2025-26 team win projection contest

[Crow]

A simple display of just Darko, BPM and Vegas would be useful.Where do they differ most?

[Mike G]

After 28 games (54 to go) OKC is 25-3, having lost 2 of their last 3.
If they continue losing 2/3 of games, they'll still win 43 games: 54/3 = 18 , on top of 25 = 43
This isn't a valid prediction, but we might consider the recent results to weigh more than very early games.

So here are a set of projections, using the Win% of their last 3, 6, 9, etc games as if it might recur in their remaining games.

Code: Select all

last  W  L   Win%   W+  +25
_3    1  2   .333   18   43
_6    4  2   .667   36   61
_9    7  2   .778   42   67
12   10  2   .833   45   70
15   13  2   .867   47   72
18   16  2   .889   48   73
21   18  3   .857   46   71
24   21  3   .875   47   72
27   24  3   .889   48   73

The avg of that final column is 67. Drop the first and last, and it's 69.5

[DQuinn1575]

8 other teams have started 25-3 with an average year-end total of 66.4 wins, with a low of 61 (10 SAS) and high of 72, with 4 titles.
So 60+ looks real safe.
We do have 3 champ teams since the merger at 16-12 - 04 DET, 06 MIA, 21 MIL - so still hope for a lot of teams.

[Crow]

2020-21 Bucks appear to have had a strong net margin at that point, even after losing 4 straight near mid-February.

Pistons near neutral.

Heat quite negative but Shaq had barely played in first 20 games. He wasn't the most important player but he was somewhat helpful.

[Mike G]

Leader board:

Code: Select all

. Atl       Bos       Brk       Cha       Chi
47  Veg   52  B-R   29  B-R   33  Wal   36  Wal
44  bpm   46  DKO   28  bpm   31  DKO   35  bpm
43  DKO   42  bpm   28  DKO   31  bpm   35  DKO
42  Wal   41  Wal   26  Wal   31  B-R   35  B-R
39  B-R   41  Veg   21  Veg   27  Veg   32  Veg
                           
. Cle       Det       Ind       Mia       Mil
57  DKO   56  B-R   39  Wal   46  B-R   43  Wal
57  Veg   46  Veg   39  Veg   40  bpm   43  Veg
53  Wal   44  DKO   36  bpm   39  Veg   42  DKO
53  bpm   43  bpm   36  DKO   38  Wal   38  bpm
42  B-R   43  Wal   22  B-R   34  DKO   33  B-R
                           
. NYK       Orl       Phl       Tor       Was
55  B-R   52  Veg   44  B-R   43  B-R   25  DKO
54  Veg   50  DKO   43  Veg   40  DKO   24  Wal
52  Wal   49  B-R   37  DKO   39  Veg   21  Veg
51  DKO   49  Wal   35  bpm   38  bpm   19  bpm
47  bpm   48  bpm   34  Wal   38  Wal   18  B-R
    
                       
. Dal       Den       GSW       Hou       LAC
45  bpm   56  B-R   49  bpm   56  B-R   50  DKO
44  DKO   54  Veg   47  Wal   54  Veg   49  bpm
41  Veg   52  DKO   47  Veg   52  DKO   48  Veg
39  Wal   51  bpm   45  B-R   50  bpm   46  Wal
32  B-R   50  Wal   44  DKO   50  Wal   30  B-R
                           
. LAL       Mem       Min       NOP       OKC
49  Veg   43  Wal   50  Veg   34  Wal   67  Wal
47  B-R   43  bpm   49  B-R   33  bpm   66  B-R
46  bpm   41  Veg   47  bpm   32  DKO   65  bpm
44  Wal   38  DKO   46  Wal   31  Veg   63  Veg
39  DKO   38  B-R   45  DKO   28  B-R   60  DKO
                           
. Phx       Por       Sac       SAS       Uta
43  B-R   38  bpm   39  bpm   54  B-R   29  B-R
35  bpm   37  Wal   36  Wal   45  DKO   28  Wal
32  Wal   37  B-R   35  DKO   45  Veg   28  DKO
33  DKO   36  DKO   35  Veg   42  Wal   24  bpm
32  Veg   34  Veg   25  B-R   42  bpm   19  Veg

Walrus has sat at #3 or 4 for the last 3 weeks, and closer to these others than to anyone below.

[Crow]

Thanks. That is easier to review.

