from Amazon
Friday, January 09, 2009
Modeling Baseball Player Ability with a Nested Dirichlet Distribution
A forecasting/aging paper by Brad Null. Excellent paper, and I recommend it. The author even went out of his way to acknowledge the sampling bias, and references one of my discussion threads. I like him just for not limiting himself to just academic papers. A couple of quick takeaways:
- I would have preferred a smoothed function by component (see Figure 14)
- the binomial approach, first introduced (to me) by Voros, would make more sense with K/BB split away first (see Figure 7)
(I doubt it would make even a hair of a difference, but it would make more sense I think)
- even with all these fantastic breakdowns, the final result is on par with Chone, ZiPS, PECOTA, and a bit ahead of Marcel (see Figure 22)
(Hat tip: Victor)
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date