Page 10 of 13

Re: 2022-23 team win projection contest

Posted: Wed Feb 15, 2023 5:14 pm
by nbacouchside
I've been tumbling down the standings all year. Better luck next year, I guess, lol

Re: 2022-23 team win projection contest

Posted: Wed Feb 15, 2023 10:07 pm
by Mike G
Standings, schmandings.
On Jan. 5, you were in 2nd place but 1.01 out of the lead. You're now 0.74 out of the lead.

Re: 2022-23 team win projection contest

Posted: Fri Feb 17, 2023 12:26 pm
by Mike G
At the break:

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   4.89   6.00   .64      538R   5.89   7.56   .50
KPel   4.93   6.35   .55      BIPM   6.16   7.33   .66
vzro   5.21   6.66   .54      AnBa   6.32   7.94   .52
avgA   5.24   6.23   .60     vegas   6.40   7.68   .51
emin   5.45   7.14   .53      TmRk   6.45   7.64   .50
LEBR   5.54   6.37   .58      22Re   6.48   7.74   .33
DRKO   5.60   6.68   .54      nuFi   6.54   8.11   .46
trzu   5.63   6.82   .59      4141   6.71   9.43   
EBPI   5.64   7.18   .50      EExp   6.99   8.28   .46
ncs.   5.68   6.90   .51      538E   7.31   9.01   .28
MPra   5.82   7.35   .61      2022   8.21   9.78   .33
Crow   5.85   6.82   .58               
https://www.basketball-reference.com/fr ... _prob.html
Current 'standings' as projected ^^ to the end:

Code: Select all

proj  west  over      proj  east  over
54.8   Den   5.4      56.6   Bos   0.7
50.5   Mem   2.0      54.9   Mil   3.9
45.1   Sac   9.4      52.0   Phl  -0.2
43.6   Phx  -7.4      51.9   Cle   4.9
43.1   Dal  -2.8      47.0   Brk   1.3

42.9   LAC  -4.4      45.1   NYK   3.3
42.9   NOP  -3.1      43.2   Mia  -6.0
41.9   OKC  16.8      40.8   Was   5.5
41.1   GSW  -7.3      40.0   Atl  -6.3
41.1   Min  -7.4      39.6   Tor  -7.4

40.4   Uta   8.6      37.8   Chi  -0.3
39.6   Por   5.2      35.8   Ind   4.8
37.9   LAL  -1.9      34.1   Orl   6.9
19.9   Hou  -3.4      24.5   Cha  -8.5
19.5   SAS  -8.9      22.6   Det  -3.3
"over" is the difference between "projected" and the avg of 6 APBR entries.

UPDATE Feb.27

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   5.03   5.96   .65      538R   5.98   7.51   .51
KPel   5.16   6.37   .56      MPra   5.99   7.46   .60
vzro   5.31   6.64   .55      BIPM   6.09   7.30   .66
avgA   5.37   6.30   .60      22Re   6.41   7.77   .34
LEBR   5.55   6.35   .59      AnBa   6.41   7.87   .53
emin   5.70   7.35   .51     vegas   6.56   7.78   .50
EBPI   5.74   7.11   .51      TmRk   6.59   7.73   .49
DRKO   5.76   6.76   .53      nuFi   6.63   8.09   .47
ncs.   5.81   6.96   .51      4141   6.65   9.55   
trzu   5.91   6.92   .58      EExp   7.03   8.24   .47
Crow   5.97   6.88   .57      538E   7.48   9.04   .29
UPDATE Mar.01
Pelton goes 4-0 -- losses by Dal, LAL; wins by Ind, Was -- to get back in it.

