2012-2013 Playoff Predictions

[v-zero]

I propose a set of predictions just like the regular season predictions, that is: how many games teams will win (and lose) in the playoffs. This will require some heavy maths or, better yet, simulations. I intend to run mine tonight. If people want to just build team strengths as they see them for the playoffs, I can then run them through a simulation to get predictions if you can't/don't want to do it yourself - as long as the team strengths are posted before saturday's games then that should be adequate.

[boooeee]

I'm in, assuming I have time to run my simulations before Saturday.

Here are some tiebreakers (according to my understanding of the rules):

- Bulls/Rockets in the Finals: Rockets get homecourt advantage, going 2-0 against the Bulls in the regular season.
- Bulls/Lakers in the Finals: They split the head to head. I think the next tiebreaker is conference record, even though the teams are from different conferences. If that's the case, then The Bulls would have home court (which doesn't really seem fair, given the relative strength of the two conferences).

I'm basing this on the NBA's website: http://www.nba.com/standings/team_recor ... d_Cnf.html.

[Mike G]

Mia 4, Mil 0
NYK 4, Bos 1
Ind 4, Atl 2
Brk 4, Chi 3
OKC 4, Hou 0
SAS 4, LAL 1
Den 4, GSW 2
LAC 4, Mem 3

Mia 4, Brk 1
NYK 4, Ind 3
OKC 4, LAC 1
SAS 4, Den 3

Mia 4, NYK 2
OKC 4, SAS 2

OKC 4, Mia 3


OKC 16-6
Mia 15-7
NYK 10-8
SAS 10-8
Den 7-6
Ind 7-6
LAC 5-7
Brk 5-7
Mem 3-4
Chi 3-4
GSW 2-4
Atl 2-4
Bos 1-4
Hou 1-4
Mil 0-4
LAL 0-4

[PD123]

OKC 4, LAL 0
SAS 4, Hou 1

It's Spurs-Lakers and Thunder-Rockets.

[Bobbofitos]

OKC 4 HOU 1
SAS 4 LAL 1
DEN 4 GSW 2
LAC 2 MEM 4

OKC 4 MEM 2
SAS 4 DEN 3

OKC 2 SAS 4

MIA 4 MIL 0
NYK 4 BOS 1
IND 4 ATL 1
BRK 2 CHI 4

MIA 4 CHI 2
NYK 4 IND 2

MIA 4 NYK 1

SAS 4 MIA 3

[v-zero]

Here are my team efficiency differentials for the playoffs:

EAST:

NYK 6.51675491102
BRK 4.96213168402
BOS 1.14155546223
IND 5.55420253713
CHI 1.85081676637
MIL 1.50873011534
MIA 11.2240636131
ATL 4.07809788666

WEST:

OKC 10.3315937495
DEN 4.27102194087
LAC 8.09369729467
GSW 3.64426835257
LAL 2.30627398469
SAS 9.26833287958
MEM 4.87452072322
HOU 4.6525130855

I will post the expected wins as soon as I can, but they are entirely determined by the above values.

[v-zero]

Ok, here are the wins and losses, and the probability of winning it all:
WINS LOSSES CHAMPIOSHIP CHANCE
NYK 7.910208 6.312232 0.040176
BRK 5.135542 5.47465 0.010022
BOS 2.188016 4.391896 0.000176
IND 5.737202 5.72705 0.015128
CHI 2.628796 4.568294 0.000406
MIL 1.255956 4.121238 0.000204
MIA 13.10399 5.98238 0.48063
ATL 3.77877 4.97638 0.003422
OKC 10.055872 6.379802 0.21931
DEN 4.568196 5.348936 0.005136
LAC 6.196058 5.674248 0.046758
GSW 3.720788 4.965528 0.002138
LAL 1.85632 4.286546 0.000436
SAS 10.376372 6.499156 0.169524
MEM 3.19111 4.824364 0.004218
HOU 2.23022 4.400716 0.002316

[PD123]

OKC-4 HOU-1
SAS-4 LAL-1
DEN-4 GSW-0
LAC-4 MEM-2

OKC-4 LAC-2
SAS-2 DEN-4

OKC4 DEN-2


MIA-4 MIL-1
NYK-4 BOS-1
IND-4 ATL-1
BRK-4 CHI-2

MIA-4 BRK-0
NYK-4 IND-2

MIA-4 NYK-3

MIA-4 OKC-2

MIA: 16-6
NYK: 11-7
IND: 6-5
BRK: 4-4
CHI: 2-4
ATL: 1-4
BOS: 1-4
MIL: 1-4

OKC: 14-9
SAS: 6-5
DEN: 8-4
LAC: 6-6
MEM: 2-4
GSW: 0-4
LAL: 1-4
HOU: 1-4

[DSMok1]

Here are my playoff ratings, for reference (haven't simmed out yet):

Tm Value
OKC 13.52
MIA 11.59
SAS 9.94
LAC 8.33
DEN 6.59
MEM 5.83
NYK 5.81
HOU 4.94
IND 4.67
GSW 2.88
BRK 2.57
ATL -0.63
CHI -1.03
LAL -2.37
BOS -3.31
MIL -3.65

[DSMok1]

Based on the above, and an unusually low HCA of +2.3 (found via the same regression), and giving the median prediction as my primary prediction for each series (rather than the mode), here are my predictions:
MEDIAN:.......MODE:
OKC 4 HOU 1 (4-0)
SAS 4 LAL 0 (4-0)
DEN 4 GSW 2 (4-1)
LAC 4 MEM 2 (4-1)

OKC 4 LAC 1 (4-1)
SAS 4 DEN 2 (4-1)

OKC 4 SAS 2 (4-1)

MIA 4 MIL 0 (4-0)
NYK 4 BOS 1 (4-0)
IND 4 ATL 1 (4-1)
BRK 4 CHI 2 (4-1)

MIA 4 BRK 1 (4-0)
NYK 4 IND 3 (4-3)

MIA 4 NYK 1 (4-1)

OKC 4 MIA 3 (4-2)

[v-zero]

So the results are in, the measure is RMSE of wins per team.

Bobbofitos: 2.65
v-zero: 3.14
Mike G: 4.06
DSMok1: 4.22
PD123: 4.68

For those who might prefer MAE for whatever reason, the order doesn't change.

I will perform some sims at some point to get a grip of where the theoretical limit for this error is.

[Bobbofitos]

So the results are in, the measure is RMSE of wins per team.

Bobbofitos: 2.65
v-zero: 3.14
Mike G: 4.06
DSMok1: 4.22
PD123: 4.68

For those who might prefer MAE for whatever reason, the order doesn't change.

I will perform some sims at some point to get a grip of where the theoretical limit for this error is.

If only I flipped SAS and MIA

[Crow]

Pretty strong margin of victory anyways. Congrats.

[DSMok1]

I blame Westbrook... if he wasn't there, I had Mia over SAS in the finals...

[v-zero]

Well if we're trotting out the excuses...I thought the Clippers wouldn't totally collapse under the pressure