Announcement

Collapse
No announcement yet.

Raptors: Career-Arc Regression Model Estimator with Local Optimization (CARMELO)

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

  • #16
    Based on the title, I assumed it was an Onion piece....

    Comment


    • #17
      Scraptor wrote: View Post
      http://fivethirtyeight.com/features/...rs-in-the-nba/

      Scan this list of franchise players and tell me this model isn't broken. I don't care how smart Nate Silver and co are. If your model produces a list that tells you Elfrid Payton and Marcus Smart will be the 13th/14th most valuable players over the next six seasons, you should check your math.
      This is the website that claimed Andrew Wiggins would be worse than James Posey.

      They are an embodiment of everything that the "traditionalists" hate on about the "analytics guys". That is to say, they just plug in numbers and don't really give any thought to context or common sense.

      This model gives a bad name to legit statistical tools and i wish they would just piss off with it already.
      Last edited by KHD; Tue Oct 27th, 2015, 09:38 PM.

      Comment


      • #18
        Dwight Howard's best comparable: Emeka Okafor

        That reminds me of all the intense debates before their draft. It looks like we are back full circle!

        Comment


        • #19
          KHD wrote: View Post
          This is the website that claimed Andrew Wiggins would be worse than James Posey.

          They are an embodiment of everything that the "traditionalists" hate on about the "analytics guys". That is to say, they just plug in numbers and don't really give any thought to context or common sense.

          This model gives a bad name to legit statistical tools and i wish they would just piss off with it already.
          Yep, most ESPN and related analytics writers are just too extreme. Collect some math, throw numbers together into various combinations, see what you get, and hopefully, hopefully there's some real insight from this. If not, at least it will get some page views.

          At least Hollinger was a good writer and talker, and in his later years he started to have some real basketball insights, he was becoming a little bit more like Zach Lowe on his good days. These young replacements do what Hollinger did, but without ability to write quite as well or the experience.

          Comment


          • #20
            When analytics go horribly wrong...

            Comment


            • #21
              KHD wrote: View Post
              This is the website that claimed Andrew Wiggins would be worse than James Posey.

              They are an embodiment of everything that the "traditionalists" hate on about the "analytics guys". That is to say, they just plug in numbers and don't really give any thought to context or common sense.

              This model gives a bad name to legit statistical tools and i wish they would just piss off with it already.
              All this comes down to sample size though. I would much rather have numbers that backup some sort of prognostication than some guy saying "I've got a gut feeling about this guy...". With Wiggins, we have to understand to take any rookie projection with a grain of salt because projections are based on a single seasons worth of statistics. Statistically speaking, Wiggins did not have a great rookie season.
              your pal,
              ebrian

              Comment


              • #22
                Basketball-Reference.com has a similar type of stat, I think it's called Similarity Scores, and it just lists a bunch of other players with similar career arcs. It's not like it's an end-all type of stat, but it does help to give an estimation.

                I'd rather look at numbers than listen to say a Bryan Colangelo's "He's the next Dirk Nowitzki!" comment and think there's some reliability behind it.
                your pal,
                ebrian

                Comment


                • #23
                  BobLoblaw wrote: View Post
                  Yep, most ESPN and related analytics writers are just too extreme. Collect some math, throw numbers together into various combinations, see what you get, and hopefully, hopefully there's some real insight from this. If not, at least it will get some page views.

                  At least Hollinger was a good writer and talker, and in his later years he started to have some real basketball insights, he was becoming a little bit more like Zach Lowe on his good days. These young replacements do what Hollinger did, but without ability to write quite as well or the experience.
                  At least these guys, unlike the Dave Berri's of the world, are appropriately cautious and humble about this model - they don't claim it's perfect and if it looks wrong it's because you are too stupid to understand.

                  Still, sports economists are mostly wrong and useless just like every other type of economist. This should surprise no one.

                  Comment

                  Working...
                  X