My personal revised projections:
Atlanta Hawks 43
Boston Celtics 46
Brooklyn Nets 18
Charlotte Hornets 30
Chicago Bulls 45
Cleveland Cavaliers 53
Dallas Mavericks 30
Denver Nuggets 58
Detroit Pistons 51
Golden State Warriors 45
Houston Rockets 58
Indiana Pacers 23
Los Angeles Clippers 35
Los Angeles Lakers 45
Memphis Grizzlies 30
Miami Heat 45
Milwaukee Bucks 41
Minnesota Timberwolves 49
New Orleans Pelicans 25
New York Knicks 52
Oklahoma City Thunder 71
Orlando Magic 45
Philadelphia 76ers 46
Phoenix Suns 38
Portland Trail Blazers 45
Sacramento Kings 28
San Antonio Spurs 43
Toronto Raptors 46
Utah Jazz 26
Washington Wizards 20
1230
2025-26 team win projection contest
Re: 2025-26 team win projection contest
EExp and DARKO have owned the top spot all month, neither for more than 4 days straight.
Top to bottom diff is season-low.
https://www.basketball-reference.com/fr ... _prob.html
Cross-posted from "Improved" thread; players who have over- or under-produced (in eWins) over last season's rates, by .20 or more.
If you wonder why a team is over- or under-achieving, some answers are here.
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
EExp 6.57 8.15 .47 Mgoo 7.27 8.83 .40
DRKO 6.60 8.06 .48 eWin 7.39 9.12 .36
vegas 6.87 8.15 .52 medi 7.55 9.22 .42
bpmW 7.04 8.37 .45 Crow 7.70 9.49 .37
knar 7.12 8.37 .43 WShr 7.72 9.36 .34
avgA 7.13 8.62 .42 KPel 7.83 9.66 .35
dtka 7.17 8.65 .41 25pr 8.11 9.68 .30
Walr 7.25 8.40 .44 DQin 8.78 11.21 .25
perW 7.26 9.64 .31
https://www.basketball-reference.com/fr ... _prob.html
Cross-posted from "Improved" thread; players who have over- or under-produced (in eWins) over last season's rates, by .20 or more.
Code: Select all
eW+ Atlanta Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.76 Jalen Johnson .651 25.5 10.5 5.9 2.30 .562 15.9 10.4 4.1 1.37
.31 Nickeil A-Walker .563 19.6 3.3 3.5 1.03 Min .573 12.1 4.5 3.2 .61
-.22 Mouhamed Gueye .565 10.5 7.3 2.3 .58 .513 10.1 9.9 1.5 1.05
-.23 Luke Kennard .637 11.9 3.2 2.1 .30 Mem .627 12.8 4.4 4.0 .70
-.29 Dyson Daniels .494 8.3 6.3 5.5 .74 .540 12.1 6.5 3.7 1.05
eW+ Boston Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.57 Jaylen Brown .578 30.9 6.8 3.9 2.15 .547 22.6 6.4 4.4 1.53
.40 Neemias Queta .610 13.5 12.9 2.2 1.53 .664 13.3 9.3 1.6 .97
.27 Jordan Walsh .561 7.6 8.8 1.6 .79 .461 5.2 5.6 1.4 .03
-.22 Anfernee Simons .543 16.3 2.9 2.7 .72 Por .552 18.6 3.1 4.5 1.03
-.32 Sam Hauser .464 7.5 5.2 1.7 .21 .631 14.3 5.2 1.3 .72
eW+ Brooklyn Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.44 Michael Porter .590 22.7 9.6 3.1 1.66 Den .610 18.3 7.8 1.8 1.11
.36 Nic Claxton .631 17.3 9.5 4.0 1.66 .561 12.3 10.6 2.6 1.20
-.27 Ziaire Williams .556 10.7 3.9 .5 .26 .543 12.5 7.1 1.7 .77
eW+ Charlotte Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.27 Moussa Diabaté .677 14.1 12.4 1.2 1.34 .596 9.9 11.9 1.3 .90
-.23 Tre Mann .455 10.8 5.2 3.9 .35 .534 16.1 4.0 3.4 .72
