This topic contains 4 replies, has 4 voices, and was last updated by AvatarAvatar sweaterflex 6 years, 6 months ago.

  • Author
    Posts
  • #67201
    AvatarAvatar
    Rdf1212
    Participant

     Catching up on the forum I was reading the debate between CP3 and Wall.. It got me thinking.. Many of the arguments were analytical based but there wasnt much discussion from a winning standpoint. I saw many people argue CP3 is the most underrated player in the league. Is it possible CP3 running your team may work well from an analytics standpoint individually but doesn’t translate to team playoff success? Is there a correlation between who has the best analytics/efficiency on your team, which position they play and how many wins the team averages? It’s obvious LBJ can get his and provide wins to the team but is it possible these efficiency and analytics dont always  translate to winning basketball?  I would love to see efficiency stats compared to position:wins to see if there’s a correlation between what position is most efficient on your team and how many wins on average it can provide.

    0
  • #1105584
    Robb_CRobb_C
    Robb_C
    Participant

    Positions are meaningless first off, and secondly Chris Paul is a PG and cant really control what a player does with the ball after it leaves Pauls hand.. Does Paul put players in positions to score? Absolutely, A big reason why the Rockets signed Paul was his ability to create in the mid range for his teammates something Harden lacked.. Data/Analytics proves this true.

    If you do the homework you’ll see the reasons for Chris Pauls struggles is he tends to over dominate the ball instead of pushing the tempo, its a recipe for disaster if your team isn’t hitting shots..

    Some believe this is why the Rockets and Paul are a great marriage.. they hide each other’s weaknesses.. 

    Essentially thats the purpose of data/analytics approach is being able to break things down to their least common denominator.

     

     

    0
  • #1105585
    AvatarAvatar
    sweaterflex
    Participant

     This Jonathan Tjarks article addresses the Chris Paul analytics issue pretty well. The NBA has had some 70ish NBA champion teams, and the game has changed a great deal, but historically the point guard being the best player on a team is less likely to lead to a champion, though it’s not a super strong relationship. Magic was the best player on 3/5 of his champion teams, Billups was the best on that Pistons team, I give Zeke credit for one of those 80s Pistons teams, Steph in 2015, Walt Frazier on one of those Knicks teams (credit to Willis Reed), and Bob Cousy on the Celts in the 50s rounds out the list as far as I can tell. If you don’t "count the ringzz," only two of the top 20 all time in Win Shares are PGs, though that’s pretty arbitrary.

    The only thought that occurs to me is that it is difficult for point guards to be as signficant of contributors on both sides of the court (pgs rarely win DPOY) as centers and athletic wings.

     

    0
  • #1105592
    AvatarAvatar
    BeastMode716
    Participant

     in the Analytics basket

    Remember these statistics were originally designed for Major League Baseball & track stats over a Long 162 game season where Individual match ups are Not an issue

    When we’re talking about Basketball or Football Every single Game the Individual match up takes presedence over all else

    This means it is almost impossible to track trends b/c variables change from game to game 

    In addition, when you pull apart the algorithm’s there are inconsistencies for example, in the Defense Analytics algorythm a Steal is worth Double a blocked shot – this is patently ridiculous b/c players w/ lots of steals are usuallt poor defenders so they must cheat in the passing lanes – for example, Steph Cirryu & Allen Iverson were both steals leaders for exactly this reason.

    Furthermore, a Top shot Blocker is Completely disrupting the opposing offense by Altering probably 3 shots for every one he blocks & he intimidates opponents to Not come into the lane & attack the basket. None of this shows up in the stats. 

    The thinking was a steal changes possession Every time but a black does not – that is flawed logic & there are many examples just like this when you pull these Valuations apart.

    This is why it is important to make Analytics Part of your overall assessment – it’s a Very important part but I’d refrain from making arguments based entirely on analytics

     

     

     

     

     

     

     

    0
    • #1105596
      AvatarAvatar
      sweaterflex
      Participant

      Using blocks and steals as determinants of defensive value in combined stats is going by the wayside, though this is one of the flaws in using PER and DBPM. For every Russell Westbrook and Rajon Rondo gambling in passing lanes there is a JaVale McGee and Hassan Whiteside going for a block and missing a rotation.

      RPM fixes a lot of these issues- Rubio and Marcus Smart get a lot a steals and play killer man to man defense and are valued properly, whereas Curry and Westbrook are essentially net neutral despite high steal rates. RPM factors in wingspan, which can sometimes unnecessarily reward/punish some players, but is generally pretty accurate (this may be why IT3 rates lower than similarly bad defenders). Nylon Calculus’s rim protection stats do a good job measuring impact of bigs.

      As a side, steals are a bigger deal than blocks. Possessions after live ball turnovers have a higher likelihood of turning in to transition buckets on the other end. A lot of blocks are spiked in to the stands or recovered by the shooting team, leading to very little change in outcome. This Neil Paine article is a good read.

       

      0

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