Announcement

Collapse
No announcement yet.

Raptors analytics... and an internal rift? [post #67]

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #46
    The one flaw with all this stuff is that you guys are talking about how we need to do this and that based on the ghost players.

    That would be awesome if that's all it took. The reality is, good teams will adjust to your defense. If DeRozan follows his ghost to a tee, then the player he is guarding won't do the same he did in the video. Suddenly, the simulation changes again and the ghost is doing something else.
    your pal,
    ebrian

    Comment


    • #47
      Craiger wrote: View Post
      This is one huge problem I've had with all this recent discussion surrounding Raptors and analytics. Last offseason Rucker was tooting the Raptors analytical horn about Andrea, Fields, Lowry and the team in general.

      So what went wrong? Did he make mistakes in analyzing the data, did the analytical message get lost along the way, did someone (Casey or Colangelo) ignore the analytical advice?

      Combination of all 3?
      Reading a little more I find parts that may answer my questions a bit.... or maybe make me ask more questions

      "When you ask coaches what's better between a 28 percent 3-point shot and a 42 percent midrange shot, they'll say the 42 percent shot," Rucker says. "And that's objectively false. It's wrong. If LeBron James just jacked a 3 on every single possession, that'd be an exceptionally good offense. That's a conversation we've had with our coaching staff, and let's just say they don't support that approach."
      Yes a 28% 3pt shot is greater than a 42% mid range shot. BUT its also much worse than a 60% shot at the bucket (which an 'average player' shoots).

      Why are we comparing a bad shot to a worse shot, when in theory we want to aim for the best shot possible?

      If Lebron James jacked a 3pt shot on ever possession that would be exceptionally good offense (58.3% efg) but its also not as good as Lebron shooting between 3-9ft (59.8) and neither are even remotely close to Lebron at the rim (79.2!!)

      Now I actually think discussions like this area good thing because if the Raps aren't going to take shots close to the rim they should take 3pters instead (in general) as they are on average the 2nd most efficient shot. But thats only if guys can hit those shots.

      But the analytics team argues that even sub–35 percent 3-point shooters should jack more 3s
      The Raps three biggest users of offense are Gay, Demar and Andrea.

      Gay eFG% from 3 - 37%
      Demar eFG% from 3 - 37%
      Bargs eFG% from 3 - 46.6%

      Bargnani taking a 3 is the equivalent of an 'average' player shooting but still below the average 3 pt shooter, yet still a better choice than almost all other shots (except at the rim) for him. Demar shoots EVERY OTHER SHOT more efficiently than he does a 3, and for Gay it should be his 2nd last option (before taking 16-23fters).

      So, a basic look at analytics tells us the quote above is patently untrue for the Raptors biggest users of offense, with the exception of Andrea.

      But that asks a bigger question - why is the team using players who all shoot at a rate below average so much? Even if your top scorers are shooting their best or better shots, they are still, in general, worse than finding someone else to take them instead.

      So is this a problem of:

      1) analytics making a mistake - a 3pter is not always the better shot, and definetely not for all players
      2) players not listening - you can ask a player to do X and Y but it won't matter if they don't do it
      3) Casey avoiding the analytics - his statement on 3 pt shooting (it doesn't count if you miss) is rather ridiculous. That applies to every shot.
      4) Colangelo giving the team the wrong players or demanding the wrong players play - those 3 also happen to be the 3 highest payed Raptors (Demar next year), and all called or referred to as 'franchise player' at one point or another, despite the above information hardly being new.

      (the last one is the scariest, as if you have bad players or a bad system you are now just trying to maximize 'bad'.)
      Last edited by Craiger; Wed Mar 20, 2013, 08:57 AM.

      Comment


      • #48
        ezz_bee wrote: View Post
        +1
        mind=boggled

        The article depresses me more than anything. You can't pin the resistance to analytics solely at Casey's feet, since Bryan hired him in the first place, but either way, I don't think the organisation, whether its the front office or the coaching staff are getting as many points/wins as they could from analytics.
        Like the article mentions, these analytics are essentially uncharted waters and teams have to decide for themselves how to use them. Just because you have the system in place doesn't guarantee that you're getting beneficial information.

