2024-25 team win projection contest

[knarsu3]

The explanation is here: https://x.com/knarsu3/status/1848785953768247488

I predict LEBRON (or EPM) in year Y+1 using the previous 3 seasons worth of LEBRON/EPM data + some other data like role + Age etc. I might try a time decay version eventually, tho I'd basically have something along DARKO at that point lol. But the predictive versions of LEBRON that I've done seem to do well

I have a thread here: https://x.com/knarsu3/status/1763321501766627328 where I did testing and you can see how the predictive versions of the metrics did here: https://x.com/knarsu3/status/1763610143168544944

Yeah both were predictive versions of the metrics and used the previous 3 seasons + an aging curve and some other data. They weren't the prior seasons individual metrics which probably would not do as well.

Pretty sure Darko would be winning and LEBRON is about where Vegas is this year or a tad better

I think the LEBRON or EPM over-unders cooked, altho I guess it’s like a predictive version of both more than just the metrics themselves iirc right?

So was it like, for the predictive versions

was it more so a time decay or less weight on early season type of deal or a regression using the year results seperately?

[knarsu3]

From various testing I've done over the years, I don't think 1 year tells you anything about how good a metric is because I think there's a lot of variance in how projections do from year to year (as there is in the NBA standings). I don't even think 3 years is a large enough sample. IMO to truly see how well whatever projection system is doing, you want at least like 5 years.

Of course, the easiest way to test this would be to look at the year to year correlation for entries that have been in this contest for years.

Pretty sure Darko would be winning and LEBRON is about where Vegas is this year or a tad better

Last year, of 18 entries, DARKO finished 12th and LEBRON 17th.
They were 2.85 and 3.18, respectively, from the lead at the end.

Why should we think they do much better this time around?

Two years ago I won the contest, last year I finished 10th, and this year I am currently in the lead. What I am saying is that it is certainly possible that LeBron and Darko are doing much better this year. I would love to see them included.

[Crow]

I think you meant this link:

https://www.espn.com/nba/story/_/id/443 ... -solutions

[Crow]

Current draft odds distribution
(from worst record then up):

Team 1: 14.0%
Team 2: 14.0%
Team 3: 14.0%
Team 4: 12.5%
Team 5: 10.5%
Team 6: 9.0%
Team 7: 7.5%
Team 8: 6.0%
Team 9: 4.5%
Team 10: 3.0%
Team 11: 2.0%
Team 12: 1.5%
Team 13: 1.0%
Team 14: 0.5%



Possible re-work:
Team 1: 10%
Team 2: 10%
Team 3: 10%
Team 4: 10%
Team 5: 10%
Team 6: 9%
Team 7: 8%
Team 8: 7%
Team 9: 6%
Team 10: 5%
Team 11: 4%
Team 12: 3%
Team 13: 2%
Team 14: 1%
5% chance of second drawing from top 16 teams with even chances.

Everyone later than 6th get improved chances. Less incentives to push for top 5 and jockey within top 5.

If you don't like tanking or extreme tanking, there has to be less incentive to do it.


Also could consider rules that teams can not get #1 pick with own pick in back to back years or more than 2 top 3 picks within 3 years with own pick. To reduce desirability to fight to stay at top of lottery for multiple years. If initially win a pick not allowed above, automatically moved to 7th (or later, if more than one has to be bumped down by these rules). Disincentive applied by these rules before a pick is conveyed to another.

Repeat extreme tanking is imo more offensive than one off.

If being excited about draft lottery is a good thing, give every fanbase a little hope for a surprise.

[DSMok1]

John Hollinger had an interesting article recently about the incentives to tank. He indicated that at this point it's not the appeal of the lottery or draft picks that's driving bad teams to not play their best players. It's that if you are out of playoff contention anyway, the incentive is to rest and protect your veterans that are under contract for next year from injury.

There is no incentive to play your best players or to try to win games. There is incentive to play some of your young players to get them experience and to find out what you have in those players.

[TeemoTeejay]

Never minded tanking
If ur team stinks then its a fun way to cheer and have a standings watch lol

[TeemoTeejay]

[/quote]


Makes sense, but 5 years is tough just in terms of time length lol, don’t wanna make it and have to wait till I’m married to say it’s legit

Makes sense it takes awhile to have some confidence in it though, retrodiction testing vs real world always gonna be different

[Mike G]

To remove some of the tanking incentive, just draft the 2nd round in the reverse order of the 1st round: The 30 pick also gets 31. First pick has to wait til the end.
About two good players per year are picks from 31 to 60; most are early in the round.

