2019-20 team win projection contest

Home for all your discussion of basketball statistical analysis.
jgoldstein34
Posts: 249
Joined: Fri Sep 30, 2016 6:38 pm

Re: 2019-20 team win projection contest

Post by jgoldstein34 »

nbacouchside wrote: Tue Sep 03, 2019 2:41 pm
Crow wrote: Mon Sep 02, 2019 8:19 pm How much of that Pelicans win projection is Zion?
I used Jacob Goldstein's PIPM projections for rookies. Zion ranks as the most valuable player on the Pelicans this year, so he's a big part of it.

The other big part of it is that Jaxson Hayes grades highly (I'm dubious on this), and there just aren't many guys expected to get minutes (via Kevin Pelton's minutes projections) that aren't at least pretty solid by the Tracking RAPM Estimate numbers I put together.
One thing that would help them look more normal is adjusting the scale of my PIPM projections to fit TRE. I haven't checked the exact difference, but visually PIPM seems to have a larger spread so a rookie heavy team would be benefiting.
nbacouchside
Posts: 151
Joined: Sun Jul 14, 2013 4:58 am
Contact:

Re: 2019-20 team win projection contest

Post by nbacouchside »

jgoldstein34 wrote: Wed Sep 04, 2019 9:53 pm One thing that would help them look more normal is adjusting the scale of my PIPM projections to fit TRE. I haven't checked the exact difference, but visually PIPM seems to have a larger spread so a rookie heavy team would be benefiting.
Finished tweaking TRE and then re-scaled the PIPM rookie projections to match stdev spread of TRE. Got updated results.

Image
eminence
Posts: 223
Joined: Sun Sep 10, 2017 8:20 pm

Re: 2019-20 team win projection contest

Post by eminence »

edit
Last edited by eminence on Mon Oct 21, 2019 1:44 pm, edited 1 time in total.
Crow
Posts: 10536
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2019-20 team win projection contest

Post by Crow »

Hawks at 28 isn't what Hawks fans and certain T. Young hipsters want to hear.

'Nets at 35? Can they handle that?
eminence
Posts: 223
Joined: Sun Sep 10, 2017 8:20 pm

Re: 2019-20 team win projection contest

Post by eminence »

I really like the Hawks going forward, seem to have a real vision for their team (have to be careful not to force that vision, but I think for now they're doing okay), just don't see it this season.

On the Nets, I kind of expect them to not be able to take it (not a Kyrie guy) probably one of the higher variance teams in that range though (maybe DeAndre really takes off again like next to CP3, but I'm not projecting it).

Wound up coming around on the Magic/Pacers a bit, though possibly still a bit lower than average.
eminence
Posts: 223
Joined: Sun Sep 10, 2017 8:20 pm

Re: 2019-20 team win projection contest

Post by eminence »

Coming around a bit on the Hawks, think the rooks look relatively ready and it seems to be meshing pretty well. Still looking at 30-35 wins, but feeling better about where they're at.

Would it be better for me to post a new post for my updated predictions or just update my 1st post? (they're finalized but I'll wait until a couple days prior to the season to post).
Crow
Posts: 10536
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2019-20 team win projection contest

Post by Crow »

Oh, you could do either but posting new and marking it as updated / revised should be fine.

Post by Oct.22 start of games.

(Slight delay of a day or two probably tolerated with explanation if not perceived / challenged as seeking advantage. Such as a noticeably different from pack win projection after a game 1 injury. Projections published elsewhere and publicly dated by this time can be brought in a little later on honor. Way later might be a problem.)
Mike G
Posts: 6144
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: 2019-20 team win projection contest

Post by Mike G »

Subject to revisions and reviews, based on K Pelton minutes linked upthread.

Code: Select all

west  wins     east   wins
LAL   54.5      Phl   53.8
Hou   53.6      Mil   51.8
LAC   52.0      Orl   44.9
Uta   50.6      Ind   44.7
Por   46.7      Bos   42.6
SAS   45.4      Det   42.3
Den   44.9      Brk   41.2
Dal   40.6      NYK   40.7

GSW   40.5      Tor   37.6
NOP   39.8      Cle   36.7
Sac   39.6      Chi   36.7
Min   36.5      Mia   36.1
Phx   35.8      Atl   35.2
OkC   29.0      Was   28.3
Mem   22.2      Cha   25.6
Kevin P has some funny minute projections, but I'll go with it.
I've made no attempt to account for player chemistry or near-identical superstars on the same team -- Harden/RW, George/Kawhi -- and suspect they can't maintain productivity together.
After scaling to 41 wins per team, the conference imbalance is extreme. So I added 0.7 win to each East team and cut the West by 0.7 each.

EDIT Oct. 17. Kevin shuffled some minutes, mostly to the benefit of weaker teams. Cavs gained 1.6 expected wins, with Sindarius Thornwell's PT going elsewhere.

Dubs add Marquese Chriss and the #39 pick, give even more time to GRob2, reduce McKinnie, drop to #10 in the west.
Once flush with scorers, they now have Curry, then D'Angelo, then ... Cauley-Stein?

EDIT2 - Oct. 22. Zion minutes reduced and a few other changes.
Crow
Posts: 10536
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2019-20 team win projection contest

Post by Crow »

Lowest estimate I've seen yet for Celtics.

