (1.08/1.13), Defense (1.01/1.02) - Net -0.04
    Player G - Offense (1.12/1.09), Defense (1.05/0.98) - Net -0.04
    Player H - Offense (1.10/1.13), Defense (1.02/0.99) - Net -0.06
  2. ButlerNut

    ButlerNut Well-Known Member VIP Member

    I can only imagine the comments on this thread...the Pro Nate and the Anti Nate. The Pro PJ and the Anti PJ. This should be a fun and wild ride.
  3. the_speakers_lab

    the_speakers_lab Well-Known Member

        1. Fowler
        2. McD
        3. Kamar
        4. Kelan
        5. Wideman
        6. Thompson
        7. Baddley
        8. PJ
    FWIW, these are somewhat skewed by the amount of minutes ppl. My rankings are simply bc advanced analytics love Fowler and McD
  4. DarthHoltmann

    DarthHoltmann Member

    Welp, I might as well log back in for the new season. There appear to be 2 defensive studs. Worst offense has to be Thompson. That's E. D is maybe Baddley? G is the biggest defensive liability. McDermott? C and F are pretty much opposites of each other. It's got to be the 2 big mean. My gut says Wideman C and Fowler F but I've read you all talk about how the advanced numbers love Fowler so I'm going to swap them even though it doesn't make sense to me. Martin is the obvious choice for A, although I suspect it's someone else that makes this order more interesting. Surely Kamar isn't last, so I'll put him B and Jorgensen last.

    A: Martin
    B: Baldwin
    C: Fowler
    D: Baddley
    E: Thompson
    F: Wideman
    G: McDermott
    H: Jorgensen
  5. Insane Dawg

    Insane Dawg Well-Known Member

    I expect Hinkle to ace this test. I'm not weighing in.
  6. I Am Butler

    I Am Butler Well-Known Member

    A) Wideman
    B) Baldwin
    C) Martin
    D) Jorgenson
    E) Thompson
    F) Baddley
    G) McDermott
    H) Fowler
  7. I Am Butler

    I Am Butler Well-Known Member

    Only if he cheats . . . ;)
  8. Hinkle

    Hinkle Well-Known Member

    Ha. This is great and super difficult because I haven't seen Butler's on/off data from last year broken down like this. I'd love to put a bunch of thought into it and try to come closer by thinking more about who comes in when X player goes out, and who got easy minutes based on opponent, but I know I'm not going to have any time after noon today until tomorrow.

    Here's my shot. Kamar was the one I had the most trouble placing. I have no idea where to put him and I sort of ran out of slots:
    A. McDermott
    B. Martin
    C. Fowler
    D. Baddley
    E. Thompson
    F. Wideman
    G. Jorgenson
    H. Baldwin
  9. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    They are adjusted per possession, so the stats aren’t skewed by number of minutes. They are skewed by type of minutes though, which we can’t currently adjust for. I can adjust for Butler lineups but not opposing teams or lineups. What I’m actually more interested in the numbers themselves is seeing how fans perceive players and match them up. I personally would have only gotten 2, maybe 3 of these right, so I don’t expect anyone to actually match them all up or even get that close.

    Sent from my iPhone using Butler Hoops mobile app
  10. dawgs2014

    dawgs2014 Well-Known Member

    Plz no

    Sent from my iPhone using Butler Hoops
  11. WestLoopDog

    WestLoopDog Well-Known Member VIP Member

    I hope I got at least one right :). I suspect that Thompson may actually have had a positive OFF differential score, because he helped facilitate, even though I didn't line him up with that option.
    A= McDermott
    B= Martin
    C= Fowler
    D= Baldwin
    F= Wideman
    G= Jorgenson
    H= Baddley
  12. ButlerGoalie

    ButlerGoalie Well-Known Member

    A Baddley
    B Martin
    C Wideman
    D Baldwin
    E Thompson
    F Fowler
    G Jorgensen
    H McDermott
  13. the_speakers_lab

    the_speakers_lab Well-Known Member

    Yeah, they are skewed bc (especially) Kelan never leaves the court unless it's non-conference cakewalks and he's going to play the whole game during losses. +/- stats are always going to mostly affect the "star" players more negatively
  14. dawgs2014

    dawgs2014 Well-Known Member

    Want to prefect my statement by saying that I have no research to support my opinion, but I think there’s a strong case to be made that there is no correlation. There are very few players that play meaningful minutes that play only against all starters or only all bench players.

    Also, if you’re good, there’s a very good chance you’re good because of your stars which would enhance their scores.

    I just think over increased sample sizes the effect of what you’re factoring in is effectively neutralized.

    Sent from my iPhone using Butler Hoops
  15. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Kelan Martin did not score well on these stats. But I've looked into some of the commons reasons why, such as the ones you've provided here, which absolutely can be valid in some cases, and they don't really hold up. If I eliminate all of the opponents with a KenPom rating over 100 from the data completely his net efficiency actually doubles. And since it's a negative number that's not a good thing. In those games he played an average of 36.3 minutes per game during a win. In a loss he averaged 37.1 minutes, or 2% more minutes. If I narrow the sample size even smaller to only the 17 A games (top 50, adjusted for location), as identified by KenPom, the data gets even worse. Lineups with Kelan we outscored by 0.08 points per possession while lineups without him outscored opponents by 0.19 points per possession. Now, we are getting into small sample sizes there (there were only 90 possessions without Kelan in those 17 games, or the equivalent of a little more than a full games worth), but my point is that the lineups without Kelan weren't feasting on weak opponents and therefore boosting their stats. It's actually the opposite.

    Edit: incorrectly listed number of possessions originally. I had it as 180 possessions, but that was offense + defense possessions and that’s not how I represented it.
    Last edited: Nov 7, 2018
  16. Insane Dawg

    Insane Dawg Well-Known Member

    The bumb!
  17. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    You actually got 3 right, on 3 consecutive players, tied for the most with DarthHoltmann. However, if I scored it by how close you were on average for each player, I’m guessing you would have scored better if you flipped this upside down!

    Sent from my iPhone using Butler Hoops mobile app
  18. estar20dawg

    estar20dawg Well-Known Member

  19. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Very close. The correct answer is:

    A Fowler
    B Baddley
    C Jorgensen
    D Baldwin
    E Thompson
    F Wideman
    G McDermott
    H Martin
    ButlerNut and estar20dawg like this.
  20. Insane Dawg

    Insane Dawg Well-Known Member

    I knew they should have cut Martin.
    bumba, bmradio99, ButlerNut and 2 others like this.

Share This Page