Darko and BPM with close to same average errors but there are differences. An average difference of just over 2.5 wins assigned. Bigger differences with Cavs, Heat, Bucks, Wizards, Warriors, Thunder, Lakers, Griz, Jazz, Kings.

[Mike G]

10-day low for the field. LAC >> Hou was huge; also Dal > Den

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.27   7.91   .53      perW   7.09   9.15   .39
bpmW   6.34   8.00   .52      eWin   7.18   8.64   .44
vegas  6.43   8.00   .56      dtka   7.27   8.73   .44
Walr   6.48   8.01   .52      Crow   7.27   9.07   .43
knar   6.55   7.94   .52      25pr   7.39   9.44   .36
avgA   6.74   8.27   .49      medi   7.62   9.21   .44
Mgoo   6.82   8.42   .48      KPel   7.93   9.75   .36
WShr   6.98   8.71   .44      DQin   8.62  10.57   .33
EExp   7.03   8.36   .48         

Now 13 of 16 entries are within 1.00 of the lead.
Our avgA with largest yet lead over EExp.

Exponent leaders:

Code: Select all

e = .25 --        e = 1.64         e = 2.78 
4.38   bpmW      7.39   DRKO      8.83   knar
4.48   DRKO      7.48   knar      8.84   DRKO
4.86   Walr      7.48   bpmW      8.91   Walr
4.93  vegas      7.49  vegas      8.91  vegas
5.10   perW      7.52   Walr      8.95   bpmW
5.18   Crow      7.88   Mgoo      9.19   EExp
5.24   knar      7.91   EExp      9.46   Mgoo
                     
e = 3.98          e = 6.9          e = 20 ++ 
9.94   DRKO      11.7  vegas      14.4   bpmW
9.97  vegas      11.8   Walr      14.6   EExp
9.97   knar      11.8   DRKO      14.7  vegas
9.98   Walr      11.8   bpmW      15.0   Walr
10.07  bpmW      11.9   EExp      15.0   DRKO
10.21  EExp      12.0   knar      15.5   Mgoo
10.72  Mgoo      12.7   Mgoo      15.6   eWin

DARKO leads from .50 to 5.0, except for the interval 2.5 to 3.2 (knarsu)
Vegas is consigned to the 5.0 - 8.5 range.
BPM rules the high and low ends; both are tenuous.

[Crow]

Current BRef guesses on winning a title:

Thunder 53%
6 contenders, about 42%
9 longshots, about 5%.

Lakers in the longshot category, given 0.1% chance.

[Mike G]

In 2 weeks, knarsu has surged from #8 to #4, from .75 off the lead to just .23 back.

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.33   7.98   .52      perW   7.21   9.24   .38
bpmW   6.52   8.16   .51      eWin   7.27   8.78   .43
vegas  6.52   8.11   .55      Crow   7.31   9.17   .42
knar   6.56   8.06   .51      dtka   7.36   8.88   .42
Walr   6.59   8.16   .50      25pr   7.41   9.57   .35
avgA   6.76   8.40   .48      medi   7.68   9.33   .43
Mgoo   6.85   8.55   .46      KPel   8.00   9.85   .36
WShr   7.05   8.88   .42      DQin   8.66  10.68   .32
EExp   7.12   8.47   .47      

Actually a couple of others have improved even more, relative to #1.
On Dec. 12 everyone but medi had their best/lowest errors of the season. And since then, the pack has significantly tightened.
Ranked by improvement in how far off the lead everyone is in this interval:

Code: Select all

impr  off:  12-12 12-26
.89   25pr   1.96  1.07
.62   WShr   1.34   .72
.41   knar    .64   .23
.39   medi   1.73  1.34
.37   Crow   1.35   .98
.21   avgA    .63   .42
.20   bpmW    .38   .19
.19   DRKO    .19   .00
.09   Walr    .35   .26
.05   dtka   1.08  1.03
.03   Mgoo    .55   .52
.00   KPel   1.67  1.67
.00   perW    .88   .88
-.09  EExp    .70   .79
-.18  DQin   2.15  2.33
-.19  vegas   .00   .19
-.25  eWin    .69   .94

The dummy at top -- last year's Pythagorean 'wins' regressed toward .500 -- is the only one with lower error than 12-12.

[Crow]

Worst BRef sim for Heat was apparently 33 wins. Will want to see actual and projections in 6 weeks and final results.