OKC trading heavy and early or expected lousy even before then?
Mike G
Posts: 6144
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: 2019-20 team win projection contest

Post by Mike G »

Celtics add up like this. Last year's minutes, this year's according to KP, last year's eWins/484 (1.00 is avg) and eWins = e484*Min/484

Code: Select all

Bos 2019-20   min19   Min?   e484   eW
Kemba Walker   2863   2652   1.79   9.8
Jayson Tatum   2455   2496   1.11   5.7
Jaylen Brown   1913   2310    .85   4.1
Marcus Smart   2200   2280    .71   3.3
Gordon Hayward 1863   1860   1.04   4.0

Enes Kanter    1640   1848   1.72   6.6
Daniel Theis    908   1155   1.09   2.6
Robert Williams 283    936   1.00   1.9
Grant Williams  r 22   936    .53   1.0
Brad Wanamaker  343    780    .86   1.4

Vincent Poirier r 80   780    .31    .5
Romeo Langford  r 14   633    .64    .8
Semi Ojeleye    594    624    .16    .2
Carsen Edwards  r 33   390    .45    .4
Rookies are shown by r and their draft pick#. Rookie eW/484 = (1/r)^0.43*2.0
The #1 pick (Zion) thus is assigned e484 = 2.00, the #2 = 1.49 etc.
These parameters give all rookies a projected eWin total of ~60, which is a modest average year. Probably more will see minutes than what Kevin suggested, esp. later in the season.

Subjectively, Horford was a huge loss. Morris was almost their best player in the postseason. Rozier and Baynes... Kanter and some rookies are not at all equivalent.

edit -- Just realized I have neither Tor or GSW making the playoffs.
tarrazu
Posts: 91
Joined: Mon Aug 04, 2014 5:02 pm

Re: 2019-20 team win projection contest

Post by tarrazu »

With lack of cap space across league and the number of first round picks OKC currently has, are they as much sellers as everyone makes them out to be?
Crow
Posts: 10536
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2019-20 team win projection contest

Post by Crow »

There will be offers for Gallinari. Not sure how much Thunder can get back. Some chance it could end up a buyout, if they don't like the offers.

Adams probably stays at least til next summer. Paul probably til next summer or maybe after that.
Mike G
Posts: 6144
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: 2019-20 team win projection contest

Post by Mike G »

About 124 players will be with a new team and getting some minutes, according to the K Pelton spreadsheet.
Unfortunately, the talent again migrated from East to West, worsening that disparity. Even with the West guys expected to get more minutes in the East, the projected impact favors the West.

Code: Select all

east to West  eWins    eWins   west to East
Kawhi Leonard  9.7      8.0   Julius Randle
H Whiteside    9.2      6.6   Enes Kanter
D'A Russell    9.0      4.5   Derrick Rose
Boj Bogdanovic 5.5      4.0   T.J. Warren
Dewayne Dedmon 4.4      3.7   Elfrid Payton
J.J. Redick    4.0      3.4  Al-Farouq Aminu
Ed Davis       3.5      2.7   Davis Bertans
Dwight Howard  3.2      2.6   Meyers Leonard
Danny Green    3.1      2.3  Markieff Morris
Pau Gasol      3.0      2.3   Evan Turner
Shabazz Napier 2.7      2.1   Damian Jones
Frank Kaminsky 2.7      1.7   Trey Burke
Noah Vonleh    2.7      1.7   Raul Neto
Kent Bazemore  2.6      1.6   Justin Holiday
Emm. Mudiay    2.6      1.6   Moritz Wagner
DeM. Carroll   2.6      1.5  Stanley Johnson
Aron Baynes    2.3      1.5   Taj Gibson
Cory Joseph    1.7      1.4   C.J. Miles
Trevor Ariza   1.6      1.2   Isaiah Thomas
Jeff Green     1.5       .7  Wilson Chandler
Mario Hezonja  1.5       .7   Kyle Korver
Omari Spellman 1.4       .5   Garrett Temple

total E to W  82.9     58.3   total W to E
These are just last year's eWin rates X this year's expected minutes.
If Durant were playing in Brk, it would be closer. But up and down the line, the East gets shafted.
Except for the Knicks -- they got a few of the western defectors. Randle could be allstar; Payton could bust out...
Mike G
Posts: 6144
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: 2019-20 team win projection contest

Post by Mike G »

Players who got 1000+ minutes last year and don't seem to be on a roster this season. Corrections are encouraged.

Code: Select all

2143  Darren Collison   Ind
1988   Josh Jackson     Phx
1885   Dwyane Wade      Mia
1757   Darius Miller    NOP
1578   Andre Iguodala   GSW
1481   Iman Shumpert    Hou
1439   Wayne Selden  Mem-Chi
1436   Jeremy Lin       Tor
1402   Tyreke Evans     Ind
1246   Shelvin Mack  Mem-Cha
1245   Nikola Mirotic NO-Mil
1218   Jonas Jerebko    GSW
1211   Jamal Crawford   Phx
1123   Lance Stephenson LAL
1104   John Wall        Was
1083   Justin Jackson   Sac
1071   Devin Harris     Dal
1064  Jonathon Simmons  Phl
1015   Nik Stauskas  Por-Cle
1003   Tony Parker      Cha
Post Reply