eW+ Chicago Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.35 Josh Giddey .548 20.0 9.9 8.2 1.90 .562 15.1 9.3 6.9 1.42
.22 Kevin Huerter .590 15.8 5.2 3.0 .92 .542 11.8 4.3 2.6 .59
-.22 Nikola Vučević .595 16.8 11.2 3.3 1.40 .606 19.6 11.2 3.2 1.68
eW+ Cleveland Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.40 Donovan Mitchell .634 32.3 5.2 5.0 2.32 .567 25.8 5.2 4.9 1.84
.21 Jaylon Tyson .632 13.8 5.8 2.1 .88 .519 10.2 6.8 2.7 .53
-.26 Evan Mobley .546 17.9 10.2 3.8 1.65 .622 22.3 11.0 3.2 1.94
-.34 Larry Nance .404 4.8 7.1 2.4 .06 Atl .628 13.2 7.7 2.2 .97
eW+ Dallas Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.26 Max Christie .634 14.9 4.3 2.7 .72 .562 11.1 4.3 2.0 .43
-.25 Klay Thompson .470 11.8 4.5 2.6 .38 .550 16.1 4.5 2.2 .81
eW+ Denver Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
1.00 Nikola Jokić .754 38.2 14.0 10.1 4.07 .651 29.2 13.0 8.0 2.89
.44 Aaron Gordon .676 26.5 7.2 1.6 1.71 .638 18.3 6.4 3.3 1.07
.40 Jamal Murray .566 23.2 5.5 5.8 1.72 .577 19.3 4.0 4.8 1.21
.21 Tim Hardaway .673 18.0 2.7 1.9 .79 Det .561 12.7 3.3 1.7 .43
-.66 Cameron Johnson .459 8.6 4.0 2.4 .13 Brk .622 20.9 5.3 3.4 1.32
eW+ Detroit Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.38 Jalen Duren .685 27.1 14.7 1.8 2.46 .688 17.6 15.6 3.1 1.89
eW+ Golden State Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.34 Jimmy Butler .646 24.0 6.5 5.1 2.04 .608 20.2 6.3 5.4 1.65
.22 Stephen Curry .622 32.2 4.5 3.9 2.22 .610 27.4 5.3 6.0 1.92
.21 Moses Moody .631 17.9 4.0 1.9 1.02 .571 14.2 4.2 1.8 .69
-.21 Buddy Hield .556 11.5 3.1 2.7 .38 .557 15.1 5.1 2.2 .78
-.28 Al Horford .458 5.7 5.9 2.4 .23 Bos .560 11.2 8.2 2.5 .84
-.29 Draymond Green .484 8.1 7.7 6.7 .62 .526 9.8 8.0 6.2 .98
eW+ Houston Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.41 Reed Sheppard .637 18.3 4.1 3.8 1.20 .464 9.1 4.3 3.2 .49
-.22 Amen Thompson .525 16.3 6.8 4.6 1.19 .591 15.4 10.0 3.6 1.44
-.23 Tari Eason .640 16.3 8.1 2.0 .99 .561 15.6 9.8 1.7 1.42
eW+ Indiana Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
-.20 Aaron Nesmith .511 14.0 5.2 1.5 .70 .646 17.2 5.7 1.4 .98
-.27 RayJ Dennis .430 8.3 3.5 3.9 .12 .441 9.6 5.3 5.1 .94
-.28 Jay Huff .477 7.9 6.4 1.4 .50 Mem .659 18.6 5.8 1.3 1.19
-.33 Jarace Walker .402 7.9 6.4 3.1 .24 .584 11.6 6.3 2.5 .66
-.34 Tony Bradley .641 11.7 7.3 1.4 .59 .635 16.5 11.6 1.4 1.47
eW+ LA Clippers Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.27 James Harden .630 26.3 6.7 7.7 2.07 .569 22.6 6.1 7.8 1.77
-.20 Nicolas Batum .471 4.8 5.2 1.2 -.01 .630 7.8 5.5 1.8 .41
-.24 Bradley Beal .471 11.5 1.6 2.6 -.05 Phx .591 17.1 3.7 3.4 .91
-.26 Brook Lopez .558 10.9 4.9 1.1 .32 Mil .617 14.5 5.5 1.7 .94
-.30 Ivica Zubac .644 18.5 13.0 2.2 1.56 .632 19.4 14.4 2.6 1.92
-.35 Kris Dunn .560 8.8 3.7 2.6 .20 .529 8.6 5.2 3.7 .71
-.57 John Collins .582 14.5 6.6 .8 .56 Uta .618 19.9 10.0 1.9 1.40
