Page 7 of 13
Re: 2022-23 team win projection contest
Posted: Thu Dec 15, 2022 10:37 am
by Mike G
On avg, entry errors increased by .28 overnight.
Code: Select all
. avg err rmse r^2 avg err rmse r^2
KPel 4.94 6.31 .478 emin 6.28 7.83 .418
DRKO 5.51 6.81 .476 538E 6.63 8.76 .262
22Re 5.61 7.30 .285 TmRk 6.77 8.24 .402
ncs. 5.63 6.95 .458 4141 6.85 8.53
LEBR 5.66 6.83 .489 BIPM 6.87 8.09 .584
dtka 5.84 6.90 .514 vegas 6.87 8.37 .406
Crow 5.85 7.17 .514 538R 6.88 8.02 .419
trzu 5.92 7.32 .510 nuFi 7.19 8.55 .386
avg. 5.96 7.01 .490 MPra 7.27 8.77 .446
vzro 6.03 7.20 .427 AnBa 7.55 8.73 .420
EBPI 6.16 7.32 .440 EExp 7.61 8.93 .364
https://www.basketball-reference.com/fr ... _prob.html
Anatomy of a bad night: 10 games, 7 of which featured an overachieving team vs an under (relative to most predictions here).
Code: Select all
badbad goodbad goodgood
Det>Cha NYK>Chi Mia>OKC
Ind>GSW Den>Was
Orl>Atl LAC>Min
Sac>Tor
Por>SAS
Cle>Dal
Not everyone would agree on every good/bad, of course. Dummies 4141 and 22Re are fine with it; and the leader, mostly unscathed, has his biggest lead so far.
Re: 2022-23 team win projection contest
Posted: Fri Dec 16, 2022 1:21 am
by Crow
Fwiw, I appear to be tied for 2nd highest / best on r2.
Re: 2022-23 team win projection contest
Posted: Sun Dec 18, 2022 1:32 pm
by Mike G
Code: Select all
. avg err rmse r^2 avg err rmse r^2
KPel 4.95 6.40 .473 538E 6.61 8.81 .261
DRKO 5.54 6.81 .480 TmRk 6.77 8.19 .412
22Re 5.60 7.35 .289 538R 6.85 8.06 .417
ncs. 5.63 6.91 .469 vegas 6.85 8.32 .415
dtka 5.77 6.77 .532 4141 6.95 8.63
trzu 5.78 7.22 .523 MPra 6.99 8.56 .470
Crow 5.81 7.12 .522 BIPM 7.08 8.12 .579
LEBR 5.82 6.89 .485 nuFi 7.21 8.59 .384
avg. 5.89 6.98 .496 AnBa 7.50 8.69 .427
vzro 6.08 7.21 .430 EExp 7.64 8.93 .366
EBPI 6.08 7.34 .441 2022 8.52 9.91 .289
emin 6.22 7.80 .426
Re: 2022-23 team win projection contest
Posted: Sun Dec 18, 2022 4:06 pm
by Crow
0.21 behind 3rd for me.
Re: 2022-23 team win projection contest
Posted: Sun Dec 18, 2022 8:30 pm
by Mike G
Does 4th out of 6 somehow feel better than 6 out of 19?
Re: 2022-23 team win projection contest
Posted: Sun Dec 18, 2022 10:35 pm
by Crow
4 out of 6 individuals who directly entered here doesn't sound like that much, except that 3 are top 6.
7 out of 23 total tracked is not bad, though I am often higher. Recent gains are encouraging and notable to me. Making use of the tracking data.
Lots of time left.
Re: 2022-23 team win projection contest
Posted: Sun Dec 18, 2022 11:03 pm
by Crow
A link to the full projection table at least once (if convenient) would provide easier awareness of why / where leads and deficits are for the leaders and beyond.
Re: 2022-23 team win projection contest
Posted: Mon Dec 19, 2022 9:49 am
by Mike G
Is this readable?
