Nicholas I Dobson
Nicholas I Dobson
Script here https://github.com/nicidob/bbgm/blob/master/predict_Composite.ipynb
like #404 but without PCA vectors. Just changing the mean and noise behind the normal generation method. script to generate is https://github.com/nicidob/bbgm/blob/master/real-progs-next.ipynb
Another pass for the same stuff as #320, except this one is mostly data-driven. All the models built were in https://github.com/nicidob/bbgm/blob/master/real-progs.ipynb linear models for progs as a function of age....
This is an alternative method of computing OVR. Instead of using massive scale player plus-minus, it uses team-level results. By looking at game results and taking a minute-averaged rating for...
A while ago, I made a statistical OVR stat (fairly straightforward regression from per36 or advanced stats to OVR). I think it may actually have a use: in an alternative...
NBA-driven regressions for a bunch of constants. Using 2019 (auto?) and 2020 (alexnoob?) rosters. Aligned real players with ratings. Filter below 1500 minutes. Linear regression to the correct ratingsArray power...
The script to generate the model is in https://github.com/nicidob/bbgm/blob/master/shooting_stats.ipynb I re-created this PR because there were some conflicts that I hopefully resolved. Plus there's a husky bug fix in here.
Yeah so this replaces the linear fatigue energy relationship with a nonlinear one. Specifically, a sigmoid. This allows players to play at nearly their peak level, until the inevitable decline....
just figured I'd place this PR here. not because it's done but for the record.
I still believe the playing time mechanic is confusing to most new players. I think this is partially due to the fact that roster order also implicitly sets playing time....