You may be pleased to know that, while drko leads at exponents 1 and 2 and everything above; you have the lead at e<.15.
This is due to 4 direct hits -- on Bos, Brk, NYK, Den -- and a couple other near-best guesses.
It's an unstable position though. Whether these teams win or lose their next game, you've lost them as zero-errors.
UPDATE Nov 12
Here's the contest mixing everyone from within and without, plus 2 more 'dummies':
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 4.59 5.54 .50 EBPI 5.95 6.93 .31
LEBR 5.11 6.27 .37 4141 6.09 7.68
dtka 5.13 6.64 .37 vzro 6.14 7.44 .35
KPel 5.16 6.61 .32 ChKl 6.22 7.87 .30
ncs. 5.36 6.61 .36 eWin 6.38 7.75 .08
trzu 5.41 6.69 .35 vegas 6.39 7.52 .29
avgA 5.45 6.68 .36 emin 6.75 8.20 .33
Crow 5.51 7.01 .32 EExp 6.76 7.95 .28
23py 5.51 7.12 .18 IanL 6.94 8.50 .28
23re 5.68 7.03 .20 DQin 7.03 8.71 .17
medi 5.89 7.03 .46 NuFi 7.22 9.61 .16
23re is last year's win total regressed halfway to 41.
23re = (W+41)/2
23py is last year's Pythagorean "should have won" based on point differential; also averaged with 41.
UPDATE Nov.13 -- Darko's biggest lead yet.
Code: Select all
. avg err rmse r^2 . avg err rmse r^2
DRKO 4.57 5.65 .52 23re 6.03 7.22 .22
dtka 5.16 6.71 .38 vzro 6.19 7.42 .37
trzu 5.22 6.73 .38 ChKl 6.25 7.81 .33
KPel 5.31 6.74 .34 vegas 6.32 7.61 .30
LEBR 5.32 6.45 .38 4141 6.59 8.05
ncs. 5.38 6.73 .37 EExp 6.67 7.99 .30
avgA 5.40 6.76 .38 eWin 6.79 8.06 .08
Crow 5.53 7.09 .34 IanL 6.83 8.47 .30
medi 5.75 6.82 .49 emin 6.97 8.26 .34
23py 5.87 7.28 .20 DQin 7.13 8.86 .18
EBPI 5.97 7.09 .32 NuFi 7.42 9.78 .16