This topic contains 12 replies, has 5 voices, and was last updated by AvatarAvatar thparadox 10 years, 1 month ago.

  • Author
    Posts
  • #53762
    AvatarAvatar
    Bajeebz
    Participant

    EPV, or Expected Possession Value, is the latest statistical measuring stick to materialize from the advanced metrics movement. According to Kurt Goldsberry from Grantland.com, “EPV seeks to place a value on every single movement of every single player every second they are on the court.”

    I’m not sure exactly how it works. I do know that EPV ranks Chris Paul is in first place with a score of (+3.48).

    What’s most concerning to me as a Wolves fan is this: Ricky Rubio (-3.33) and Kevin Love (-2.38) were ranked as the two worst players on the EPV scale. These statistics came from last year, when both were injured for much of the year, but it makes me wonder if EPV can ex plain the Wolves’ below .500 record considering they’re +4.1 point differential this season.

    Russell Westbrook and Paul George, both considered top 10 NBA players also scored poorly on the EVP rankings.

    The sites below provide information on EPV (side effects may include nausea, upset stomach, tremors, and in rare cases, death)

    As always, thoughts?

    http://www.startribune.com/sports/wolves/244048511.html
    http://grantland.com/features/expected-value-possession-nba-analytics/

    0
  • #869182
    AvatarAvatar
    B-ball fan
    Participant

     It should be mentioned, that the actual stats they used are limited by the arenas that have SVU cameras, which they admitted benefitted Chris Paul and hurt LeBron James.  

    This is definitely a promising means of analyzing players, especially point guards.  It would be great if this data eventually became tallied and publicly available sometimes in the future.  

    I like the shot satisfaction stat.  It is interesting to have a statistic that can effectively grade a player’s decision making.  While this stat obviously is largely dependent on a player’s supporting cast, but it could be nice for comparing players on the same or similar teams.  Interesting that Jamal Crawford and Steph Curry are so high.  Considering Curry’s tendency to take shots that would be bad for almost any other player in the league makes his inclusion on the list pretty amazing.

     

    0
  • #869288
    AvatarAvatar
    B-ball fan
    Participant

     It should be mentioned, that the actual stats they used are limited by the arenas that have SVU cameras, which they admitted benefitted Chris Paul and hurt LeBron James.  

    This is definitely a promising means of analyzing players, especially point guards.  It would be great if this data eventually became tallied and publicly available sometimes in the future.  

    I like the shot satisfaction stat.  It is interesting to have a statistic that can effectively grade a player’s decision making.  While this stat obviously is largely dependent on a player’s supporting cast, but it could be nice for comparing players on the same or similar teams.  Interesting that Jamal Crawford and Steph Curry are so high.  Considering Curry’s tendency to take shots that would be bad for almost any other player in the league makes his inclusion on the list pretty amazing.

     

    0
  • #869229
    AvatarAvatar
    DolanCare
    Participant

     I just read the article… I sugguest that you skim it, but it’s definitely a good read. EPV is absolutely a promising metric. But right now, the concept is pretty raw. It computed all actions that a player has done. While this data has infinite potential, EPV basically determines if a player is able to hit the open man and/or make the shot themselves. 

    The article discusses how screens and swing passes are never credited in the boxscore, but the article also didn’t explain how EPV quantifies these kinds of selfless moves. It also doesn’t bring up the fact that having great teammates would dramatically help a player’s EPV. Especially when those teammates can catch-and-shoot and/or finish up around the rim. According to this article, one’s EPV should skyrocket if you have Blake Griffin on you team. 

    Again, I’m not knocking this new statistic. In fact, if it continues to grow along side SportsVU cameras it could very well be the future of assessing talent. But I wouldn’t put any stock on how to catalogs current players. 

     

     

     

     

    0
  • #869335
    AvatarAvatar
    DolanCare
    Participant

     I just read the article… I sugguest that you skim it, but it’s definitely a good read. EPV is absolutely a promising metric. But right now, the concept is pretty raw. It computed all actions that a player has done. While this data has infinite potential, EPV basically determines if a player is able to hit the open man and/or make the shot themselves. 

    The article discusses how screens and swing passes are never credited in the boxscore, but the article also didn’t explain how EPV quantifies these kinds of selfless moves. It also doesn’t bring up the fact that having great teammates would dramatically help a player’s EPV. Especially when those teammates can catch-and-shoot and/or finish up around the rim. According to this article, one’s EPV should skyrocket if you have Blake Griffin on you team. 

    Again, I’m not knocking this new statistic. In fact, if it continues to grow along side SportsVU cameras it could very well be the future of assessing talent. But I wouldn’t put any stock on how to catalogs current players. 

     

     

     

     

    0
  • #869391
    AvatarAvatar
    TallmanNYC
    Participant

    Loves EVP from last year is low because he was missing a ton of shots. This year he is a beast and his EVP is going to be high. Love helps the Wolves get wins tremendously.