eW+ LA Lakers Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.69 Austin Reaves .623 28.4 5.6 7.0 2.19 .605 20.4 4.8 5.1 1.35
.56 Luka Dončić .599 32.5 9.5 7.4 2.94 .576 26.5 8.4 6.6 2.21
eW+ Memphis Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
-.21 Jaylen Wells .460 9.6 5.1 1.8 .24 .553 12.6 5.0 1.9 .54
-.36 Ja Morant .468 17.5 4.4 8.2 1.14 .553 24.2 5.1 7.1 1.66
-.53 Jaren Jackson .581 19.5 6.5 1.5 1.09 .581 24.4 7.1 2.0 1.73
eW+ Miami Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.40 Jaime Jaquez .579 17.6 7.6 4.6 1.37 .535 12.8 7.2 3.6 .87
.31 Norman Powell .651 30.1 4.5 2.2 1.84 LAC .606 24.5 3.6 2.0 1.36
.30 Davion Mitchell .568 10.9 3.2 7.2 .80 .568 9.5 2.9 5.3 .44
.24 Pelle Larsson .662 16.5 3.7 3.5 .82 .537 9.5 3.9 2.4 .41
.22 Si. Fontecchio .690 18.3 4.0 1.3 .81 Det .537 10.0 6.0 1.4 .37
-.22 Andrew Wiggins .586 17.3 5.1 2.5 .97 .557 19.5 5.5 2.7 1.22
eW+ Milwaukee Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.40 Giannis A .646 35.7 12.9 6.1 3.34 .609 31.7 12.1 5.8 2.85
.39 Ryan Rollins .630 19.2 4.6 5.0 1.33 .595 14.0 4.3 3.7 .89
-.21 Gary Trent .534 10.8 2.2 2.2 .30 .586 13.8 2.9 1.3 .57
-.22 Amir Coffey .512 3.8 3.4 1.3 -.35 LAC .606 13.7 3.2 1.3 .50
-.44 Bobby Portis .558 12.5 9.0 1.5 .64 .538 16.2 10.8 2.4 1.38
eW+ Minnesota Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.83 Julius Randle .646 28.5 8.1 5.7 2.38 .582 20.6 8.3 4.7 1.46
.29 Jaden McDaniels .637 20.0 5.3 2.2 1.26 .557 13.0 6.9 2.0 .92
.28 Anthony Edwards .610 31.9 5.6 3.7 2.29 .585 26.9 5.9 4.0 1.79
-.20 Naz Reid .538 12.7 7.6 2.3 .91 .572 16.9 7.9 2.5 1.22
-.31 Rudy Gobert .707 11.9 11.3 1.3 1.02 .670 14.4 12.5 1.7 1.38
eW+ New Orleans Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
-.31 Jordan Poole .526 15.9 2.1 3.2 .50 Was .582 21.8 3.5 4.4 1.21
eW+ New York Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.48 Mikal Bridges .607 15.2 4.6 4.4 1.36 .581 16.0 3.4 3.0 .80
.22 Landry Shamet .678 15.8 3.1 1.3 .68 .598 11.5 2.8 1.0 .25
.20 Jalen Brunson .580 27.6 3.7 6.0 1.93 .593 25.5 3.2 6.3 1.67
eW+ Oklahoma City Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.72 Ajay Mitchell .555 21.3 5.0 4.4 1.47 .581 12.9 3.6 3.1 .60
.58 Shai G-A .639 41.7 5.7 6.7 3.48 .624 37.2 5.2 5.9 2.88
.25 Chet Holmgren .682 28.2 10.5 2.1 2.13 .587 20.1 10.5 2.3 1.73
.24 Is. Hartenstein .687 19.6 14.8 3.6 1.94 .591 14.8 13.7 4.4 1.65
-.21 Ousmane Dieng .392 5.0 5.1 1.3 -.22 .534 10.4 6.3 2.1 .56
-.34 Luguentz Dort .421 6.0 6.0 1.7 .01 .584 12.4 5.0 1.8 .68
-.35 Jaylin Williams .470 7.4 8.3 3.8 .60 .589 11.9 11.0 4.5 1.29
eW+ Orlando Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.28 Tr. Da Silva .622 16.1 5.7 1.3 .91 .518 10.3 5.9 2.2 .49
.23 Jalen Suggs .687 22.9 6.7 6.5 1.78 .529 18.9 5.7 4.3 1.29
-.20 Goga Bitadze .722 11.6 9.7 1.9 1.12 .619 13.0 12.8 3.1 1.54
-.24 Franz Wagner .573 22.4 6.6 3.5 1.60 .550 24.6 6.8 4.6 1.83
