Announcement

Collapse
No announcement yet.

Everything Bargnani

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

  • ebrian wrote: View Post
    It's a good point actually, since Bargnani is really known for his offense... or at least, he's supposed to be. I've moved up the offensive categories and defensive have been pushed down. I've also bolded the ones where Dirk is significantly better than Bargnani.

    Dirk Nowitzki (Andrea in brackets):

    Per game --
    PER: 7th (32nd)
    Offensive Rating: 2nd (29th)
    Offensive Win Shares: 2nd (33rd)
    Field Goal Percentage: 29th (34th)
    Assists: 11th (34th)
    Total rebounds: 26th (35th)
    Blocks: 22nd (30th)
    Steals: 13th (33rd)
    Total Rebound %: 33rd (35th)
    Defensive Rating: 23rd (35th)
    Defensive Win Shares: 13th (32nd)

    Usage Percentage: 5th (13th)
    One of the things that really stands out for me is the assists. Yes, he's bad at rebounding, but it's just as damning that he's this bad at assisting as well. We knew that of course, but this is one more of those things he showed promise in in his first years (in my opinion) that didn't develop at all.

    Comment


    • Soft Euro wrote: View Post
      I don't think there are many people, if any, who (still) think that. I don't think anybody argued that here either.
      All the more reason why arguing against the methodology was such a waste of time/effort, if we all agree on the conclusion.

      "2 plus 2 equals 4"

      "You're crazy! 3 plus 1 equals 4!"

      Comment


      • CalgaryRapsFan wrote: View Post
        As I've read through this thread, I think I've come up with a way to explain the issues that some people have had with various arguments made by various posters on both sides of the debate - for the record, I thought Jimmie did a good job trying to explain it yesterday.

        I also think I should start by saying that my assumption is based on the fact that so much blame & "hate" has been dumped on Bargnani over the years - some well deserved and some well beyond his control (ie: draft position, salary) - that the "piling on" can get frustrating to take, even when the bulk of the pile is quite legit.

        Basically, I think the personal stats (like those in ebrian's recent posts) and unbiased observations of game action are accepted as factual evidenciary proof of Bargnani's "suckage". By this, I mean stats like PER, Offensive Rating, Offensive Win Share, Defensive Win Share, as examples, and observations such as poor help defense. Those stats are based purely on the indivdual and all pass the eye-test, without bias.

        However, I think people take issue with circumstantial evidence, such as the team's record in games when he scores a certain # of points, for example. These sorts of coincidental stats show no direct correlation between Bargnani and the outcome. These sorts of 'stats' can be found to support any argument about any player, good or bad. For example, there might be a statline that shows the team is 8-0 all-time when Bargnani scores 7 points and has 3 rebounds, but that doesn't mean the team should adopt a strategy to ensure Bargnani hits exactly that statline and then gets benched, since it's purely coincidental.

        Even the most vocal Bargnani supporter, past or present, can accept unrefutable evidence based on stats that observation can validate. However, after all the years of "Bargnani bashing", it can get frustrating to have that compounded by purely circumstantial evidence, be it hand-picked coincidental statlines or factors beyond Bargnani's control.

        Hopefully that makes sense and helps shed some light on the ongoing discussion...
        Well said.

        I agree almost completely except for the use of so called advanced stats like PER and WS. I don't like those at all, and I personally don't consider these advanced stats at all, but reductionist stats. Real advanced stats are adding information instead of reducing information to a single number which is derived in a debatable fashion. Most of those don't add anything to boxscore stats. Off those kind of stats I'm more receptive to things like Offensive and Defensive Rating over the course of more than one season as they signal trends that boxscore stats don't necessarily give.

        Comment


        • CalgaryRapsFan wrote: View Post
          As I've read through this thread, I think I've come up with a way to explain the issues that some people have had with various arguments made by various posters on both sides of the debate - for the record, I thought Jimmie did a good job trying to explain it yesterday.

          I also think I should start by saying that my assumption is based on the fact that so much blame & "hate" has been dumped on Bargnani over the years - some well deserved and some well beyond his control (ie: draft position, salary) - that the "piling on" can get frustrating to take, even when the bulk of the pile is quite legit.

          Basically, I think the personal stats (like those in ebrian's recent posts) and unbiased observations of game action are accepted as factual evidenciary proof of Bargnani's "suckage". By this, I mean stats like PER, Offensive Rating, Offensive Win Share, Defensive Win Share, as examples, and observations such as poor help defense. Those stats are based purely on the indivdual and all pass the eye-test, without bias.

          However, I think people take issue with circumstantial evidence, such as the team's record in games when he scores a certain # of points, for example. These sorts of coincidental stats show no direct correlation between Bargnani and the outcome. These sorts of 'stats' can be found to support any argument about any player, good or bad. For example, there might be a statline that shows the team is 8-0 all-time when Bargnani scores 7 points and has 3 rebounds, but that doesn't mean the team should adopt a strategy to ensure Bargnani hits exactly that statline and then gets benched, since it's purely coincidental; no game is ever decided by a single player's statline.