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   5.06   5.92   .65      MPra   6.02   7.39   .60
KPel   5.08   6.28   .56      BIPM   6.11   7.35   .65
vzro   5.28   6.60   .55      22Re   6.40   7.72   .34
avgA   5.45   6.26   .60      AnBa   6.49   7.89   .53
LEBR   5.54   6.25   .60      4141   6.61   9.51   
DRKO   5.84   6.73   .53      TmRk   6.66   7.75   .49
emin   5.85   7.31   .51     vegas   6.66   7.81   .50
EBPI   5.90   7.16   .50      nuFi   6.72   8.10   .46
538R   5.91   7.47   .51      EExp   7.03   8.21   .47
ncs.   5.91   6.91   .51      538E   7.48   8.97   .29
Crow   5.94   6.83   .58      2022   8.13   9.67   .34
trzu   5.96   6.89   .58               

Re: 2022-23 team win projection contest

Posted: Thu Mar 02, 2023 12:48 am
by Crow
Every game counts and affects standings / contest but these results over last 10 games are probably the most notable:

Knicks and Raptors 8W-2L
Bulls and Nets 3-7
Mavs, Clippers, Pelicans 4-6

Re: 2022-23 team win projection contest

Posted: Thu Mar 02, 2023 3:29 pm
by Mike G
Actually, OKC could lose (or win) every remaining game and not change the contest standings. Nobody picked them to win over 27 games, and they've already won 28.
The Kings are similarly above the worst-case scenario seen at b-r.com's projection generator.

Other teams that, without major lineup changes, are very unlikely to matter: Cha, Mia, Min, Phx, SAS all well below anyone's guesses; NYK, Orl pretty safely above.
Suns do have a rather major lineup addition.

Re: 2022-23 team win projection contest

Posted: Thu Mar 02, 2023 3:56 pm
by Crow
You are right that certain teams will not affect relative standings. They will still affect absolute errors, fwiw.

Re: 2022-23 team win projection contest

Posted: Fri Mar 03, 2023 2:16 pm
by Mike G

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   5.01   5.83   .67      BIPM   5.98   7.25   .66
vzro   5.17   6.48   .57      trzu   5.99   6.86   .58
KPel   5.20   6.29   .57      MPra   6.00   7.37   .61
LEBR   5.35   6.12   .62      22Re   6.39   7.70   .36
avgA   5.43   6.19   .61      AnBa   6.49   7.82   .54
DRKO   5.82   6.69   .54      TmRk   6.66   7.71   .49
emin   5.84   7.28   .52     vegas   6.66   7.76   .50
EBPI   5.86   7.07   .52      nuFi   6.70   8.01   .48
Crow   5.89   6.81   .58      4141   6.72   9.58   
ncs.   5.89   6.84   .53      EExp   6.95   8.11   .48
538R   5.91   7.47   .51      538E   7.47   8.85   .31
UPDATE Mar.07

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   4.96   5.96   .65      Crow   6.05   6.79   .59
KPel   5.16   6.48   .55      538R   6.11   7.64   .50
avgA   5.35   6.29   .61      BIPM   6.16   7.46   .64
vzro   5.42   6.78   .54      22Re   6.48   7.90   .34
LEBR   5.66   6.40   .59      AnBa   6.54   7.93   .53
emin   5.73   7.27   .52      TmRk   6.64   7.74   .50
DRKO   5.80   6.76   .54     vegas   6.65   7.77   .51
MPra   5.82   7.25   .62      nuFi   6.82   8.17   .46
EBPI   5.91   7.32   .49      4141   6.93   9.70   
trzu   5.94   6.87   .59      EExp   7.10   8.20   .48
ncs.   5.94   6.98   .52      538E   7.63   9.13   .28
Overall errors and top-to-bottom spread are largest since mid-Jan.

UPDATE Mar.11 -- it's looking like a runaway.