[Crow]

Need another 2-4 weeks like last 2.

[Mike G]

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.43   7.92   .52      EExp   7.10   8.47   .45
knar   6.49   7.94   .51      perW   7.16   9.17   .38
Walr   6.49   8.05   .50      eWin   7.19   8.75   .42
bpmW   6.50   8.07   .50      25pr   7.28   9.37   .35
vegas  6.53   8.17   .53      dtka   7.30   8.76   .42
avgA   6.77   8.33   .47      Crow   7.35   9.16   .42
Mgoo   6.84   8.52   .45      medi   7.85   9.39   .42
WShr   6.97   8.69   .43      KPel   8.09   9.88   .34
.                             DQin   8.73  10.54   .32

Lowest error of the year for 25pr. Early this month, they were .08 from the bottom.
Meanwhile, five are within .10 of the top.

UPDATE Dec. 28

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.52   8.03   .50      EExp   7.24   8.60   .43
knar   6.56   8.00   .50      perW   7.26   9.35   .35
bpmW   6.65   8.20   .48      eWin   7.29   8.84   .40
Walr   6.65   8.15   .48      dtka   7.46   8.87   .40
vegas  6.69   8.39   .51      Crow   7.51   9.38   .39
avgA   6.90   8.48   .45      25pr   7.53   9.42   .34
Mgoo   6.94   8.67   .43      medi   8.09   9.61   .39
WShr   7.10   8.79   .41      KPel   8.24  10.04   .32
.                             DQin   8.84  10.73   .30

Logjam at the top and exponent scramble:

Code: Select all

from - to     leader
..   < 0.45   BPM
0.46 - 0.59   vegas
0.60 - 0.63   knarsu
0.64 - 1.78   DARKO
1.80 - 9.9    knarsu
10.0 - 69     BPM
70   < ...    DARKO

UPDATE Dec. 31

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.36   7.88   .52      EExp   7.17   8.44   .45
bpmW   6.55   8.12   .49      eWin   7.28   8.76   .42
vegas  6.61   8.20   .53      perW   7.29   9.41   .35
knar   6.62   7.87   .52      dtka   7.35   8.68   .43
Walr   6.62   8.03   .50      Crow   7.46   9.34   .39
.                             25pr   7.50   9.35   .35
avgA   6.92   8.40   .46      medi   8.08   9.53   .40
Mgoo   6.94   8.59   .44      KPel   8.34  10.02   .33
WShr   7.00   8.77   .42      DQin   8.76  10.67   .31

Empty line to note the separation between the top 5 and everyone else. KPel with season worst.
While DARKO currently enjoys a .19 lead, the next 4 are within a good day of closing that gap.
Here they are relative to the b-r.com projections of 15 teams presently in contention:

Code: Select all

W  Brk      Cha       Chi       NYK       Orl
31  kna   34  kna   36  Wal   55  B-R   52  Veg
31  B-R   33  Wal   36  kna   54  Veg   51  kna
28  bpm   32  B-R   35  bpm   52  Wal   50  DKO
28  DKO   31  DKO   35  DKO   51  DKO   49  Wal
26  Wal   31  bpm   34  B-R   48  kna   48  bpm
21  Veg   27  Veg   32  Veg   47  bpm   48  B-R
                           
. Phl       Was       Den       GSW       Hou
43  Veg   25  DKO   54  B-R   49  bpm   56  B-R
42  B-R   24  Wal   54  Veg   48  kna   54  Veg
37  DKO   23  kna   52  DKO   47  Wal   52  DKO
35  bpm   21  Veg   51  bpm   47  Veg   50  kna
34  Wal   20  B-R   50  Wal   44  B-R   50  bpm
34  kna   19  bpm   45  kna   44  DKO   50  Wal
                           
. LAL       Mem       Min       OKC       Por
49  Veg   43  Wal   50  Veg   67  Wal   38  bpm
46  bpm   43  bpm   49  B-R   65  bpm   37  Wal
45  B-R   41  Veg   47  bpm   64  B-R   37  B-R
44  Wal   41  kna   47  kna   63  kna   36  kna
41  kna   38  B-R   46  Wal   63  Veg   36  DKO
39  DKO   38  DKO   45  DKO   60  DKO   34  Veg

All else being equal, a single team could rise or fall enough to dramatically change these configurations. All else being equal, should the Nets collapse or the Warriors surge, or -- most probable -- the Nuggets sink:

Code: Select all

Brk=21 (31)      Den=44 (54)      GSW=49 (44)
vegas  6.26      knar   6.34      bpmW   6.40
DRKO   6.43      DRKO   6.57      DRKO   6.51
Walr   6.60      bpmW   6.68      knar   6.56
bpmW   6.67      Walr   6.72      Walr   6.61
Mgoo   6.85      vegas  6.92      vegas  6.62
knar   6.97                  

In parentheses are current projections. And in each case, the selected teams don't have to change by this much to promote these top entries.