-.36 Tyus Jones .286 1.9 2.3 3.1 -.22 Phx .583 12.2 3.2 5.8 .76
-.55 Desmond Bane .532 16.5 5.0 4.4 .90 Mem .594 20.1 7.1 4.9 1.49
eW+ Philadelphia Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.75 Tyrese Maxey .586 27.2 4.8 6.0 2.14 .554 21.1 3.3 4.8 1.39
-.24 Jabari Walker .481 6.0 8.5 .9 .15 Por .592 12.3 9.4 1.3 .91
-.25 Adem Bona .464 3.4 8.0 .5 .27 .696 12.5 9.4 .8 .94
eW+ Phoenix Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.66 Grayson Allen .632 20.8 3.7 4.2 1.41 .623 13.9 4.2 2.5 .64
.58 Devin Booker .602 28.6 4.4 6.1 1.91 .579 22.0 3.9 5.6 1.35
.41 Dillon Brooks .581 24.8 4.2 2.0 1.54 Hou .551 14.7 4.7 1.7 .71
.22 Ryan Dunn .558 11.0 8.0 2.3 .86 .515 9.9 6.7 1.2 .54
-.26 Mark Williams .643 15.8 12.6 1.1 1.58 Cha .635 20.5 14.1 2.8 1.99
eW+ Portland Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.41 Jrue Holiday .560 15.5 5.9 7.3 1.36 Bos .561 12.9 5.3 4.2 .87
.32 Jerami Grant .617 20.0 3.7 2.3 1.07 .515 13.2 4.0 2.0 .64
.31 Deni Avdija .620 25.5 7.2 4.1 1.74 .592 18.8 8.9 3.9 1.40
.31 Shaedon Sharpe .527 22.5 6.9 2.1 1.40 .545 18.3 5.3 2.6 1.00
-.31 Toumani Camara .518 9.8 5.7 2.5 .42 .566 11.1 6.6 2.1 .75
eW+ Sacramento Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
-.20 Dennis Schröder .491 10.7 4.4 5.3 .62 Det .524 14.2 3.5 5.8 .84
-.21 Zach LaVine .648 21.4 3.8 1.8 .97 .631 22.8 4.3 3.5 1.22
-.23 Keon Ellis .619 9.7 2.7 .8 .26 .657 11.6 4.0 1.7 .69
-.27 DeMar DeRozan .575 16.8 3.6 2.8 .91 .559 19.4 4.1 3.6 1.19
-.28 Malik Monk .581 14.4 2.1 2.1 .53 .543 16.4 4.5 5.1 1.03
-.45 Do. Sabonis .554 14.9 13.8 3.1 1.39 .644 19.5 15.3 5.2 1.98
eW+ San Antonio Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.62 Stephon Castle .570 19.4 7.0 7.4 1.57 .513 15.6 4.7 4.4 .85
.48 Vi. Wembanyama .597 26.8 14.3 3.6 2.74 .586 23.5 11.5 3.2 2.18
-.29 Devin Vassell .525 13.1 4.1 2.3 .58 .548 15.9 4.4 2.7 .91
eW+ Toronto Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.45 Scottie Barnes .584 20.5 8.8 4.8 1.99 .516 17.5 8.6 5.2 1.48
.23 Brandon Ingram .574 20.6 6.6 3.5 1.38 NOP .555 20.0 6.1 4.5 1.12
-.20 Ochai Agbaji .414 3.9 4.8 1.2 -.15 .598 12.6 5.0 1.6 .62
eW+ Utah Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.76 Lauri Markkanen .610 27.4 6.1 1.8 1.81 .563 18.0 7.1 1.4 1.03
.49 Keyonte George .554 20.2 4.1 6.2 1.27 .530 14.8 4.4 5.0 .73
eW+ Washingoton Eff% Sco Reb Ast e484 Eff% Sco Reb Ast e484
.55 Alex Sarr .572 18.6 9.4 3.4 1.65 .477 12.6 8.4 2.6 .94
.49 Kyshawn George .623 17.1 6.6 4.0 1.06 .498 8.6 5.3 2.6 .42
-.23 CJ McCollum .525 15.2 3.7 3.0 .74 NOP .544 18.8 4.2 3.5 1.03
-.29 Bub Carrington .451 5.6 4.7 4.2 -.02 .510 9.0 4.7 4.2 .39
-.30 Cam Whitmore .532 12.4 3.9 .8 .39 Hou .543 17.3 6.7 1.7 1.04
-.39 Khris Middleton .553 11.4 5.3 2.6 .48 .580 16.9 5.5 5.2 1.25
Re: 2025-26 team win projection contest
UPDATE Nov. 19 -- not much change in the order, and overall avg's continue to worsen.