Code: Select all

31 - Nic Claxton
Andrew Nembhard
Bojan Bogdanović
32 - Ivica Zubac
33 - Joe Harris
Hassan Whiteside
Jalen Brunson
34 - Jae Crowder
35 - DeAndre Jordan
P.J. Tucker
Draymond Green
Herbert Jones
36 - Ersan İlyasova
Mitchell Robinson
Malcolm Brogdon
37 - Gary Trent Jr.
38 - Daniel Gafford
Spencer Dinwiddie
Chandler Parsons
39 - Jaylen Wells
Jerami Grant
Khris Middleton

41 - Nikola Jokić
43 - Isaiah Hartenstein
45 - Lou Williams
Goran Dragić
Dillon Brooks
46 - Norman Powell
Danny Green
Jordan Clarkson
47 - Paul Millsap
49 - Isaiah Joe

55 - Aaron Wiggins
Patty Mills
56 - Amir Johnson
57 - Trayce Jackson-Davis
Marcin Gortat
60 - Sam Merrill
Isaiah Thomas

39 from the last 20 years' drafts.
https://www.basketball-reference.com/dr ... _2013.html

Code: Select all

.  avg err   rmse   r^2         avg err   rmse   r^2
TmTj   7.13   9.2   .48      vegas  7.95  10.0   .41
dtka   7.13   9.1   .48      WShr   7.96   9.7   .41
avgA   7.32   9.4   .44      Crow   7.97   9.9   .44
medi   7.43  10.3   .40      emin   8.06  10.2   .40
bpmW   7.50   9.3   .46      24pr   8.12  10.4   .32
DQin   7.55   9.7   .44      24py   8.55  11.5   .32
eExp   7.78  10.0   .46      eWin   8.89  11.0   .23
KPel   7.79   9.2   .58      perW   9.06  11.1   .22
.                            bmgm   9.09  12.0   .32

[TeemoTeejay]

Peltons R^2 is nuts

[Mike G]

This was the Jan. 31 installment:

Code: Select all

.  avg err   rmse   r^2      .  avg err   rmse   r^2
TmTj   5.77   8.1   .57      WShr   6.73   8.8   .47
avgA   6.11   8.3   .54      medi   6.73   9.2   .48
dtka   6.22   8.1   .56      emin   6.91   9.2   .47
bpmW   6.30   8.3   .53      Crow   6.91   8.9   .52
vegas  6.37   8.6   .53      24pr   7.02   9.8   .35
DQin   6.45   8.6   .53      eWin   7.47   9.9   .33
KPel   6.49   8.0   .55      perW   7.85  10.2   .30
eExp   6.63   8.7   .57      24py   8.08  10.9   .35
.                            bmgm   8.64  11.2   .37

Pelton was 4th in r^2 then, and 9 were over .50

UPDATE Mar. 26

Code: Select all

.  avg err   rmse   r^2          avg err   rmse   r^2
dtka   7.17   9.2   .48     vegas   7.97   10.1   .41
TmTj   7.19   9.3   .48      WShr   8.00    9.7   .40
avgA   7.33   9.5   .44      Crow   8.01   10.0   .44
medi   7.49  10.4   .40      emin   8.11   10.3   .40
bpmW   7.57   9.4   .45      24pr   8.13   10.4   .32
DQin   7.62   9.8   .44      24py   8.56   11.5   .32
KPel   7.79   9.3   .57      eWin   8.96   11.1   .23
eExp   7.82  10.1   .46      bmgm   9.09   12.0   .32
.                            perW   9.15   11.2   .22

[Mike G]

The Pacers have joined the "outsiders" -- teams whose Win projection is above or below all our predictions -- bringing the # of such teams back up to 17. It had dropped to 13 at some point.