[Crow]

Leader board:

Code: Select all

. Atl       Bos       Brk       Cha       Chi
47  Veg   52  B-R   29  B-R   33  Wal   36  Wal
44  bpm   46  DKO   28  bpm   31  DKO   35  bpm
43  DKO   42  bpm   28  DKO   31  bpm   35  DKO
42  Wal   41  Wal   26  Wal   31  B-R   35  B-R
39  B-R   41  Veg   21  Veg   27  Veg   32  Veg
                           
. Cle       Det       Ind       Mia       Mil
57  DKO   56  B-R   39  Wal   46  B-R   43  Wal
57  Veg   46  Veg   39  Veg   40  bpm   43  Veg
53  Wal   44  DKO   36  bpm   39  Veg   42  DKO
53  bpm   43  bpm   36  DKO   38  Wal   38  bpm
42  B-R   43  Wal   22  B-R   34  DKO   33  B-R
                           
. NYK       Orl       Phl       Tor       Was
55  B-R   52  Veg   44  B-R   43  B-R   25  DKO
54  Veg   50  DKO   43  Veg   40  DKO   24  Wal
52  Wal   49  B-R   37  DKO   39  Veg   21  Veg
51  DKO   49  Wal   35  bpm   38  bpm   19  bpm
47  bpm   48  bpm   34  Wal   38  Wal   18  B-R
    
                       
. Dal       Den       GSW       Hou       LAC
45  bpm   56  B-R   49  bpm   56  B-R   50  DKO
44  DKO   54  Veg   47  Wal   54  Veg   49  bpm
41  Veg   52  DKO   47  Veg   52  DKO   48  Veg
39  Wal   51  bpm   45  B-R   50  bpm   46  Wal
32  B-R   50  Wal   44  DKO   50  Wal   30  B-R
                           
. LAL       Mem       Min       NOP       OKC
49  Veg   43  Wal   50  Veg   34  Wal   67  Wal
47  B-R   43  bpm   49  B-R   33  bpm   66  B-R
46  bpm   41  Veg   47  bpm   32  DKO   65  bpm
44  Wal   38  DKO   46  Wal   31  Veg   63  Veg
39  DKO   38  B-R   45  DKO   28  B-R   60  DKO
                           
. Phx       Por       Sac       SAS       Uta
43  B-R   38  bpm   39  bpm   54  B-R   29  B-R
35  bpm   37  Wal   36  Wal   45  DKO   28  Wal
32  Wal   37  B-R   35  DKO   45  Veg   28  DKO
33  DKO   36  DKO   35  Veg   42  Wal   24  bpm
32  Veg   34  Veg   25  B-R   42  bpm   19  Veg

BRef rolling projection at this date at one of the extremes in this group in 22 of 30 cases. How will things look at the end? Will one or more of the contest projections beat this version of BRef's time advantaged projection? Will the average of the 4? Time will eventually tell.

It could be a split decision and the split will be interesting.

[Mike G]

WinShares and knarsu with their smallest errors of the season; DARKO lead is biggest for anyone since Dec. 10.

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.19   7.78   .53      EExp   7.04   8.35   .47
bpmW   6.44   8.07   .50      perW   7.16   9.33   .36
Walr   6.46   7.97   .51      eWin   7.18   8.72   .42
knar   6.46   7.84   .52      dtka   7.22   8.62   .44
vegas  6.59   8.11   .54      Crow   7.28   9.30   .40
avgA   6.72   8.35   .47      25pr   7.40   9.34   .36
Mgoo   6.84   8.54   .45      medi   8.07   9.45   .41
WShr   6.91   8.71   .43      KPel   8.21  10.00   .33
EExp   7.04   8.35   .47      DQin   8.64  10.62   .32

I just discovered the betting line on BETMGM, another "vegas" thing. Last year they were in the contest as bmgm; I thought at first they could not be serious, but in the last few weeks they got off the floor and did better than PER & eWins.
This year, they would be among the leaders and currently .10 better than the Vegas we've been seeing.
And this is after scaling their predictions downward, as their total sums to 1234 (vs 1230).