    Rubio on the other hand has a serious scoring problem. The inability to score from anywhere on the court is a problem and yes it is part of the reason the Wolves lose.

    But the main problem for the Wolves is their bench. Rubio will probably improve his shooting percentage a bit. And he brings a lot of other things to the table. So he isn’t even close to being the worst player in the league.  

    0
  • #869285
    AvatarAvatar
    TallmanNYC
    Participant

    Loves EVP from last year is low because he was missing a ton of shots. This year he is a beast and his EVP is going to be high. Love helps the Wolves get wins tremendously.

    Rubio on the other hand has a serious scoring problem. The inability to score from anywhere on the court is a problem and yes it is part of the reason the Wolves lose.

    But the main problem for the Wolves is their bench. Rubio will probably improve his shooting percentage a bit. And he brings a lot of other things to the table. So he isn’t even close to being the worst player in the league.  

    0
  • #869395
    AvatarAvatar
    thparadox
    Participant

    You don’t know how it works? Perhaps read the paper too, and then ask some questions?

    http://www.sloansportsconference.com/wp-content/uploads/2014/02/2014_SSAC_Pointwise-Predicting-Points-and-Valuing-Decisions-in-Real-Time.pdf

    TallmanNYC makes a great point. EPV has to be placed in the context of the team (just like +/-, and defensive ratings)

    Chris Paul would probably do very well in these rankings regardless of team, because he is such a great shooter and scorer… whereas Rubio is unlikely to look good because his shooting is painful, and so he is heavily dependent on his team. Even with Love and Kevin Martin, the wolves are a bit of a mess when it comes to spacing this year.

    I think we will see further stats which break down the EPV into shooting, passing, etc. There is some discussion of this in the paper, where Carmelo comes out as a bit of a "selfish shooter". It’s really interesting how well Jamal Crawford both of these metrics. Also notable, there are 3 blazers on the list. When you consider Batum is also on their team, it helps to explain their success this season.

    Coaching will obviously have a major impact. It’s smart that they used the Spurs as an example. Because the Spurs run the same plays repeatedly, the stats will be much more meaningful for their team.

    The opposite might be a team like the Rockets, who don’t really run plays, and have some real disparities from game to game. For example, Dwight scoring 25 points one night, with 20 FTa. So EPV would probably be less reliable for them.

    From the paper:

     

    Top 10 EPVA                                          Bottom 10 EPVA

     

     

    Dirk Nowitzki 2.60                                      Kevin Love -2.38 

    Deron Williams 2.52                                      Russell Westbrook -2.07 

    Stephen Curry 2.50                                       Evan Turner -1.90 

    Jamal Crawford 2.50                                  Austin Rivers -1.84 

    Greivis Vasquez 2.46                                 Rudy Gay -1.75 

    LaMarcus Aldridge 2.40                             Jrue Holiday -1.51 

    Steve Nash 2.09                                          Paul George -1.49 

    Wesley Matthews 2.06                              Singleton -1.48 

    Damian Lillard 1.95                                   Roy Hibbert -1.44 

     

    Top 10 shot satisfaction                               Bottom 10 shot satisfaction

     

    Steve Nash 0.340                                          Daniel Gibson -0.082 

    Pablo Prigioni 0.335                                    Ricky Rubio -0.067 

    Chris Paul 0.334                                          Patrick Beverley -0.046 

    Jamal Crawford 0.310                                  Michael Beasley -0.033 

    Jared Dudley 0.286                                       Andre Miller -0.005 

    Martell Webster 0.283                                  Luc Richard Mbah a Moute -0.005 

    Stephen Curry 0.258                                     George Hill 0.001 

    Amir Johnson 0.256                                     Evan Turner 0.001 

    Patrick Mills 0.255                                       Glen Davis 0.010 

     

     

    0
  • #869289
    AvatarAvatar
    thparadox
    Participant

    You don’t know how it works? Perhaps read the paper too, and then ask some questions?

    http://www.sloansportsconference.com/wp-content/uploads/2014/02/2014_SSAC_Pointwise-Predicting-Points-and-Valuing-Decisions-in-Real-Time.pdf

    TallmanNYC makes a great point. EPV has to be placed in the context of the team (just like +/-, and defensive ratings)

    Chris Paul would probably do very well in these rankings regardless of team, because he is such a great shooter and scorer… whereas Rubio is unlikely to look good because his shooting is painful, and so he is heavily dependent on his team. Even with Love and Kevin Martin, the wolves are a bit of a mess when it comes to spacing this year.

    I think we will see further stats which break down the EPV into shooting, passing, etc. There is some discussion of this in the paper, where Carmelo comes out as a bit of a "selfish shooter". It’s really interesting how well Jamal Crawford both of these metrics. Also notable, there are 3 blazers on the list. When you consider Batum is also on their team, it helps to explain their success this season.

    Coaching will obviously have a major impact. It’s smart that they used the Spurs as an example. Because the Spurs run the same plays repeatedly, the stats will be much more meaningful for their team.

