2015-16 Team win projections

[Crow]

Subjective projections from ESPN crowd:

http://espn.go.com/nba/story/_/id/13431 ... -standings
http://espn.go.com/nba/story/_/id/13431 ... -standings

[Crow]

Pacers and Celtics may be among the tougher projections. Haven't studied it closely but found myself surprisingly wondering if Lakers might outpeform. Just a question. 26 wins isn't much.

[Mike G]

Good regression to the mean; there's a -0.49 correlation between last year's Wins and additional wins next season.
Teams with 49+ wins last year are expected to win 4.3 fewer games in 2016, on avg
Those who won 37-46 games should win 1.6 more on avg
The worst teams, with 33 or fewer wins last year avg another 3.4 in these projections.

Code: Select all

west  2015  2016  impr       east  2015  2016  impr
GSW    67    60    -7        Cle    53    59     6
SAS    55    57     2        Atl    60    50   -10
Hou    56    56     0        Chi    50    50     0
LAC    56    56     0        Mia    37    47    10
OKC    45    55    10        Was    46    46     0

Mem    55    51    -4        Tor    49    44    -5
NOP    45    47     2        Mil    41    44     3
Dal    50    41    -9        Bos    40    40     0
Uta    38    40     2        Ind    38    39     1
Phx    39    35    -4        Cha    33    35     2

Por    51    31   -20        Det    32    35     3
Sac    29    31     2        NJN    38    30    -8
Den    30    27    -3        Orl    25    30     5
LAL    21    26     5        NYK    17    25     8
Min    16    24     8        Phl    18    19     1

avg    43.5  42.5  -1        avg    38.5  39.5   1

Early projection forecasts the east-west disparity to be reduced by 40%

[Crow]

If the divisional adjustment occurs it would roll the edge back to 2011-2 levels. I kinda doubt it goes that far but between 2013-4 and 2014-5 the west's edge did fall 20% from its recent period high.

[ampersand5]

Andrew Johnson's projections are up:

http://nyloncalculus.com/2015/09/30/hig ... s-2015-16/

My only piece of feedback is for how easy it is to include the effect of fatigue (back to back/four in five), I'm not sure why it's not being added to the numbers.

[ampersand5]

Kevin Ferrigan's:

[Mike G]

how easy it is to include the effect of fatigue (back to back/four in five)...

Do some teams have a lot more of these than others?

[ampersand5]

how easy it is to include the effect of fatigue (back to back/four in five)...

Do some teams have a lot more of these than others?

I haven't seen the numbers for the 2016 season yet, but typically yes.

[Crow]

Andrew, did you consider age adjusting offense and defense separately given their different RAPM age curves? Have you look at which coaches minute allocations vary from the model the most? Overall, for vets - young guys split, good players - bad, perimeters - bigs, etc.? Both could improve the model, with sufficient time.

[bbstats]

Are we going to do the contest in this thread?

[Crow]

Could. I created the thread with that option. But if someone sees any reason to start fresh, go for another thread.

[Mike G]

Using Kevin Pelton's minutes,
viewtopic.php?f=2&t=9005
... I get an average of 43.2 eWins per team. This is 45.4 wins expected: It's a pretty optimistic distribution of minutes.
Scaling every team down equally, I get these; you may consider eW to be a 50% regression to the mean, and the columns represent a range.

Code: Select all

west    xW      eW         east    xW      eW
LAC    67.5    54.3        Mia    56.1    48.6
SAS    63.3    52.1        Atl    55.4    48.2
GSW    61.9    51.5        Cle    53.9    47.4
Okl    60.6    50.8        Tor    44.4    42.7
Hou    57.6    49.3        Chi    42.4    41.7

Mem    46.9    44.0        Ind    42.2    41.6
NOP    46.7    43.9        Det    41.9    41.5
Sac    45.2    43.1        Mil    40.6    40.8
Uta    42.1    41.6        Bos    34.8    37.9
Dal    35.3    38.2        Cha    33.1    37.0

Por    33.2    37.1        Was    32.8    36.9
Phx    32.8    36.9        Orl    30.6    35.8
Min    28.0    34.5        Brk    23.3    32.2
Den    24.2    32.6        NYK    20.4    30.7
LAL    22.0    31.5        Phl    10.8    25.9

avg    44.5    42.7        avg    37.5    39.3

Miami rules the East! Bosh gets 60% more minutes, Wade 11% more, Whiteside 90% more than last year. Headed to the bench are Chalmers, Haslem, Tyler Johnson, and their woeful rates.

Out west, it looks like the Clippers' year. They added quality depth with Pierce, Smoove, Lance, and Aldrich. Actually, there's great uncertainty with those guys. But no boost required in anyone's minutes -- just continued health.

I didn't do any age adjustment. In another thread, we see that aging players produce more wins with a given point differential. So these factors may just about offset.

[Dr Positivity]

* deleted because of post later in thread

[tacoman206]

I'll post my team projections in here once I finish up. Just wondering, do you guys also have a contest for individual player "stat line" projections? If so, I'd like to enter that one as well.

[Crow]

Haven't had one of those, but start a thread up if you like.