Some exponent leaders:
Some exponent leaders:
Code: Select all
e = .25- e = .65 e = 1.46 e = 2.86 e = 14
4.80 perW 6.06 EExp 7.61 DRKO 9.47 vegas 14.2 knar
5.09 EExp 6.23 perW 7.65 vegas 9.54 knar 15.1 DRKO
5.41 vegas 6.23 DRKO 7.65 EExp 9.54 EExp 15.1 dtka
5.46 dtka 6.31 vegas 7.92 bpmW 9.54 DRKO 15.5 bpmW
5.48 DRKO 6.51 bpmW 7.96 knar 9.71 Walr 15.7 EExp
5.56 knar 6.54 dtka 8.10 Walr 9.72 bpmW 15.9 vegas
5.69 bpmW 6.56 knar 8.10 avgA 9.97 dtka 16.2 Walr
Re: 2025-26 team win projection contest
Could you walk thru the interpretation and meaning of this data and are exponents more than .2 or .4 from 1 at all "realistic"?
Re: 2025-26 team win projection contest
Low exponents minimize the large errors and optimize direct hits.
Big exponents exaggerate larger errors and don't care about near misses.
Exponent of 2.0 is also known as Root Mean Squared Error, and is included in the regular updates. I'm not sure how this is more predictive or meaningful than Mean Absolute Error (e = 1.0), but it's a lot harder to articulate to the average person.
My intuitive feeling is that .50 < e < 2.00 are more meaningful than anything outside those endpoints. I just find it interesting how different leaderboards are generated along the continuum.
Sometimes, like now, there are several "leaders", and often there are just 2. I don't recall just one entry owning all exponents.
Big exponents exaggerate larger errors and don't care about near misses.
Exponent of 2.0 is also known as Root Mean Squared Error, and is included in the regular updates. I'm not sure how this is more predictive or meaningful than Mean Absolute Error (e = 1.0), but it's a lot harder to articulate to the average person.
My intuitive feeling is that .50 < e < 2.00 are more meaningful than anything outside those endpoints. I just find it interesting how different leaderboards are generated along the continuum.
Sometimes, like now, there are several "leaders", and often there are just 2. I don't recall just one entry owning all exponents.
Re: 2025-26 team win projection contest
Again a lead change after 3 days. Smallest top-bottom spread so far.
I recall a contestant saying (in the last year or two) they designed their predictions for a good RMSE (e=2), though I don't see how one would do that. Perhaps just be conservative?
The tendency over the season seems to be that the winner at e=1 is also best at e=2. Sometimes it's a split winner.
It's also been suggested to average the two, or to use an intermediate exponent -- 1.414...?
The intermediate could be a 3rd "winner"
UPDATE Nov. 21 -- 4 games incl. Orl>LAC by 28 and Mem>Sac by 41.
These teams' projections just changed by 3 to 4 wins, and some teams that didn't even play by almost a win: Having played LAC once and Sac twice, Denver's SOS now looks notably lighter.
Overall largest errors of the month.
UPDATE Nov. 22
Teams have played 14 to 18 games (avg 15.7), and I am fairly confident that these current b-r.com projections will be better predictors of final Win totals than any pre-season prediction.
Relative to the avg of our guesses:If Miami wins >45 games, these first 16 games are more indicative than what we thought in mid October.
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 6.83 8.36 .47 Walr 7.60 8.73 .43
EExp 6.84 8.44 .46 eWin 7.62 9.33 .36
vegas 6.93 8.33 .52 Crow 7.78 9.72 .36
bpmW 7.12 8.62 .44 WShr 7.81 9.52 .34
avgA 7.28 8.87 .41 medi 7.83 9.58 .40
dtka 7.30 8.93 .40 KPel 7.94 9.86 .35
perW 7.30 9.87 .30 25pr 8.27 9.79 .31
knar 7.34 8.61 .43 DQin 8.82 11.37 .25
Mgoo 7.40 9.03 .40
The tendency over the season seems to be that the winner at e=1 is also best at e=2. Sometimes it's a split winner.
It's also been suggested to average the two, or to use an intermediate exponent -- 1.414...?
Code: Select all
exponent leader
< .65 PER Wins
.65-.95 ESPN Experts
.97-1.7 DARKO
1.7-3.3 vegas
> 3.3 knarsuUPDATE Nov. 21 -- 4 games incl. Orl>LAC by 28 and Mem>Sac by 41.
These teams' projections just changed by 3 to 4 wins, and some teams that didn't even play by almost a win: Having played LAC once and Sac twice, Denver's SOS now looks notably lighter.
