Blogroll
- Andrew Gelman’s Blog
- Big Computing
- Big Data Apps
- By The Numbers: Charles Blow's Blog
- CHANCE magazine
- Data Blog
- Data Mining and Predictive Analytics Blog
- Data Mining with MATLAB
- Federal Statistics
- Freakonomics Blog
- Frontloading HQ
- Hal Stern’s Blog
- Harvard Stats Blog
- http://amathew.com/
- Ian Ayres website
- Jalayer Academy
- John Myles White blog
- Malcolm Gladwell Archive
- Mr Science Show
- Politically Incorrect Statistics
- Population Counts
- Results Disoriented
- Social Science Statistics Blog
- Sourish Das's Blog
- StatAcumen
Charts and Graphs
Completely Unrelated to Statistics
Data Art
Jobs
Politics
Public Health and Medicine
R
SAS Blogs
Sports
Stats and the Law
Twitter
Interesting. I ran this same list recently. The list looks very different than fWar and rear. As long as WAR remains a nonstandard metric I like to see as many options as possible.
The tops players in openWAR are pretty much the same. What are some of the big differences that you are talking about?
There are some major differences between openWAR and the other WAR implementations. 1) openWAR is completely open and reproducible. 2) openWAR is based on the plate appearance level, rather than aggregate statistics.
The difference in aggregate stats may be the cause. For example, there’s Josh Donaldson who is currently ranked 3rd or 4th in rWAR and fWAR but way down the list in openWAR. This is a result of his fWAR “Def” number being much higher than his openWAR fieldingRAA.
One thing is very clear though, I think openWAR does a MUCH better job with pitchers than any of the alternative systems!
Another benefit to openWAR is that if you don’t like something, you can see exactly how we did it and criticize it or go ahead and suggest an improvement. Everything is open.
Pingback: openWAR in 2014 | Exploring Baseball Data with R