          Even the most vocal Bargnani supporter, past or present, can accept irrefutable evidence based on stats that observation can validate. However, after all the years of "Bargnani bashing", it can get frustrating to have that compounded by purely circumstantial evidence, be it hand-picked coincidental statlines or factors beyond Bargnani's control.

          Hopefully that makes sense and helps shed some light on the ongoing discussion...
          I disagree with the bolded. I think the stats do indeed show a correlation....they just don't necessarily establish causation.

          The correlation being that it's harder to win games with Bargnani a key contributor (which is different than saying Bargnani is the sole reason why we lose).

          Comment


          • Nilanka wrote: View Post
            All the more reason why arguing against the methodology was such a waste of time/effort, if we all agree on the conclusion.

            "2 plus 2 equals 4"

            "You're crazy! 3 plus 1 equals 4!"
            We keep disagreeing about this. You give two ways of coming to a conclusiong that are both right. Let me try (with emphasis on 'try') to give a similar analogy. Suppose you have a table with 4 pies. Now someone says "see, there are 3 on the left and 2 on the right side of the table; together that makes 4 pies." I agree that there are 4 pies, so I should not say anything about the way this guy came to the same conclusion as me?

            Comment


            • Nilanka wrote: View Post
              I disagree with the bolded. I think the stats do indeed show a correlation....they just don't necessarily establish causation.

              The correlation being that it's harder to win games with Bargnani a key contributor (which is different than saying Bargnani is the sole reason why we lose).
              You're absolutely right. I realized that I meant they only show a correlation (same thing as being coincidental), without establishing a causality - hence, circumstantial evidence.

              There's clearly enough hard evidence, so there's no reason to even have to resort to the more circumstantial evidence at this point. And that's coming from one of the most vocal supporters of Bargnani, up until this season.

              Comment


              • Nilanka wrote: View Post
                I disagree with the bolded. I think the stats do indeed show a correlation....they just don't necessarily establish causation.

                The correlation being that it's harder to win games with Bargnani a key contributor (which is different than saying Bargnani is the sole reason why we lose).
                Except of course, that when looked at these stats over the complete career of Bargnani the team is actually much better when he scores 20 points or more, which makes this much smaller sample size extra suspect.

                An added critique of these kind of stats is that they're very crude, e.g. it would be much better to look at his effeciency in relation to the amount of shots he takes.

                Comment


                • Soft Euro wrote: View Post
                  We keep disagreeing about this. You give two ways of coming to a conclusiong that are both right. Let me try (with emphasis on 'try') to give a similar analogy. Suppose you have a table with 4 pies. Now someone says "see, there are 3 on the left and 2 on the right side of the table; together that makes 4 pies." I agree that there are 4 pies, so I should not say anything about the way this guy came to the same conclusion as me?
                  I understand the point you're trying to make. Our difference in opinion stems from whether or not we accept Matt's evidence to be truthful or not.

                  I believe that they are indicative of Bargnani's value to this team, and that they shouldn't be dismissed based on the somewhat poor counter examples provided in this thread.

                  It's not universally accepted that isolated stats don't provide value. Therefore Matt's numbers can't be considered "wrong".

                  Comment


                  • Soft Euro wrote: View Post
                    Except of course, that when looked at these stats over the complete career of Bargnani the team is actually much better when he scores 20 points or more, which makes this much smaller sample size extra suspect.

                    An added critique of these kind of stats is that they're very crude, e.g. it would be much better to look at his effeciency in relation to the amount of shots he takes.
                    Looking at Bargnani's whole career makes sense (statistically speaking), but doing so doesn't take into account that a player can change (in this case, for the worse) since early in his career. At what point does a larger sample size becomes a less accurate indicator of Bargnani's value today?

                    Comment


                    • Nilanka wrote: View Post
                      I understand the point you're trying to make. Our difference in opinion stems from whether or not we accept Matt's evidence to be truthful or not.

                      I believe that they are indicative of Bargnani's value to this team, and that they shouldn't be dismissed based on the somewhat poor counter examples provided in this thread.

                      It's not universally accepted that isolated stats don't provide value. Therefore Matt's numbers can't be considered "wrong".
                      I think the point is that there's enough statsitical and observable reasons to want Bargnani traded, that there's no reason to even resort to the cherry-picked stat such as the team's record in games when Bargnani scores a certain number of points.

                      Basing an opinion on circumstantial arguments only weakens your point and invites criticism. I think that's what p00ka had issue with, since you can find similar statlines to "prove" the same argument for players like Jordan and Kobe, that are being used to support the stance of #tradeBargnani, but anybody would be a complete buffoon to jump on a #tradeJordan or #tradeKobe bandwagon. That's why such arguments stink much more of "hating", even when they're being presented as statistical, with no intended "hating".