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 6.88 8.55 .46 Walr 7.66 8.92 .42
EExp 6.89 8.61 .45 eWin 7.72 9.52 .35
vegas 7.12 8.48 .51 medi 7.91 9.66 .40
bpmW 7.18 8.88 .43 Crow 7.92 9.92 .35
perW 7.30 10.09 .29 WShr 7.94 9.75 .33
avgA 7.37 9.09 .40 KPel 8.05 10.13 .33
knar 7.44 8.84 .42 25pr 8.47 10.05 .29
Mgoo 7.51 9.27 .39 DQin 9.06 11.55 .24
dtka 7.54 9.21 .39
UPDATE Nov. 22
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 6.80 8.66 .46 Walr 7.49 8.86 .43
EExp 6.97 8.61 .46 eWin 7.52 9.41 .37
bpmW 7.06 8.79 .45 medi 7.88 9.78 .40
vegas 7.11 8.42 .52 WShr 8.00 9.73 .34
perW 7.21 9.99 .31 Crow 8.00 10.11 .34
knar 7.28 8.85 .43 KPel 8.06 10.14 .34
avgA 7.29 9.09 .41 25pr 8.59 10.17 .29
Mgoo 7.47 9.17 .41 DQin 9.07 11.56 .25
dtka 7.48 9.25 .39
Relative to the avg of our guesses:
Code: Select all
over west avgA proj. over east avgA proj
10.2 Phx 33.9 44.1 15.7 Mia 37.3 53.0
8.3 SAS 42.5 50.8 14.0 Tor 36.8 50.8
8.0 Hou 49.7 57.7 10.1 Det 42.8 52.9
5.4 Uta 24.0 29.4 8.7 Phl 37.2 45.9
5.1 Den 50.6 55.7 7.0 Chi 35.4 42.4
4.4 Por 36.1 40.5 2.3 Atl 43.4 45.7
4.0 LAL 44.0 48.0 1.9 Bos 43.1 45.0
-0.7 Min 46.7 46.0 1.9 NYK 48.3 50.2
-2.5 OKC 62.7 60.2 1.5 Cha 30.2 31.7
-7.0 GSW 48.3 41.3 0.5 Brk 25.7 26.2
-8.7 Mem 42.8 34.1 -1.5 Orl 48.9 47.4
-10.5 NOP 32.8 22.3 -3.1 Mil 41.2 38.1
-13.4 Sac 38.2 24.8 -3.2 Cle 53.0 49.8
-16.9 LAC 49.3 32.4 -3.8 Was 20.5 16.7
-18.7 Dal 45.6 26.9 -19.4 Ind 39.3 19.9
Re: 2025-26 team win projection contest
CraftedNBA projection from yesterday:
EASTERN CONFERENCE
Cavaliers 52-30
Knicks 51-31
Pistons 50-32
Heat 50-32
Magic 48-34
Celtics 47-35
Raptors 46-36
Hawks 46-36
Bulls 40-42
76ers 38-44
Bucks 33-49
Pacers 25-57
Hornets 25-57
Nets 22-60
Wizards 15-67
WESTERN CONFERENCE
Thunder 68-14
Rockets 59-23
Nuggets 55-27
T-Wolves 53-29
Warriors 50-32
Lakers 48-34
Suns 41-41
Clippers 41-41
Spurs 41-41
Blazers 37-45
Grizzlies 31-51
Mavericks 26-56
Jazz 22-60
Kings 22-60
Pelicans 22-60
EASTERN CONFERENCE
Cavaliers 52-30
Knicks 51-31
Pistons 50-32
Heat 50-32
Magic 48-34
Celtics 47-35
Raptors 46-36
Hawks 46-36
Bulls 40-42
76ers 38-44
Bucks 33-49
Pacers 25-57
Hornets 25-57
Nets 22-60
Wizards 15-67
WESTERN CONFERENCE
Thunder 68-14
Rockets 59-23
Nuggets 55-27
T-Wolves 53-29
Warriors 50-32
Lakers 48-34
Suns 41-41
Clippers 41-41
Spurs 41-41
Blazers 37-45
Grizzlies 31-51
Mavericks 26-56
Jazz 22-60
Kings 22-60
Pelicans 22-60
Re: 2025-26 team win projection contest
We'll see better in longer run and in the end how everything did. From pre-season and from this time period.
Overall and on Thunder. My revised projection of 71 wins. Will it prove too high or too low?
BRef with Heat at 53. Closest?
Spurs 51 or 41 or where in-between?
Most differences are moderate or less.
Overall and on Thunder. My revised projection of 71 wins. Will it prove too high or too low?
BRef with Heat at 53. Closest?