Code: Select all

w  Atl   w  Bos   w  Brk   w  Cha   w  Chi   w  Cle   w  Det   w  Ind
44  pW   62  em   37  KP   34  pW   37 b-r   66 b-r   45 b-r   48 b-r
41  WS   61  DQ   34  WS   33  em   35  WS   52  dt   33  eW   48  pW
41  dt   61 b-r   31  eW   31  Cr   35  bW   50  mr   33  pW   47  mr
41  bW   61  mr   30  pW   31  eW   34  eW   51  KP   32  TT   47  Cr
40 b-r   60  dt   29  bW   30  DQ   34  pW   50  Cr   32  bW   47  KP
39  DQ   59  Cr   29  DQ   30  WS   32  KP   49  em   32  WS   46  dt
39  eW   57  TT   26 b-r   29  KP   31  TT   48  DQ   29  KP   44  em
38  mr   57  WS   26  dt   29  dt   30  dt   48  TT   28  em   44  TT
37  TT   56  bW   26  mr   28  bW   29  Cr   46  WS   28  dt   43  bW
35  Cr   52  pW   24  TT   26  TT   28  DQ   46  bW   27  Cr   43  WS
34  em   52  KP   23  em   24  mr   27  em   45  eW   26  DQ   42  DQ
32  KP   52  eW   18  Cr   21 b-r   25  mr   44  pW   26  mr   41  eW
                                             
w  Mia   w  Mil   w  NYK   w  Orl   w  Phl   w  Tor   w  Was      
47  em   50  dt   58  em   49  Cr   55  mr   34  KP   34  eW      
45  KP   49  DQ   53  Cr   48  mr   53  eW   32  pW   33  pW      
45  eW   48  em   51 b-r   46  dt   52  pW   32  WS   28  WS      
43  bW   48  Cr   50  TT   44  WS   51  TT   32  em   28  bW      
43  dt   47  KP   50  dt   44  KP   47  DQ   32  bW   24  KP      
43  WS   46 b-r   48  mr   43  DQ   47  dt   30 b-r   21  dt      
42  mr   45  pW   45  KP   43  TT   46  WS   30  eW   19  em      
42  DQ   45  TT   44  WS   42  em   46  Cr   30  dt   19  Cr      
42  TT   44  mr   43  DQ   41  bW   46  bW   29  mr   18 b-r      
41  pW   44  bW   43  bW   39  eW   45  KP   29  DQ   18  mr      
40  Cr   44  WS   41  eW   39 b-r   43  em   29  TT   15  TT      
37 b-r   43  eW   38  pW   34  pW   27 b-r   25  Cr   14  DQ      
                                             
w  Dal   w  Den   w  GSW   w  Hou   w  LAC   w  LAL   w  Mem   w  Min
56  DQ   53  em   52  Cr   53 b-r   47 b-r   48 b-r   51  KP   57  em
53  pW   51  mr   49  DQ   47  eW   45  DQ   45  TT   49 b-r   55  DQ
52  eW   51  dt   49  TT   46  mr   44  WS   44  Cr   49  Cr   53  eW
52  TT   51 b-r   48  mr   45  bW   40  bW   44  dt   48  em   52  mr
51  Cr   51  eW   47  em   44  TT   40  dt   43  DQ   45  DQ   51  Cr
50  KP   49  pW   47 b-r   44  Cr   36  KP   42  em   43  dt   50  WS
50  bW   49  DQ   46  bW   44  WS   36  TT   41  KP   43  eW   50  bW
47  WS   49  Cr   45  dt   43  pW   35  mr   39  mr   43  mr   50  dt
47  dt   47  KP   45  KP   40  KP   35  pW   39  bW   42  pW   48  TT
47  mr   47  WS   44  WS   38  dt   33  Cr   38  WS   41  bW   47 b-r
43  em   47  bW   41  eW   36  DQ   32  em   37  pW   41  TT   46  pW
40 b-r   46  TT   40  pW   33  em   31  eW   36  eW   34  WS   43  KP
                                             
w  NOP   w  OKC   w  Phx   w  Por   w  Sac   w  SAS   w  Uta      
49  mr   69 b-r   51  TT   36 b-r   50  Cr   38  em   33  pW      
49  em   62  mr   50  mr   33  KP   49  KP   37  pW   32  eW      
48  Cr   61  DQ   49  dt   31  eW   49  bW   36  eW   32  WS      
45  DQ   59  TT   48  Cr   28  pW   48  pW   36  bW   31  mr      
44  bW   58  em   48  pW   27  bW   47  eW   35 b-r   29  KP      
44  pW   55  WS   47  bW   26  Cr   47  DQ   34  WS   29  DQ      
44  TT   54  bW   47  WS   26  WS   47  TT   34  TT   29  dt      
44  WS   54  Cr   46  em   26  DQ   47  dt   34  KP   29  bW      
44  eW   54  dt   45  KP   24  TT   46  WS   31  DQ   28  em      
42  dt   54  KP   45  eW   24  mr   44  em   31  dt   27  TT      
42  KP   52  pW   43  DQ   23  em   42  mr   31  mr   26  Cr      
23 b-r   49  eW   38 b-r   22  dt   41 b-r   29  Cr   18 b-r      

[Crow]

Second round successes are often bigs, shooters, defenders. PGs slightly under-represented. High usage guys even lower.