Code: Select all
tm 12-18 538R 538E emin ncs EExp EBPI vegas dtka trzu vzro Crow TmRk DRKO nuFi MPra AnBa LEBR BIPM KPe
Atl 40.1 51 44 44 45 46 50 46 50 48 44 47 45 45 44 46 47 45 50 46
Bos 56.5 57 56 56 57 54 57 53 55 57 58 53 53 57 56 49 52 61 62 54
Brk 46.4 45 43 51 47 40 47 50 44 47 42 44 50 46 52 44 48 40 58 46
Cha 23.4 41 39 32 32 38 42 33 33 32 39 30 35 34 39 25 37 35 30 41
Chi 38.5 35 35 35 41 44 36 42 38 38 37 40 42 38 43 34 44 41 40 38
Cle 52.8 44 39 50 43 45 46 47 46 48 48 47 47 44 43 48 46 46 43 43
Det 27.2 23 29 27 27 28 24 29 24 24 31 23 30 24 24 24 27 25 19 26
Ind 37.9 35 30 33 34 23 32 22 31 28 32 28 24 33 28 24 23 32 32 38
Mia 38.9 51 52 48 51 51 45 48 47 50 49 50 48 46 50 52 49 50 45 46
Mil 52.3 49 49 55 48 54 48 52 52 50 52 49 53 50 54 51 52 54 63 50
tm 12-18 538R 538E emin ncs EExp EBPI vegas dtka trzu vzro Crow TmRk DRKO nuFi MPra AnBa LEBR BIPM KPe
NYK 45.8 40 42 36 43 39 36 38 43 42 43 44 38 40 41 42 40 42 38 42
Orl 31.1 23 27 28 28 23 25 27 26 28 26 28 27 29 21 36 25 30 24 31
Phl 46.3 50 49 52 52 50 56 51 53 54 51 51 50 53 49 57 51 49 49 48
Tor 40.3 50 46 40 49 47 42 47 48 50 44 52 46 46 44 50 48 49 45 47
Was 32.6 32 34 34 37 34 39 35 38 33 40 30 35 34 35 33 40 41 39 40
Dal 45.2 50 51 44 47 49 50 48 46 47 46 46 48 47 47 47 50 45 46 43
Den 44.1 54 43 51 47 53 47 52 49 49 48 53 51 47 48 58 52 48 53 48
GSW 40.6 49 57 52 44 55 49 52 47 50 47 51 51 50 53 56 52 44 43 42
Hou 29.2 19 26 24 27 22 25 23 20 23 18 27 23 30 18 20 19 23 13 27
LAC 40.6 47 43 49 45 50 42 51 45 49 45 51 52 48 53 52 51 40 50 44
tm 12-18 538R 538E emin ncs EExp EBPI vegas dtka trzu vzro Crow TmRk DRKO nuFi MPra AnBa LEBR BIPM KPe
LAL 36.9 32 34 42 39 42 41 44 43 37 43 35 44 41 47 41 44 42 40 37
Mem 51.4 52 51 47 47 51 47 48 50 48 47 52 49 48 49 53 50 45 47 46
Min 36.5 47 45 53 47 49 46 48 49 50 45 47 49 50 46 47 48 49 44 46
NOP 51.7 42 44 43 46 43 45 45 46 48 44 49 45 47 43 48 45 50 55 48
OKC 35.9 24 27 23 25 25 23 24 26 24 26 27 24 26 22 23 18 27 22 27
Phx 49.9 49 49 50 54 56 50 52 50 52 51 50 52 47 51 49 52 50 52 49
Por 44.1 38 23 40 28 35 32 40 36 37 33 33 39 36 41 37 38 33 35 37
Sac 46.8 33 32 32 37 34 37 34 37 37 32 39 34 37 31 39 37 36 39 37
SAS 21.9 30 40 31 31 24 28 22 26 25 33 25 22 29 31 19 16 29 24 31
Uta 44.8 39 46 28 35 26 42 23 32 29 39 29 24 29 31 26 29 32 30 35
Rounding to integers shouldn't have much effect, I think.
Re: 2022-23 team win projection contest
Posted: Mon Dec 19, 2022 5:00 pm
by Crow
Yes it is workable readable, thanks.
Compared to Pelton, I have a huge advantage currently on Hornets, moderate on Cavs. And Griz, Spurs and Wizards. He has huge advantage on Pacers and moderate on Heat. Nuggets, Raptors, Warriors, Jazz and currently Clippers, though that one is changing.