Spurs 51 or 41 or where in-between?
Most differences are moderate or less.
Re: 2025-26 team win projection contest
Overnight the average improvement is .10 -- ranging from .21 (KPel) to -.09 (perW, worse)
Our avgA has crept to within .22 of EExp, down from .57 on Tuesday.
Nov. 24 -- Here's the playoff outlook at this point of the season, according to b-r.com
Columns: current SRS, chance of reaching the finals, chance of winning the title, and the ratio -- chance of winning if they get there.
https://www.basketball-reference.com/fr ... _prob.html
total conference SRS at Finals = sum of (SRS*reach%)
Minor anomalies in the East: NY and Atl with less chance (W/R) than some weaker teams.
The West has just caught up with the East, and both are projecting to avg 41.0 wins per team.
East has 7 of the 11 best teams (SRS>4) and only 4 of the 9 worst (< -4)
Our avgA has crept to within .22 of EExp, down from .57 on Tuesday.
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 6.76 8.58 .46 eWin 7.44 9.24 .39
EExp 6.88 8.51 .47 Walr 7.48 8.84 .44
vegas 6.93 8.32 .53 medi 7.71 9.70 .40
bpmW 7.00 8.67 .46 Crow 7.84 10.04 .34
avgA 7.10 8.98 .42 WShr 7.84 9.61 .35
knar 7.18 8.69 .45 KPel 7.85 10.02 .35
perW 7.30 9.93 .32 25pr 8.47 10.03 .31
Mgoo 7.33 9.03 .42 DQin 8.92 11.48 .25
dtka 7.37 9.11 .41 Columns: current SRS, chance of reaching the finals, chance of winning the title, and the ratio -- chance of winning if they get there.
Code: Select all
west SRS reach win W/R east SRS reach win W/R
OKC 12.1 .480 .373 .78 Mia 7.7 .272 .106 .39
Hou 11.2 .292 .203 .70 Det 5.7 .149 .046 .31
Den 7.6 .121 .071 .59 Cle 5.3 .128 .037 .29
SAS 5.0 .050 .025 .50 Tor 5.1 .119 .035 .29
Min 2.8 .019 .007 .37 NYK 5.1 .100 .026 .26
Phx 2.6 .018 .007 .39 Orl 4.5 .086 .024 .28
LAL 1.9 .014 .005 .36 Bos 4.2 .066 .018 .27
GSW -.1 .004 .001 .25 Atl 2.7 .040 .007 .18
Por -.9 .003 ---- --- Phl 2.5 .035 .008 .23
tot: Chi -.8 .006 .001 .17
west 10.4 1.00 .69 east 5.6 1.00 .31 total conference SRS at Finals = sum of (SRS*reach%)
Minor anomalies in the East: NY and Atl with less chance (W/R) than some weaker teams.
The West has just caught up with the East, and both are projecting to avg 41.0 wins per team.
East has 7 of the 11 best teams (SRS>4) and only 4 of the 9 worst (< -4)
Re: 2025-26 team win projection contest
BRef's win projection for Thunder up 8 wins in 2 weeks. Method gets better with more data but early method proving to be too regressed there.... and still, imo.
Up to 7 teams projected with 50+ wins, 8 below 30 wins. The top number might increase. The bottom may decline, depending on degree of tanking and injury management and not trying really hard and development and poor lineup management.
Up to 7 teams projected with 50+ wins, 8 below 30 wins. The top number might increase. The bottom may decline, depending on degree of tanking and injury management and not trying really hard and development and poor lineup management.
Re: 2025-26 team win projection contest
OKC has actually gotten a lot better of late, so it's not just more data or less regression by b-r.com
Plugging in the current SRS of their opponents, it's seen that they've played just one team from the top 12, Hou in game 1.
The game SRS column is just that game's margin of victory + opponent SRS.
Cumulative gSRS is divided by games played to get season SRS to date.
Final column is pythagorean expected wins in 82 games, ignoring current W-L.They now have about the same SRS as last year, when they won 68.
Last year, no team had a SOS more than +/-0.84 . Minus-5 at 22% of the season is remarkable.
Plugging in the current SRS of their opponents, it's seen that they've played just one team from the top 12, Hou in game 1.
The game SRS column is just that game's margin of victory + opponent SRS.
Cumulative gSRS is divided by games played to get season SRS to date.
Final column is pythagorean expected wins in 82 games, ignoring current W-L.
