Announcement

Collapse
No announcement yet.

Predicting the Raptors' 2017-18 Win Total

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

  • Predicting the Raptors' 2017-18 Win Total

    Hey all

    I just posted an article on Raps HQ with my annual statistical projection for the Raptors, plus projections for every team in the league. Feel free to go read it in full for a more detailed explanation of the process, but I'll put all the necessary info here anyway.

    https://www.raptorshq.com/2017/8/1/1...ns-projections

    The process is much the same as most years I do this, with a couple of tweaks. The goal, as ever, is to build this up with as little manual modification as possible - so none of my own opinion on rotations, minutes, roster construction comes into play at all. Player usage on each team is completely controlled by their minutes last year. And no assumptions are made on my part in terms of age decline, or young players getting better. Everyone is assumed to be the same player as last year, with the same usage, with only their minutes increased or decreased based on the team around them adding or removing high minutes players from last year.

    1) Pull 2016-17 data for every player in league: total minutes played, WS/48, BPM, RPM
    2) Compile team rosters (as of July 26th in this case) and check team minutes total against 19,800 (5x48x82).
    3) Pro-rate player minutes to get correct minutes total for each team. Individual player cap at 3000 MP.
    4) Multiply MP x WS/48 then divide by 48 to get total WS for each player.
    5) Sum player WS's for each team to get team WS total.
    6) Multiply BPM by MP for each player, divide by team MP, sum all player values for team average BPM.
    7) Same for RPM.
    8) Use pythagorean wins calculation using BPMx5 (and RPMx5) as point differential.
    9) Calculate average win total for the entire league, adjust all teams up or down to ensure 41 win average.
    10) Profit!

    This results in 3 win predictions for each team. I'll present the Raptors in detail, and for league standings will use an average win total for each team with the three results equally weighted.

    So, here is the individual player data for the Raptors.



    Those numbers result in win predictions of 56 wins (by WS), 52 wins (by BPM), and 49 wins (by RPM).

    As for the league as a whole:



    Some of that seems right, other bits are somewhat whacky. I've got explanations for some outliers (Cavs in particular) in the piece, if you are interested. Or we can just discuss them here.

    My big takeaways - if the Cavs do implode because of Kyrie, the Raptors seem as poised as anyone in the East to jump in to fill the void. And the East is awful - with 12 of the 15 West teams projected to finish with a better record than the 8th seed Knicks - and the Kings are soooo close to making it 13 of 15.
    twitter.com/dhackett1565

  • #2
    2 things that jump out at me:

    1. The Cavs fall from grace.
    2. The Knicks making the playoffs with 33 wins, lol.

    Man, there will be a lot of terrible teams in the East.

    Edit: Hoping the Pelicans falter, in the off chance that Davis requests a trade

    Comment


    • #3
      Awesome job! Only issue I can see with the Raps is that it's doubtful Jonas + Bebe + Jakob will play 3770 minutes.... That would be 46mpg for 82 games. Casey will be too tempted by Ibaka not to go small more than that. But I get that you don't want to out your own projections in there for rotations.
      It's the Spicy Show! Time to cook.

      Comment


      • #4
        Nuggets are gonna be a 4 or 5 seed. I absolutely love their team and I think Jamal Murray, Garry Harris and Nikola Jokic are all ready to make big jumps. Paul Milsap fits perfectly for them (as he would on almost any team). Obviously these jumps would not be shown or your calculations, which is fine, but my big prediction is denver surprises a lot of people.


        Sent from my iPhone using Tapatalk

        Comment


        • #5
          And I'm surprised Boston's wins are so high. Here's a question for you Dan, how do these calculations value the addition of a rookie? Do they have an impact on the team's results at all, and if so how do you incorporate them into your calculations? I'd be interested to know.


          Sent from my iPhone using Tapatalk

          Comment


          • #6
            Maury wrote: View Post
            And I'm surprised Boston's wins are so high. Here's a question for you Dan, how do these calculations value the addition of a rookie? Do they have an impact on the team's results at all, and if so how do you incorporate them into your calculations? I'd be interested to know.


            Sent from my iPhone using Tapatalk
            Rookies are not included at all, as they have 0 minutes played the prior season, so are given 0 MPG this season too.
            twitter.com/dhackett1565

            Comment


            • #7
              I remember you did this last year. I'd be curious to find the link to last year's results and see how close it came!
              It's Klaw Season. Time to hunt.

              Comment


              • #8
                KeonClark wrote: View Post
                I remember you did this last year. I'd be curious to find the link to last year's results and see how close it came!
                The link from last year is in the article. The model last year came up with 58 wins.

                Comment


                • #9
                  ogi wrote: View Post
                  The link from last year is in the article. The model last year came up with 58 wins.
                  Yeah, last year's model had two issues that I didn't bother with until this year. First, I didn't have a minutes cap on it (which put Lowry's minutes way above what he could possibly play, and he's a huge driver for the team's success), and second, I didn't adjust the win totals to make the league average 41 - so every team had an overprediction of 2-3 wins just from that alone. The other factor is that using WORP as I did there tends to overpredict wins for some reason, while using BPM/RPM (the source for WORP) tends to underpredict a little bit. It's part of why two years ago (when I used a method like the above method for the RPM stat) I underpredicted the win total, and last year I overpredicted (even applying the common sense minutes cap and correcting league wide win totals, the Raptors would have been predicted at 54 wins).
                  twitter.com/dhackett1565

                  Comment


                  • #10
                    DanH wrote: View Post
                    Hey all

                    I just posted an article on Raps HQ with my annual statistical projection for the Raptors, plus projections for every team in the league. Feel free to go read it in full for a more detailed explanation of the process, but I'll put all the necessary info here anyway.

                    https://www.raptorshq.com/2017/8/1/1...ns-projections

                    The process is much the same as most years I do this, with a couple of tweaks. The goal, as ever, is to build this up with as little manual modification as possible - so none of my own opinion on rotations, minutes, roster construction comes into play at all. Player usage on each team is completely controlled by their minutes last year. And no assumptions are made on my part in terms of age decline, or young players getting better. Everyone is assumed to be the same player as last year, with the same usage, with only their minutes increased or decreased based on the team around them adding or removing high minutes players from last year.