Code: Select all

.   avg err         rmse      
dtka   4.76      dtka   5.77
.            
.           
.            
KPel   5.07      avgA   6.13
.                LEBR   6.21
avgA   5.20      KPel   6.30
vzro   5.25         
.            
.                vzro   6.56
DRKO   5.52      DRKO   6.59
LEBR   5.53         
emin   5.57      trzu   6.76
.                Crow   6.76
EBPI   5.67      ncs.   6.78
trzu   5.76         
ncs.   5.77         
BIPM   5.79      EBPI   7.01
.                BIPM   7.10
Crow   5.96      emin   7.13
MPra   5.98         
538R   6.00         
.                MPra   7.33
.            
.                538R   7.51
AnBa   6.39         
22Re   6.42      TmRk   7.61
TmRk   6.46     vegas   7.67
vegas  6.46         
nuFi   6.55      22Re   7.78
.                AnBa   7.80
.            
.                nuFi   7.90
EExp   6.92         
4141   6.94      EExp   8.07
.            
.            
.            
.            
.            
.            
538E   7.57         
.            
.            
.                538E   9.00


Re: 2022-23 team win projection contest

Posted: Thu Mar 09, 2023 4:39 pm
by Crow
I am currently 12th on average error (ouch, could be my lowest ever) but 7th on rmse and tied for 5th on r2.

Re: 2022-23 team win projection contest

Posted: Sat Mar 11, 2023 6:54 pm
by v-zero
With that RMSE on a different planet altogether. A resounding success for the value of blending (high quality) metrics. I've had a very good season (not on this but gambling wise, which admittedly is a far more complex model) but it is always good to get a little perspective and dig into what others who have success are doing.

Re: 2022-23 team win projection contest

Posted: Sat Mar 18, 2023 9:38 am
by Mike G

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   4.93   6.00   .65      BIPM   6.03   7.47   .64
KPel   5.10   6.52   .55      MPra   6.08   7.44   .60
avgA   5.24   6.29   .61      538R   6.23   7.69   .50
vzro   5.50   6.84   .53      22Re   6.40   7.89   .34
DRKO   5.61   6.66   .55      TmRk   6.66   7.84   .49
LEBR   5.71   6.46   .59     vegas   6.66   7.91   .49
trzu   5.72   6.87   .59      AnBa   6.76   8.08   .51
EBPI   5.76   7.22   .51      nuFi   6.87   8.17   .46
ncs.   5.79   6.83   .53      EExp   6.98   8.27   .47
emin   5.79   7.36   .52      4141   7.07   9.72   
Crow   5.94   6.75   .59      538E   7.56   9.02   .30
KP was .41 off the lead just 3 days ago; now .17 off.

UPDATE Mar.21

Code: Select all

.   avg err   rmse   r^2          avg err   rmse   r^2
dtka   4.95   6.07   .64      BIPM   6.00   7.57   .63
KPel   5.32   6.58   .54      MPra   6.15   7.53   .59
avgA   5.33   6.36   .60      538R   6.27   7.73   .49
vzro   5.56   6.83   .53      22Re   6.34   7.78   .35
LEBR   5.72   6.48   .58      TmRk   6.68   7.91   .48
DRKO   5.74   6.75   .54     vegas   6.69   7.99   .48
EBPI   5.81   7.26   .50      AnBa   6.83   8.14   .51
ncs.   5.84   6.89   .53      EExp   6.88   8.29   .47
Crow   5.87   6.80   .59      nuFi   6.90   8.18   .46
trzu   5.89   6.98   .58      4141   6.97   9.69   
emin   5.93   7.46   .50      538E   7.51   8.94   .30

Re: 2022-23 team win projection contest

Posted: Tue Mar 21, 2023 9:14 pm
by Crow
6th on rmse for me, tied for 4th on r2.

Top 5 on average error in the end is fairly unlikely but some further improvement still possible..

Re: 2022-23 team win projection contest

Posted: Wed Mar 22, 2023 2:10 am
by DarkStar48
It is truly impressive how much @dtka’s CraftedNBA contest entry is so far significantly outperforming every submission at all three levels of statistical fit.