[Mike G]

The leader has his worst error of the season, as does nearly everyone.

Code: Select all

.  avg err   rmse   r^2      .  avg err    rmse   r^2
dtka   7.25   9.3   .47      WShr   8.09    9.8   .41
TmTj   7.27   9.5   .47      emin   8.10   10.4   .40
avgA   7.46   9.6   .43     vegas   8.11   10.2   .44
medi   7.51  10.6   .39      Crow   8.13   10.1   .40
bpmW   7.66   9.4   .45      24pr   8.17   10.5   .32
DQin   7.80   9.9   .43      24py   8.54   11.6   .32
eExp   7.88  10.2   .45      eWin   9.01   11.2   .23
KPel   7.97   9.5   .58      bmgm   9.14   12.1   .32
.                            perW   9.23   11.4   .22

Last year's leader posted their worst observed error on Apr. 1, 2024
They are now just .25 from the lead.
A curious separation of powers:

Code: Select all

e = .50 -         e = .78          e = 1.22         e = 1.98         e = 5.0 +
5.71   medi      6.67   dtka      7.776  TmTj      9.30   dtka      12.4   KPel
5.86   dtka      6.74   TmTj      7.779  dtka      9.41   TmTj      12.9   dtka
6.05   TmTj      6.74   medi      8.09   bpmW      9.41   bpmW      13.2   bpmW
6.46   eExp      7.22   bpmW      8.23   medi      9.42   KPel      13.3   WShr
6.58   DQin      7.28   DQin      8.30   DQin      9.80   WShr      14.1   TmTj
6.63   bpmW      7.28   eExp      8.33   KPel      9.90   DQin      14.1   DQin
6.72   emin      7.51   emin      8.43   eExp     10.11   Crow      14.5   Crow
6.92   Crow      7.59   KPel      8.50   WShr     10.16   eExp      14.6   eExp

Both the top and bottom end increase their lead as e goes further from 1.
Others are at max separation from the pack.

UPDATE Mar. 30

Code: Select all

.  avg err   rmse   r^2      .   avg err   rmse   r^2
TmTj   7.23   9.5   .47      WShr   8.09    9.8   .41
dtka   7.25   9.4   .47      emin   8.10   10.4   .40
avgA   7.43   9.6   .44     vegas   8.11   10.2   .44
medi   7.47  10.6   .39      Crow   8.13   10.1   .40
bpmW   7.67   9.5   .45      24pr   8.17   10.5   .32
DQin   7.74   9.9   .44      24py   8.54   11.6   .32
eExp   7.92  10.2   .45      eWin   9.01   11.2   .23
KPel   7.99   9.5   .58      bmgm   9.14   12.1   .32
.                            perW   9.23   11.4   .22

UPDATE Mar. 31 -- almost no change in the standings, but medi has dislodged AvgA from the #3 spot.

Code: Select all

.  avg err   rmse   r^2       .  avg err   rmse   r^2
TmTj   7.26   9.5   .47      emin   8.15   10.5   .40
dtka   7.29   9.4   .47      24pr   8.17   10.6   .33
medi   7.49  10.6   .40      Crow   8.18   10.2   .43
avgA   7.53   9.7   .44      egas   8.23   10.3   .40
bpmW   7.65   9.5   .45      24py   8.58   11.6   .33
DQin   7.80  10.0   .44      eWin   9.11   11.4   .23
eExp   7.99  10.3   .45      bmgm   9.21   12.1   .33
WShr   8.02   9.9   .41      perW   9.32   11.5   .21
KPel   8.07   9.6   .58               

If we'd left TmTj with their original submission, they'd be in 5th place.
If I'd kept my original slate (pre Pelton minutes) eWins would be middle of the pack. I shall persist in my tanking strategy.

[dtkavana]

It is my opinion that TmTj should be banned from the competition. Who let them in here anyways?!?

[Mike G]

How about, should they win, there will be the inglorious asterisk next to their achievement.