                      Comment


                      • Nilanka wrote: View Post
                        I understand the point you're trying to make. Our difference in opinion stems from whether or not we accept Matt's evidence to be truthful or not.

                        I believe that they are indicative of Bargnani's value to this team, and that they shouldn't be dismissed based on the somewhat poor counter examples provided in this thread.

                        It's not universally accepted that isolated stats don't provide value. Therefore Matt's numbers can't be considered "wrong".
                        Matt's stat regarding the Raps record when he scores 20+ points is even more telling considering Bargnani's #1 attribute to this team, has always been scoring. If the team can't win when he most does the thing he is best at, then that is obviously a problem. Rudy Gay's #1 attribute isn't steals, so the Raps record when he records 4+ steals is completely tripe (as intended). Accpeting that certain players have key areas of responsibility and the correlating affect it has on the team, is totally a fair justification of using stats to state the obvious "Bargnani is terrible"
                        Heir, Prince of Cambridge

                        If you see KeonClark in the wasteland, please share your food and water with him.

                        Comment


                        • Nilanka wrote: View Post
                          I disagree with the bolded. I think the stats do indeed show a correlation....they just don't necessarily establish causation.

                          The correlation being that it's harder to win games with Bargnani a key contributor (which is different than saying Bargnani is the sole reason why we lose).
                          While you are correct in your first sentence, you are incorrect in your second. Mathematically speaking, again.

                          The "correlation" doesn't proved that it's harder to win games with AB. It shows there is a coincidence. That is, when Bargs scores more than 20 in the sample size used, the Raptors have shown a tendency to lose more.

                          There is absolutely no statistical/mathematical way to link those two statements. I understand that may be difficult to grasp, when we all see Bargnani not getting rebounds and playing poor help defense on a regular basis, but it's true.

                          At the same time, there are no stats that can prove that there ISN'T a causal relationship between Bargs scoring 20 and the Raptors losing. It could be true, but it can't be proven with the data we currently have on hand. It may never be able to be proven. And that's why some people object to it being used as substantiated proof that 'Bargs scoring 20 = Raptors losing more'.

                          Disclaimer (for the nth time): I am NOT, by posting this, trying to defend AB. I'm just saying you should stick to beating him up over things you can say are actually, substantially true/factual. Actually, I'm not saying "you need" to do anything. You can argue his suckage based on whatever spurious criteria you like; but by the same token, don't expect everyone to buy it wholesale or agree with it all without poking some holes in the argument.
                          Definition of Statistics: The science of producing unreliable facts from reliable figures.

                          Comment


                          • Axel wrote: View Post
                            Matt's stat regarding the Raps record when he scores 20+ points is even more telling considering Bargnani's #1 attribute to this team, has always been scoring. If the team can't win when he most does the thing he is best at, then that is obviously a problem. Rudy Gay's #1 attribute isn't steals, so the Raps record when he records 4+ steals is completely tripe (as intended). Accpeting that certain players have key areas of responsibility and the correlating affect it has on the team, is totally a fair justification of using stats to state the obvious "Bargnani is terrible"
                            That's exactly the type of argument I disagree with. The telling/damning fact is not the number of points scored, but more likely it's the inefficiency with which those points were scored in most of those games, that has any statistical merit. On the surface, who's to say that those games he scored 20+ points were games in which the entire team shot poorly, so Bargnani had to be relied on more to score?

                            See, that's the problem with drawing conclusions from circumstantial evidence - you can't prove that any one player is the root cause of the outcome in a team game. Present statistics of a highly ineffcient scorer who contributes little else on either end of the court - regardless of the outcomes of games - and I would be much more receptive.
                            Last edited by CalgaryRapsFan; Fri Mar 8, 2013, 02:40 PM.

                            Comment


                            • jimmie wrote: View Post
                              The "correlation" doesn't proved that it's harder to win games with AB. It shows there is a coincidence. That is, when Bargs scores more than 20 in the sample size used, the Raptors have shown a tendency to lose more.
                              I think this sums up the gist of the debate pretty well. I don't think we can simply write it off as coincidence either. The "truth" likely lies somewhere between 'coincidence' and 'direct causation', where like you said, will probably never be proven one way or the other.

                              Comment


                              • Nilanka wrote: View Post
                                I think this sums up the gist of the debate pretty well. I don't think we can simply write it off as coincidence either. The "truth" likely lies somewhere between 'coincidence' and 'direct causation', where like you said, will probably never be proven one way or the other.
                                True, and maybe will start to let this thread wind down...? ;-)

                                The larger point is, we can find stats to 'prove' anything, as long as we don't care too much about being mathematically or factually correct in our 'proof'...
                                Definition of Statistics: The science of producing unreliable facts from reliable figures.

                                Comment

                                Working...
                                X