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

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

  1. ButlerNut

    ButlerNut Well-Known Member VIP Member

    Money:
    $4,186
    There's lies, damn lies, and statistics. I will take D through H as the starting 5 over any combination you could put together otherwise.
     
  2. Insane Dawg

    Insane Dawg Well-Known Member

    Money:
    $12,653
    Apparently LJ didn't look at these stats and agreed with you.
     
    ButlerNut likes this.
  3. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
    I got 2 correct. I'm a Genius . . .

    Really though, what is supposed to be gained from this ranking? Not being snarky, really curious. It feels counterintuitive. I dont know a lot about these particular metrics, but they clearly dont point to the "best" players.
     
  4. dawgs2014

    dawgs2014 Well-Known Member

    Money:
    $2,589
    The point is that ascertaining how good a player is is not intuitive


    Sent from my iPhone using Butler Hoops
     
  5. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
    But that's not what these stats are doing. They show how good the team performs when the player is either in or out of the lineup, which is not how good the player is. There are a lot of factors going into those numbers that aren't necessarily skill based. For instance, how often did a rotation of Baldwin, PJ, Baddley, McD, Fowler see the floor together. Going by the numbers provided, that's the best rotation (using McD as the 4 over Martin). But the reality is that the rotation was rarely, possibly never, used.

    I'd like to know how coaches use this type of information. What is the practical application of it?
     
    shoeevv likes this.
  6. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Money:
    $15,957
    I can't tell you how the coaching staff uses it really. Based on the data I'm seeing I really don't think they do. For example, I'm not sure how you could look at that data on Kelan Martin and continue to play him near 40 minutes per game like they were at the end of the season. To me that data screams let's get Kelan some more rest. The backups are more than holding their own, there's no reason not to get him some more breaks.

    As for what I like to use it for, it's mainly to raise questions to research further. What did the team do differently that made them more effective offensively without Kelan Martin? Or why does the team rebound better with Nate Fowler on the floor?

    If we look at that first question, the answer appears to be related to 3 point shooting. Butler took more 3 point shots and hit them at a higher clip with Kelan Martin benched. They turned the ball over less but picked up more assists. I don't think it's a coincidence that Butler's most effective 2 man duo last season was the combo of their best 3 pointer shooters, Sean McDermott and Henry Baddley. Butler scored 1.23 points per possession when those 2 played on the floor together, and 1.09 ppp any other time. Despite neither of those guys being ball handlers the assist total goes up and turnovers go down with them on the floor.

    So why was Butler more effective last year with Kelan on the bench? IMO it was because they did a better job of spreading the floor with shooters and sharing the ball. I wish I could look at Kelan Martin or Kamar Baldwin's individual numbers when playing with McDermott and Baddley. Wouldn't it make sense that it's pretty hard to guard either of those guys if you have a 43% 3 point shooter on one side and a 46% 3 point shooter on the other? The game has become all about spreading the floor and shooting the 3 ball. So it's not that Kelan Martin isn't a great player, because he is. It's that Butler played a more effective style of basketball without him in the game. At least that's what I take out of it when I dive into this.
     
  7. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
    Thanks Jared. It raises the question: is the solution to sit your best player more often or to adjust how he is utilized while in the game? Perhaps, more likely, a combination of the two. Resting him an extra 3-4 minutes per game but also reduces his shot volume by encouraging others to shoot. I wonder what Kelans numbers look like if you could isolate his on/off stats to only account for times Thompson was out of the game. Seems having Thompson on the floor would force Martin to create his own shot more often, as there is one less potential shooter. It also pushes the idea of Baddley replacing Thompsons minutes as the season moves a long.

    Where are you pulling this info. Is it behind the KenPom pay wall?
     