Update on the over- and under-achievers, relative to our averages.

Code: Select all

over    tm   avgA   proj     over   tm    avgA   proj
12.3   SAS   42.5   54.8     12.0   Det   42.8   54.8
10.7   Phx   33.9   44.6     10.7   Mia   37.3   48.0
6.8    Uta   24.0   30.8      8.5   Bos   43.1   51.6
6.0    Hou   49.7   55.7      7.4   Tor   36.8   44.2
3.2    Den   50.6   53.8      6.4   NYK   48.3   54.7
2.3    OKC   62.7   65.0      5.8   Brk   25.7   31.5
1.1    Min   46.7   47.8      5.2   Phl   37.2   42.4
0.7    LAL   44.0   44.7      1.6   Cha   29.9   31.5
0.0    Por   36.1   36.1      0.1   Was   20.5   20.6
-3.7   GSW   48.3   44.6     -1.0   Chi   35.7   34.7
-4.5   Mem   42.8   38.3     -1.2   Orl   48.9   47.7
-8.4   NOP   32.8   24.4     -4.3   Atl   43.4   39.1
-12.9  LAC   49.3   36.4     -7.7   Cle   53.0   45.3
-14.3  Dal   45.6   31.3     -8.2   Mil   41.2   33.0
-15.5  Sac   38.2   22.7    -19.3   Ind   39.3   20.0

"We" expected average wins to favor the West by 43-39, and it's now right around 42-40.

UPDATE Jan. 3 -- some shuffling, Walrus in the 2 spot.

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.19   7.85   .53      dtka   7.11   8.60   .45
Walr   6.32   8.04   .51      perW   7.17   9.45   .35
bpmW   6.44   8.14   .50      eWin   7.21   8.85   .42
knar   6.59   7.94   .52      25pr   7.36   9.37   .36
vegas  6.68   8.07   .55      Crow   7.40   9.34   .40
avgA   6.77   8.40   .47      medi   8.06   9.43   .42
WShr   6.93   8.80   .43      KPel   8.19  10.04   .33
Mgoo   6.93   8.64   .45      DQin   8.50  10.57   .33
EExp   6.95   8.33   .48               

We now have clear separations in the upper ranks.

[Mike G]

https://www.basketball-reference.com/fr ... _prob.html
As of now, the most likely playoff seedings in the west look like this:

Code: Select all

rk   tm   prob
1   OKC   .994
2   Hou   .488
3   SAS   .315
4   Den   .314
5   Min   .340
6   LAL   .229
7   Phx   .243
8   GSW   .199

The chance that all 8 spots fill out this way -- just multiply all 8 fractions -- comes to .00018, or one in 5536.
Switch the 2 and 3, SAS and Hou, the odds are 1:10,000+
Chance that happens and Min/Den switch spots is 1/56,000 ... others remaining the same.

UPDATE Jan. 5 -- a new entry in the contest, bmgm, was in it last year, mostly dwelling at the bottom. They are another vegas-like entity; the one formerly known as 'vegas' is now called Ebet.
The field worsened by an avg of .13 overnight, and more scrambling ensues.

Code: Select all

.   avg err   rmse   r^2      .   avg err   rmse   r^2
DRKO   6.41   8.07   .50      Mgoo   7.12   8.82   .42
bpmW   6.56   8.34   .47      dtka   7.27   8.80   .42
Walr   6.60   8.31   .47      eWin   7.30   8.97   .40
bmgm   6.71   8.31   .52      25pr   7.38   9.50   .34
knar   6.72   8.11   .49      perW   7.41   9.66   .33
Ebet   6.89   8.33   .52      Crow   7.59   9.50   .38
WShr   6.97   8.96   .40      medi   8.16   9.62   .40
avgA   7.02   8.61   .44      KPel   8.48  10.22   .31
EExp   7.02   8.50   .45      DQin   8.80  10.79   .30

Our avgA has lost its lead over EExp. KPel again with worst error to date; medi tied for their worst.