        I'll just throw this out there: we know how BC goes about putting together a basketball team and coaching staff, and it could be that he's equally inept at getting the right people to run the analytics - it could be that the people running it are just not good at what they do.

        It's either that, or the coaching staff doesn't listen. When a good part of your offence is geared around Demar shooting long twos, clearly something isn't working.

        ebrian wrote: View Post
        The one flaw with all this stuff is that you guys are talking about how we need to do this and that based on the ghost players.

        That would be awesome if that's all it took. The reality is, good teams will adjust to your defense. If DeRozan follows his ghost to a tee, then the player he is guarding won't do the same he did in the video. Suddenly, the simulation changes again and the ghost is doing something else.
        The ghost players are just the defensive principles of the team personified (or, I guess, ghostified). Teams are going to adjust, but if you keep applying your principles throughout the whole possession, you'll force a harder shot. If your principles are right.

        Craiger wrote: View Post
        Reading a little more I find parts that may answer my questions a bit.... or maybe make me ask more questions



        Yes a 28% 3pt shot is greater than a 42% mid range shot. BUT its also much worse than a 60% shot at the bucket (which an 'average player' shoots).
        But these shots are just hard to get.
        Last edited by JimiCliff; Wed Mar 20, 2013, 09:37 AM.
        "Stop eating your sushi."
        "I do actually have a pair of Uggs."
        "I've had three cups of green tea tonight. I'm wired. I'm absolutely wired."
        - Jack Armstrong

        Comment


        • #49
          JimiCliff wrote: View Post
          Like the article mentions, these analytics are essentially uncharted waters and teams have to decide for themselves how to use them. Just because you have the system in place doesn't guarantee that you're getting beneficial information.

          I'll just throw this out there: we know how BC goes about putting together a basketball team and coaching staff, and it could be that he's equally inept at getting the right people to run the analytics - it could be that the people running it are just not good at what they do.

          It's either that, or the coaching staff doesn't listen. When a good part of your offence is geared around Demar shooting long twos, clearly something isn't working.



          The ghost players are just the defensive principles of the team personified (or, I guess, ghostified). Teams are going to adjust, but if you keep applying your principles throughout the whole possession, you'll force a harder shot. If your principles are right.



          But these shots are just hard to get.
          But thats accounted for, to a degree anyways, in the %.

          If those shots were 'easy' to get, players would make nearly 100% of them. But since they are difficult to get (ie. the defense tries to prevent them) we see an enourmous drop in the %. The same applies to all shots. ie. their %s are lower than they could be because teams try and prevent them from happening.

          A more difficult but higher % shot can still be worth more than a less difficult low % shot.

          Thats not to say that all the Raps should be doing is shooting at the rim. But that should be the goal for most possessions, and if the team is unable to do it at a consistently reliable rate, they need to change how they are trying to achieve that goal - either through changing the plan (ie. plays) or the players (and/or their roles). Over a long enough period of time (sample size) we can get a very reliable idea of who is and who is not doing it, who is able to and unable to do it and who is doing it effectively.
          Last edited by Craiger; Wed Mar 20, 2013, 10:03 AM.

          Comment


          • #50
            Craiger wrote: View Post
            But thats accounted for, to a degree anyways, in the %.

            If those shots were 'easy' to get, players would make nearly 100% of them. But since they are difficult to get (ie. the defense tries to prevent them) we see an enourmous drop in the %. The same applies to all shots. ie. their %s are lower than they could be because teams try and prevent them from happening.

            A more difficult but higher % shot can still be worth more than a less difficult low % shot.

            Thats not to say that all the Raps should be doing is shooting at the rim. But that should be the goal for most possessions, and if the team is unable to do it at a consistently reliable rate, they need to change how they are trying to achieve that goal - either through changing the plan (ie. plays) or the players (and/or their roles). Over a long enough period of time (sample size) we can get a very reliable idea of who is and who is not doing it, who is able to and unable to do it and who is doing it effectively.
            My point was that it isn't a matter of should; they just can't. No team can. If you could take every shot at the rim, you'd do it in a heartbeat. But it's impossible, because basketball defence is, at it's fundamentals, about taking that shot away.