  8. WestLoopDog

    WestLoopDog Well-Known Member VIP Member

    Money:
    $3,668
    I like Baddley's play and athleticism, but I don't expect Thompson to lose a bunch of minutes to him. I do expect a big leap and a big year from Aaron this season. It may take a couple of years for him to become an outside shooting threat, but I think this year we will see an increased ability to get to the rim and finish/collect fouls. IMO the key will be for him to significantly improve his FT%, which I predict he will do. His shot isn't mechanically broken, so getting to 65-70% is certainly doable. Being able to get 6-8 points/game by improving his ability to finish and convert FTs will keep him on the floor, because his defense and ball handling are so valuable.
     
  9. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Money:
    $15,957
    This info is from Hoop Lens, which is behind a paywall. I can isolate stats to reflect any combination of players you want. I can adjust for Butler's lineup, but not for opponent unfortunately. With Martin and Thompson on the court Butler scored 1.08 ppp and gave up 1.0 (sample size about 1200 possessions). With Martin on and Thompson off Butler scored 1.14 ppp and gave up 1.05 (sample size about 940 possessions). So overall it was pretty even, but obviously favors offense without Thompson and defense with him. I expect that trend would be about the same for any combination involving Thompson.

    Edit: When I say I can't adjust for opponent, I'm talking about opponent lineups. What I can do is only run the data for specific sets of games.
     
    Last edited: Nov 10, 2018
    estar20dawg likes this.
  10. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
    What were the numbers when both Thompson and Martin were out?

    Also, does it tell you how big those samples are for each rotation?
     
  11. Insane Dawg

    Insane Dawg Well-Known Member

    Money:
    $12,653
    I would venture to say the opponents had their best defender on Martin and their least on Thompson.
     
  12. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
    Or no defender on Thompson and two defenders on Martin.
     
    DawgsMD, FRENCH DAWG and Insane Dawg like this.
  13. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Money:
    $15,957
    Too small of a sample size with both Thompson and Martin off. There were only 141 possessions and 80 of them involved at least one of Brunk, Gillen-Butler, Campbell, or David (1.21 ppp scored). If you remove those 80 possessions you get 1.39 ppp on offense and 1.14 on defense. That’s basically just under a 1 game sample size though. Maybe enough data to say it could have been tried more often, but there’s no way 1.39 is sustainable.


    Sent from my iPhone using Butler Hoops mobile app
     
  14. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Money:
    $15,957
    To follow up on this lineup, it was essentially Baldwin, Jorgensen, Baddley, McDermott and one of the bigs. No surprise that this group basically chucked 3s. 45% of their shots came from 3 and they only turned it over 8% of the time (half the normal rate).


    Sent from my iPhone using Butler Hoops mobile app
     
  15. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
     
  16. I Am Butler

    I Am Butler Well-Known Member

    Money:
    $4,119
    So three games in, and I understand the simple size remains small, but what do the on/off numbers say about Brunk and Fowler. The eye test says we are much better so far with Brunk in the game. Do the numbers bear that out?
     
  17. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Money:
    $15,957
    Nope. Brunk has actually had a negative +/- in 2 of the 3 games. I’ve actually been surprised at that as it does feel like his energy is contagious on the floor. He’s 7th on the team when adjusted per possession while Fowler is once again first. The big difference between the two is on defense, where the team is giving up 1.06 points per possession on defense with Brunk in and 0.85 points per possession with Fowler in. It’s a small sample size as you noted, but that is a huge gap. I doubt that sized gap will hold up, but I do agree that Fowler is superior to Brunk on the defensive end. While I’m very happy with the progress he has made on that end, he still gets beat frequently in the screen game.


    Sent from my iPhone using Butler Hoops mobile app
     
    Hinkle likes this.
  18. Insane Dawg

    Insane Dawg Well-Known Member

    Money:
    $12,653
    What's the offensive ratings for Fowler over Brunk?
     
  19. Jared Grubbs

    Jared Grubbs Administrator Staff Member

    Money:
    $15,957
    IMG_1111.JPG



    Sent from my iPhone using Butler Hoops mobile app
     
  20. Insane Dawg

    Insane Dawg Well-Known Member

    Money:
    $12,653
    Thanks. You need to share this with LJ. I don't think he knows this by virtue of his substitutions
     

Share This Page