Re: 2022-23 team win projection contest

Posted: Wed Mar 22, 2023 4:26 pm
by Crow
Which, if any, media writer with analytics interest will write on the story?

Pelton, Hollinger, Partnow, Schuhmann,,,?


I've had interest in, promoted, used metrics blends for many years (including in my base for this season's contest) It should get more acknowledgement, endorsement, broader use.

I wasn't convinced that Crafted PM was the best blend possible but it is certainly very good for this season, with some mix of quality and "good fortune".

Re: 2022-23 team win projection contest

Posted: Wed Mar 22, 2023 6:56 pm
by Mike G
Luck is big.
Here's a decidedly incomplete table of entries who were entered at least twice in the last 7 years. The numbers are a sort of 'percentile', as the % of entries that finished beneath you. Middle of the pack is around 50.

Code: Select all

year  2023 2022 2021 2020 2019 2018 2017      
#      19   13   16   17   18   20   23  yrs avg
Crow   58   46   56   94   28   15   57   7   51
vegas  21   62   88   18   39   50   70   7   49
538    32   77   19   24   22   90    0   7   38
cali        69   63   88   89   65   87   6   77
trzu   53   92   38   47   94        78   6   67
shad        85   81   41   50   40   30   6   55
emin   47   38   13   65   72   95        6   55
KPel   89        31   35   78             4   58
RyRi                  59   44   60   52   4   54
sndi                  53   33   85   39   4   53
dtka   95   54   50   12                  4   53
ncs    63             29         0   65   4   39
eWin         8         0        45   35   4   22

year  2023 2022 2021 2020 2019 2018 2017 
bbst             69   71   67             3   69
kmed                       17   80   96   3   64
gold                  76   56   25        3   52
lisp        31   25   82                  3   46
AnJo                       83   10   43   3   46
lnqi                            70   78   2   74
sbs                             75   48   2   61
EExp   11        94                       2   52
EBPI   68         0                       2   34
ATCt                            30   30   2   30
GK5                             35   17   2   26
Rd11                   6    6             2    6
Caliban did not show up this year, but they were in or near top 1/3 for 6 years straight.
538 is probably now 538R or a precedent form.

Re: 2022-23 team win projection contest

Posted: Wed Mar 22, 2023 7:28 pm
by nbacouchside
Mike G wrote: Wed Mar 22, 2023 6:56 pm Luck is big.
Here's a decidedly incomplete table of entries who were entered at least twice in the last 7 years. The numbers are a sort of 'percentile', as the % of entries that finished beneath you. Middle of the pack is around 50.

Code: Select all

year  2023 2022 2021 2020 2019 2018 2017      
#      19   13   16   17   18   20   23  yrs avg
Crow   58   46   56   94   28   15   57   7   51
vegas  21   62   88   18   39   50   70   7   49
538    32   77   19   24   22   90    0   7   38
cali        69   63   88   89   65   87   6   77
trzu   53   92   38   47   94        78   6   67
shad        85   81   41   50   40   30   6   55
emin   47   38   13   65   72   95        6   55
KPel   89        31   35   78             4   58
RyRi                  59   44   60   52   4   54
sndi                  53   33   85   39   4   53
dtka   95   54   50   12                  4   53
ncs    63             29         0   65   4   39
eWin         8         0        45   35   4   22

year  2023 2022 2021 2020 2019 2018 2017 
bbst             69   71   67             3   69
kmed                       17   80   96   3   64
gold                  76   56   25        3   52
lisp        31   25   82                  3   46
AnJo                       83   10   43   3   46
lnqi                            70   78   2   74
sbs                             75   48   2   61
EExp   11        94                       2   52
EBPI   68         0                       2   34
ATCt                            30   30   2   30
GK5                             35   17   2   26
Rd11                   6    6             2    6
Caliban did not show up this year, but they were in or near top 1/3 for 6 years straight.
538 is probably now 538R or a precedent form.
Caliban the GOAT!