            That's why you need to be figuring out, "ok, so when we're not shooting at the rim, where should we be shooting?" And that's where understanding the expected value of different shots and different spaces on the floor becomes a big deal.
            Last edited by JimiCliff; Wed Mar 20, 2013, 10:45 AM.
            "Stop eating your sushi."
            "I do actually have a pair of Uggs."
            "I've had three cups of green tea tonight. I'm wired. I'm absolutely wired."
            - Jack Armstrong

            Comment


            • #51
              JimiCliff wrote: View Post
              My point was that it isn't a matter of should; they just can't. No team can. That's why you need to be figuring out, "ok, so when we're not shooting at the rim, where should we be shooting?"
              Sigh. I'm not arguing against that.

              But regardless how often can't they vs how often do they choose not to? And if they can't or won't often enough why not try to use or find someone who can more often rather than settling for a poorer choice?

              This is the problem I have, as I noted in my last line in the post (ie. the scariest part) is just 'using what you are using because thats what you have' is problematic. From game to game you ofcourse have no choice but to use the players you have (although you can change how you use them). But a team can and is able to change players over the course of the season, and make decisions leading to the player they do and do not have.

              I'm not arguing against figuring out what other shots to take, what are the 2nd and 3rd or 4th best shots to take etc. Far from it. My point was, as we see through the numbers I posted, that not only are certain shots not the same for all players, sometimes those players shouldn't be the ones taking the shots (or as many of them).

              basically whats the plan while using analytics?

              1) trying to build the best team possible?
              2) trying to build the best team possible given the players we have/want to use?

              Those two ideas are NOT the same

              The question should be - how often do players do it?, how effective are they at doing it? and how often + effectively do they do it vs others? When you don't ask that last question, you get an incomplete answer.

              Demar should rarely ever be taking 3s
              Rudy should only be taking 3s instead of long 2s
              Andrea should be taking 3s instead of most shots

              And all of them should be taking less shots.

              Comment


              • #52
                ceez wrote: View Post
                Probably the most interesting article I've read in a long time. I also love how it passive aggressively shits all over our players and coaching.

                If the team is so pro analytics, why do we have so many inefficient players? This seriously boggles my mind.
                Because BC looks at it a different way?
                Twitter: @ReubenJRD • NBA, Raptors writer for Daily Hive Vancouver, Toronto.

                Comment


                • #53
                  This basically just says our front office, coaching and analytic team have different perspectives on what constitutes a winning team ... and this worries me a lot.

                  The on court performance is just a product of our dysfunctional management team.

                  I understand the value of shooting 3 pointers. BUT, in order to generate clean looks (and thus, increasing the expected value of a shot), you need to have PENETRATION.

                  You can do this through the dribble, the pass or post; the goal should be get through the first line of defense. You can't generate clean looks by having 4 people stand outside the lane and having 1 big man run around setting screens (only to ignore him anyway on the roll). Basically, what I'm saying is, GO TO THE PAINT FIRST (pass, dribble or post), collapse the defense, then kick it out for a shot. Shot in the paint & a 3 point shot have the highest expected value per possession. It's no coincidence that the Spurs are applying this principle on the court -- and executing it like a well oiled machine.

                  For a team that claims to have an advanced analytic team, we have an OLD SCHOOL coach and a roster that goes against their claims. What you get is differing opinions about how to do sh!t which ultimately leads to inconsistent results.

                  So thanks for the article. I feel worse about our team.
                  Last edited by torch19; Wed Mar 20, 2013, 02:30 PM.
                  “I don’t create controversies. They’re there long before I open my mouth. I just bring them to your attention.”

                  -- Charles Barkley

                  Comment


                  • #54
                    torch19 wrote: View Post
                    This basically just says our front office, coaching and analytic team have different perspectives on what constitutes a winning team ... and this worries me a lot.