Code: Select all
opp mov opSRS gSRS SRS pW*82
Hou 1 11.23 12.2 12.2 68.4
Ind 6 -12.18 -6.2 3.0 49.4
Atl 17 2.69 19.7 8.6 62.3
Dal 7 -7.58 -0.6 6.3 57.5
Sac 6 -8.61 -2.6 4.5 53.3
Was 19 -14.54 4.5 4.5 53.3
NOP 31 -10.87 20.1 6.7 58.5
LAC 19 -4.03 15.0 7.8 60.7
Por -2 -0.86 -2.9 6.6 58.2
Sac 31 -8.61 22.4 8.2 61.5
Mem 14 -3.24 10.8 8.4 62.0
GSW 24 -0.14 23.9 9.7 64.4
LAL 29 1.87 30.9 11.3 67.1
Cha 13 -5.10 7.9 11.1 66.7
NOP 17 -10.87 6.1 10.7 66.1
Sac 14 -8.61 5.4 10.4 65.6
Uta 32 -6.40 25.6 11.3 67.0
Por 27 -0.86 26.1 12.1 68.3
avg 16.9 -4.82 12.1
Last year, no team had a SOS more than +/-0.84 . Minus-5 at 22% of the season is remarkable.
Re: 2025-26 team win projection contest
Change in relative SRS performance is a factor, but wins are wins and the ultimate answer.
For a team that has significant turnover, last season data may not be that useful; but for almost identical Thunder, last season is quite informative. Current season stats also without Jalen Williams. Two big reasons a better projection (more likely closer to final total) is higher than the early regressed BRef projection.
.
"They now have about the same SRS as last year, when they won 68." Without JDub. And current BRef projection is... 62.
On average regression conservatism may be appropriate, but it looks bad here imo
As much as projection target accuracy may be worth checking / discussing, the answer for team wins will eventually be the final actual.
"Last year, no team had a SOS more than +/-0.84 . Minus-5 at 22% of the season is remarkable." How much this was happenstance and how much was league narrative design? The SOS is extreme tail and even 50% beyond number 2, I think narrative design may be a substantial influence.
Wonder of that is a common approach with last season champs. Quick impression is early schedule was not easy for 2024-25 Celtics or 2023-24 Nuggets but maybe for 2022-23 Warriors. Happenstance or design? Don't know. Maybe only do it for dynasty candidates? It mostly evens out, but early momentum is valuable for confidence and narrative
For a team that has significant turnover, last season data may not be that useful; but for almost identical Thunder, last season is quite informative. Current season stats also without Jalen Williams. Two big reasons a better projection (more likely closer to final total) is higher than the early regressed BRef projection.
.
"They now have about the same SRS as last year, when they won 68." Without JDub. And current BRef projection is... 62.
On average regression conservatism may be appropriate, but it looks bad here imo
As much as projection target accuracy may be worth checking / discussing, the answer for team wins will eventually be the final actual.
"Last year, no team had a SOS more than +/-0.84 . Minus-5 at 22% of the season is remarkable." How much this was happenstance and how much was league narrative design? The SOS is extreme tail and even 50% beyond number 2, I think narrative design may be a substantial influence.
Wonder of that is a common approach with last season champs. Quick impression is early schedule was not easy for 2024-25 Celtics or 2023-24 Nuggets but maybe for 2022-23 Warriors. Happenstance or design? Don't know. Maybe only do it for dynasty candidates? It mostly evens out, but early momentum is valuable for confidence and narrative
Re: 2025-26 team win projection contest
Wizards have won 6% of games so far. BRef is projecting they win 24% of the remainder.
How strong a candidate for worst record in league history?
How strong a candidate for worst record in league history?
Re: 2025-26 team win projection contest
Overall best numbers in a week, least top-to-bottom all year.
In 3 days, vegas has gone from .31 off the lead to just .12
In that same time, avgA has done the same vs EExp.
DARKO with a season-high 6 consecutive days on top.
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 6.66 8.51 .48 Walr 7.33 8.68 .46
vegas 6.78 8.18 .55 eWin 7.45 9.17 .40
bpmW 6.82 8.47 .49 WShr 7.59 9.49 .37
EExp 6.83 8.46 .49 medi 7.62 9.66 .41
avgA 6.95 8.85 .44 KPel 7.68 9.81 .37
knar 7.08 8.58 .47 Crow 7.78 9.95 .36
perW 7.12 9.81 .34 25pr 8.34 9.95 .32
Mgoo 7.21 8.94 .44 DQin 8.62 11.21 .28
dtka 7.24 9.01 .42
In that same time, avgA has done the same vs EExp.
DARKO with a season-high 6 consecutive days on top.