Re: 2022-23 team win projection contest
Posted: Thu Dec 22, 2022 11:59 am
by Mike G
A bad night for most but not all.
Code: Select all
. avg err rmse r^2 avg err rmse r^2
KPel 4.77 6.31 .48 TmRk 6.72 8.26 .40
dtka 5.68 6.73 .54 538E 6.75 8.95 .24
DRKO 5.69 6.86 .47 vegas 6.82 8.41 .40
Crow 5.71 7.17 .52 538R 6.86 8.05 .42
ncs. 5.75 6.94 .46 MPra 6.90 8.56 .47
trzu 5.82 7.23 .52 4141 6.98 8.59
LEBR 5.86 6.93 .48 BIPM 7.05 8.04 .59
avg. 5.87 7.02 .49 nuFi 7.07 8.65 .38
22Re 5.91 7.34 .29 AnBa 7.48 8.82 .41
EBPI 6.05 7.41 .43 EExp 7.64 9.08 .35
emin 6.14 7.69 .44 2022 8.38 9.94 .29
vzro 6.22 7.19 .43
Pelton's lead is perhaps the biggest ever seen here, at this point in the season.
Update Dec.24
Code: Select all
. avg err rmse r^2 avg err rmse r^2
KPel 4.80 6.11 .515 TmRk 6.60 8.26 .404
Crow 5.55 7.05 .530 538R 6.62 7.81 .447
DRKO 5.57 6.79 .483 MPra 6.64 8.49 .479
dtka 5.66 6.64 .548 538E 6.65 8.84 .258
ncs. 5.72 6.83 .479 vegas 6.67 8.40 .406
trzu 5.81 7.17 .529 4141 6.87 8.64
avg. 5.86 6.94 .502 BIPM 6.89 7.85 .609
LEBR 5.87 6.88 .486 nuFi 6.93 8.65 .378
EBPI 5.98 7.23 .455 AnBa 7.43 8.72 .423
22Re 5.98 7.29 .299 EExp 7.69 9.11 .346
vzro 6.25 7.21 .432 2022 8.07 9.82 .299
emin 6.25 7.76 .430
Re: 2022-23 team win projection contest
Posted: Thu Dec 22, 2022 4:03 pm
by Crow
I'm in 4th place now and 0.03 behind second overall by average error. I feel good about possibility of getting to 2nd.
Fwiw, Pelton's entry was not a self entry in the contest but rather a pulled in by other, outside reference. So was DARKO. Dtka is the current leader of the official contest, by average error.
By R2, I am tied for 3rd and Pelton is tied for 6th.
Re: 2022-23 team win projection contest
Posted: Thu Dec 22, 2022 11:06 pm
by Crow
1 team over 70% win rate and 7 more over 60%. Same as conclusion of last season except the top 1 is 7 percentage points lower (or 10% lower). I'd expect some increase in distribution spread.
I had 9 over 60% and none over 70%. Pelton, just 2 over 60%.
6 with 28 wins or less last season. 1 projected 6 to do that this season. Pelton just 3.
Long way from the final results.
Re: 2022-23 team win projection contest
Posted: Fri Dec 23, 2022 2:54 am
by Crow
Last season at this date there were fewer fairly strong records but more of the higher ones. Fewer poor records. Both expanded in frequency by the end of the season. If that happens, tightly regressed prediction sets will probably see increasing errors relative to less regressed ones. If the best don't get much better (and the worst worse) that could mitigate the impact of expected quantity of strong & weak teams.
Too early to say for sure but I am expecting these trends will generally help me comparatively.
Re: 2022-23 team win projection contest
Posted: Sat Dec 24, 2022 5:23 pm
by Crow
I am now the leader in the APBRmetrics contest by average error, 2nd overall. Pelton's lead over me falls by .19 in last 2 days to .75.
(I was 2nd, now currently 1st in a different, outside contest on Western Conference record rank with 55 contestants.)
Re: 2022-23 team win projection contest
Posted: Sat Dec 24, 2022 9:11 pm
by Crow
Pacers 5W-8L most recently with a pretty rough schedule in next 10 days.
Wizards continue to crash. I am still bullish on the Griz. BRef currently project 51.7 to my 52 and Pelton's 47. 2 straight wins for Raptors help.