                    The on court performance is just a product of our dysfunctional management team.

                    I understand the value of shooting 3 pointers. BUT, in order to generate clean looks (and thus, increasing the expected value of a shot), you need to have PENETRATION.

                    You can do this through the dribble, the pass or post; the goal should be get through the first line of defense. You can't generate clean looks by having 4 people stand outside the lane and having 1 big man run around setting screens (only to ignore him anyway on the roll). Basically, what I'm saying is, GO TO THE PAINT FIRST (pass, dribble or post), collapse the defense, then kick it out for a shot. Shot in the paint & a 3 point shot have the highest expected value per possession. It's no coincidence that the Spurs are applying this principle on the court -- and executing it like a well oiled machine.

                    For a team that claims to have an advanced analytic team, we have an OLD SCHOOL coach and a roster that goes against their claims. What you get is differing opinions about how to do sh!t which ultimately leads to inconsistent results.

                    So thanks for the article. I feel worse about our team.
                    Just like when BC signed Kapono, and then Mitchell decided to have him shoot twos.
                    "Stop eating your sushi."
                    "I do actually have a pair of Uggs."
                    "I've had three cups of green tea tonight. I'm wired. I'm absolutely wired."
                    - Jack Armstrong

                    Comment


                    • #55
                      JimiCliff wrote: View Post
                      Just like when BC signed Kapono, and then Mitchell decided to have him shoot twos.
                      "We need to take mo shats!"

                      Comment


                      • #56
                        Nilanka wrote: View Post
                        "We need to take mo shats!"
                        What a mastermind...
                        "Stop eating your sushi."
                        "I do actually have a pair of Uggs."
                        "I've had three cups of green tea tonight. I'm wired. I'm absolutely wired."
                        - Jack Armstrong

                        Comment


                        • #57
                          Craiger wrote: View Post
                          Actually it does.

                          You are inferrring that their conclusion about Dwight Howard applies to Andrea, which it does not. Again, that was discussed and explained by the authors at sloan. Your thinking is players did not get within 5ft of Andrea under the bucket - they on the other hand concluded Andrea does gett to the bucket to get within 5 ft.

                          That is a SIGNFICANT difference.
                          Not to belabor this whole debate we're having, but could you quote what the Authors said exactly, as I couldn't find any reference to Bargnani that you allude to.

                          As to my conclusion that the Dwight Effect (of staying further away) applies to Andrea, I would say yes. That being said, I also believe this "Effect" doesn't apply in the same capacity that it does with Howard himself.

                          For the most part, I think we'll have to agree to disagree.
                          Works for me.

                          .

                          Comment


                          • #58
                            JimiCliff wrote: View Post
                            What a mastermind...
                            What's hilarious is that Smitch has more Coach of the Year awards than Jerry Sloan and George Karl :|

                            Comment


                            • #59
                              ebrian wrote: View Post
                              The one flaw with all this stuff is that you guys are talking about how we need to do this and that based on the ghost players.

                              That would be awesome if that's all it took. The reality is, good teams will adjust to your defense. If DeRozan follows his ghost to a tee, then the player he is guarding won't do the same he did in the video. Suddenly, the simulation changes again and the ghost is doing something else.
                              It's not really a "flaw". Think of it like chess. Whatever move I make, a good player is going to respond accordingly and make a good move in response. However, if I keep making good move, after good move, it gets harder and harder for him to respond. Eventually, if I make enough good moves, he may make a mistake or, even if he doesn't, his position is at best even with me or likely worse as the game goes on.

                              Defense is the same way. You are never going to stop the other team every possession and you can't count on your plan always working. But, if you do the right things over and over again you will make it extremely difficult for the opposition to score. Eventually, over the length of quarter or a game, that will be to your advantage.

                              Comment


                              • #60
                                Nilanka wrote: View Post
                                What's hilarious is that Smitch has more Coach of the Year awards than Jerry Sloan and George Karl :|
                                doesn't that mean he's actually an ok coach. Just had sub par players to work with...and boshtrich

                                Comment

                                Working...
                                X