APBRmetrics

from
http://godismyjudgeok.com/DStats/


To-Do List
1.Salary and contract value discussions and charts
2.Multi-year APM/RAPM with aging incorporated
3.Revise ASPM based on multi-year RAPM with aging
4.ASPM within-year stability/cross validation
5.Historical ASPM Tableau visualizations
6.Create Excel VBA recursive web scraping tutorial
7.Comparison of residual exponents for rankings
8.Comparison of various "value metrics" ability to "explain" wins
9.Publication of spreadsheets used
10.Work on using Bayesian priors in Adjusted +/-
11.Work on K-Means clustering for player categorization
12.Learn ridge regression
13.Temporally locally-weighted rankings
14.WOWY as validation of replacement level
15.Revise ASPM with latest RAPM data
16.Conversion of ASPM to" wins"
17.Lineup Bayesian APM
18.Lineup RAPM
19.Learn SQL


Is #3 basically the inverse of an SPM prior informed RAPM?
Anu twists planned or unde considerration?

Any ETA on 11, 17 or 18?



This is an article I am bookmarking here to help me to remember to come back to it:
http://godismyjudgeok.com/DStats/2012/n ... 2003-2011/

No ETA on anything. Just a list of what I want to do.

I would like to re-run multi-year RAPM with aging accounted for in pre and post-processing. ASPM is already based on J.E.'s RAPM right now.

For 11.: Check out this post: http://godismyjudgeok.com/DStats/2011/g ... lustering/ and these 4 charts (click to the right to go to the next one): https://picasaweb.google.com/1183961692 ... 3768029810 .

For 17 and 18: Sooner or later I'll figure out how to do lineup RAPM. Probably use ASPM sum of lineup as a prior. Difficult to set up the whole regression, though--that's a lot of variables. Far more than for individual RAPM, actually. Also, it's an extremely sparse matrix. I need to figure out how to handle that number of variables (I would like to use R to run the numbers).

Thanks for the reply and the reminder on your cluster article. Any interest in doing one with more than 5 clusters?

Crow wrote:Thanks for the reply and the reminder on your cluster article. Any interest in doing one with more than 5 clusters?

Those other charts looked at more clusters. Need some way to figure out what number is "reasonable".

Thanks again for the reminder on the 4 charts. I forgot to click thru to the other chart pages.

Move SQL to #1.

-Chris

Eternal wrote:Move SQL to #1.

-Chris

I've started that Stanford class you recommended. My brother-in-law is in database management, so I could always ask him...or my brother who does web development for database webapps...

My interest lies more toward the visualization side, as I'm sure you've noticed. I have studied a lot about effective data visualization/proper web design.