                    1) Pull 2016-17 data for every player in league: total minutes played, WS/48, BPM, RPM
                    2) Compile team rosters (as of July 26th in this case) and check team minutes total against 19,800 (5x48x82).
                    3) Pro-rate player minutes to get correct minutes total for each team. Individual player cap at 3000 MP.
                    4) Multiply MP x WS/48 then divide by 48 to get total WS for each player.
                    5) Sum player WS's for each team to get team WS total.
                    6) Multiply BPM by MP for each player, divide by team MP, sum all player values for team average BPM.
                    7) Same for RPM.
                    8) Use pythagorean wins calculation using BPMx5 (and RPMx5) as point differential.
                    9) Calculate average win total for the entire league, adjust all teams up or down to ensure 41 win average.
                    10) Profit!

                    This results in 3 win predictions for each team. I'll present the Raptors in detail, and for league standings will use an average win total for each team with the three results equally weighted.

                    So, here is the individual player data for the Raptors.



                    Those numbers result in win predictions of 56 wins (by WS), 52 wins (by BPM), and 49 wins (by RPM).

                    As for the league as a whole:



                    Some of that seems right, other bits are somewhat whacky. I've got explanations for some outliers (Cavs in particular) in the piece, if you are interested. Or we can just discuss them here.

                    My big takeaways - if the Cavs do implode because of Kyrie, the Raptors seem as poised as anyone in the East to jump in to fill the void. And the East is awful - with 12 of the 15 West teams projected to finish with a better record than the 8th seed Knicks - and the Kings are soooo close to making it 13 of 15.
                    Great stuff always enjoy reading these each year. I think there's a chance that the cavs do take a small step back in the regular season due to the loss of kyrie (obviously dependent on what they get back for him). But on a separate note, is this the year lebron finally shows some semblance of declining? It's absolutely insane that he's managed to play at such a high level for so long without a single serious injury to sideline him. We all know it's going to have to happen eventually and probably sometime soon.

                    Comment


                    • #11
                      DanH wrote: View Post
                      Yeah, last year's model had two issues that I didn't bother with until this year. First, I didn't have a minutes cap on it (which put Lowry's minutes way above what he could possibly play, and he's a huge driver for the team's success), and second, I didn't adjust the win totals to make the league average 41 - so every team had an overprediction of 2-3 wins just from that alone. The other factor is that using WORP as I did there tends to overpredict wins for some reason, while using BPM/RPM (the source for WORP) tends to underpredict a little bit. It's part of why two years ago (when I used a method like the above method for the RPM stat) I underpredicted the win total, and last year I overpredicted (even applying the common sense minutes cap and correcting league wide win totals, the Raptors would have been predicted at 54 wins).
                      True. And it obviously can't account for trades and injuries which were both plentiful last year

                      Comment


                      • #12
                        Raps - 57 wins
                        Bos - 54 Wins
                        Cavs - 53 wins
                        Wizs - 48 wins
                        Bucks - 48 wins
                        Heat - 45 wins
                        Horts - 41 wins
                        6ers - 36 wins

                        Raps > 6ers in 5
                        Boston > Horents in 6
                        Cavs > Heat in 5
                        Wizs < Bucks in 7

                        Raps > Bucks in 6
                        Cavs > Bos in 7

                        Raps > Cavs in 7

                        GSW > Raps in 4
                        Abbas wrote:

                        First of all i was my own source

                        Comment


                        • #13
                          I'm stunned the Clippers are rated so high
                          Axel wrote:
                          Now Cody can stop posting about this guy and we have a poster to blame if anything goes wrong!!
                          KeonClark wrote:
                          We won't hear back from him. He dissapears into thin air and reappears when you least expect it. Ten is an enigma. Ten is a legend. Ten for the motherfucking win.
                          KeonClark wrote:
                          I can't wait until the playoffs start.

                          Until then, opinions are like assholes. Everyone has one and they most often stink

                          Comment


                          • #14
                            Projected 2017-18 records and standings for every NBA team - Kevin Pelton

                            http://www.espn.com/nba/story/_/id/2...2017-18-season

                            To help answer these questions and more, here's an early look at 2017-18 NBA projections using ESPN's real plus-minus (RPM).
                            6.(in just the EC) Toronto Raptors
                            Projected wins: 43.4

                            Perhaps the most surprising projection in the league, the Raptors suffer because of the loss of Patterson. RPM rates him as a more valuable contributor while on the court than starting power forward Serge Ibaka. RPM projects Toronto to drop to 12th in both offense and defense after ranking in the top 10 in both categories last season.
                            Yeah that is surprising. PP's departure costing so many wins is a slight overvalution I'd say. But I take pure stat based projections with a sprinkling of salt for flavour.
                            Last edited by Mess; Thu Aug 3rd, 2017, 12:30 PM.
                            Two beer away from being two beers away.

                            Comment


                            • #15
                              It's crazy to me that three recent ex-Raptors were in the top 10 for RPM at power forward: Amir, JJ, and now Patman.

                              http://www.espn.com/nba/statistics/r...RPM/position/6

                              But FWIW Pelton had us at 49 last year which means we outperformed his projections for four consecutive years. 43 seems low to me unless we get injured or Ibaka really shits the bed.
                              It's the Spicy Show! Time to cook.

                              Comment

                              Working...
                              X