    The opposite might be a team like the Rockets, who don’t really run plays, and have some real disparities from game to game. For example, Dwight scoring 25 points one night, with 20 FTa. So EPV would probably be less reliable for them.

    From the paper:

     

    Top 10 EPVA                                          Bottom 10 EPVA

     

     

    Dirk Nowitzki 2.60                                      Kevin Love -2.38 

    Deron Williams 2.52                                      Russell Westbrook -2.07 

    Stephen Curry 2.50                                       Evan Turner -1.90 

    Jamal Crawford 2.50                                  Austin Rivers -1.84 

    Greivis Vasquez 2.46                                 Rudy Gay -1.75 

    LaMarcus Aldridge 2.40                             Jrue Holiday -1.51 

    Steve Nash 2.09                                          Paul George -1.49 

    Wesley Matthews 2.06                              Singleton -1.48 

    Damian Lillard 1.95                                   Roy Hibbert -1.44 

     

    Top 10 shot satisfaction                               Bottom 10 shot satisfaction

     

    Steve Nash 0.340                                          Daniel Gibson -0.082 

    Pablo Prigioni 0.335                                    Ricky Rubio -0.067 

    Chris Paul 0.334                                          Patrick Beverley -0.046 

    Jamal Crawford 0.310                                  Michael Beasley -0.033 

    Jared Dudley 0.286                                       Andre Miller -0.005 

    Martell Webster 0.283                                  Luc Richard Mbah a Moute -0.005 

    Stephen Curry 0.258                                     George Hill 0.001 

    Amir Johnson 0.256                                     Evan Turner 0.001 

    Patrick Mills 0.255                                       Glen Davis 0.010 

     

     

    0
  • #869415
    AvatarAvatar
    Bajeebz
    Participant

    I said I’m not “EXACTLY” sure how it works. The concept of rating players based on the circumstantial decisions they make is not lost on me. I get it. I thought it would be cool to share. It’s just following the overall rating system that I find difficult, because advanced math is not my expertise. Also, the findings are a bit shocking. Last season, Steve Nash (while on the floor) was more effective than Paul George? Deron Williams higher than Westbrook? That’s cool. I can believe that Nash and Williams made better basketball decisions, which would make them superior if all players we’re created equal talent wise. But this isn’t a vacuum, it’s the NBA.

    They say that technology precedes art. The electric guitar was created, then Jimmy Hendrix rocked out on it. It’s cool to think these statistical innovations could open the door for a new brand of high IQ basketball, played by the most talented athletes in the world.

    0
  • #869309
    AvatarAvatar
    Bajeebz
    Participant

    I said I’m not “EXACTLY” sure how it works. The concept of rating players based on the circumstantial decisions they make is not lost on me. I get it. I thought it would be cool to share. It’s just following the overall rating system that I find difficult, because advanced math is not my expertise. Also, the findings are a bit shocking. Last season, Steve Nash (while on the floor) was more effective than Paul George? Deron Williams higher than Westbrook? That’s cool. I can believe that Nash and Williams made better basketball decisions, which would make them superior if all players we’re created equal talent wise. But this isn’t a vacuum, it’s the NBA.

    They say that technology precedes art. The electric guitar was created, then Jimmy Hendrix rocked out on it. It’s cool to think these statistical innovations could open the door for a new brand of high IQ basketball, played by the most talented athletes in the world.

    0
    • #869724
      AvatarAvatar
      thparadox
      Participant

      I’m sorry, it wasn’t supposed to sound condescending. I thought from the sentence after that you were looking for clarification, but I misunderstood.

      Hmmm, I’d say it’s more a measure of decisionmaking "value" than effectiveness. I think point guards are going to have a much higher leveraged EPV because they have the ball so much and make so many decisions. Since they start every play, they get rewarded or punished for every play (whereas the screen setter may have been crucial, but may not get credit in EPV right now).

      That’s why a CP3 or Nash will be very high, and Rubio will be very low. A guy like Paul George, who was forced to score a lot last year, and therefore turned the ball over, took some bad shots, etc. comes out looking a lot worse. Perhaps it would make more sense to adjust for position / usage / etc.

      I love advanced stats, so I was glad you posted this

      0
    • #869618
      AvatarAvatar
      thparadox
      Participant

      I’m sorry, it wasn’t supposed to sound condescending. I thought from the sentence after that you were looking for clarification, but I misunderstood.

      Hmmm, I’d say it’s more a measure of decisionmaking "value" than effectiveness. I think point guards are going to have a much higher leveraged EPV because they have the ball so much and make so many decisions. Since they start every play, they get rewarded or punished for every play (whereas the screen setter may have been crucial, but may not get credit in EPV right now).

      That’s why a CP3 or Nash will be very high, and Rubio will be very low. A guy like Paul George, who was forced to score a lot last year, and therefore turned the ball over, took some bad shots, etc. comes out looking a lot worse. Perhaps it would make more sense to adjust for position / usage / etc.

      I love advanced stats, so I was glad you posted this

      0

You must be logged in to reply to this topic. Login