I have sufficiently addressed the critique that I “under-cited” the sabermetric literature. I’m not going to comment on it any further. I didn’t do anything even borderline inappropriate. It’s nit picking. I am shutting down comments on this subject. If you want to write or talk about what an awful person I am, you are free do so in another forum.

The unfortunate part of the Birnbaum criticism is that he wanted to take exception to academia, and focused his thoughts on J.C.’s paper in particular. Other than the line I quoted, which many may read as being dismissive of otherwise quality work, the rest of the paper stands well on its own.

I only have two issues with the presentation of the paper.

The first is: “However, this finding, though widely accepted, has not been tested with sufficient statistical rigor”.

I don’t know how widely accepted it is, and I’m not sure that “sufficient statistical rigor” is a necessary requirement.

The “widely accepted” is a line thrown in, without basis. I wouldn’t claim it’s widely accepted by the mainstream, nor by the analysts. Maybe it is, but unless you substantiate it, it shouldn’t be mentioned.

As well, fielding also plays a large role, as Mike Cameron, Mark Kotsay, and Torii Hunter are prime examples of guys who got a fielding-bump to their salaries.

Seeing how much the coefficients jumped year-to-year was more of an indication that the study was lacking in something fundamental, rather than a huge shift in a market response.

The first is: “However, this finding, though widely accepted, has not been tested with sufficient statistical rigor”.

I don’t know how widely accepted it is, and I’m not sure that “sufficient statistical rigor” is a necessary requirement.

The “widely accepted” is a line thrown in, without basis. I wouldn’t claim it’s widely accepted by the mainstream, nor by the analysts. Maybe it is, but unless you substantiate it, it shouldn’t be mentioned.

Likely the most rigorous of these studies is the Allen/Hsu called Solving DIPS on my site. Even if it doesn’t meet whatever threshholds are required to make J.C.’s statement true, it is easily the most compelling of the anti-strong-DIPS position out there. That is doesn’t meet the threshholds doesn’t mean that it gets to be dismissed or rejected. The Allen/Hsu paper is a prime example of smart guys, who love baseball, who spend tons of their time, doing things that didn’t require formal training. The results, if undergone through sufficient statistical rigor, would have likely yielded the same conclusions.

The second: to the extent that a bibliography is created, other DIPS-related research should have been prevalent.

The unfortunate part of the Birnbaum criticism is that he wanted to take exception to academia, and focused his thoughts on J.C.’s paper in particular. Other than the line I quoted, which many may read as being dismissive of otherwise quality work, the rest of the paper stands well on its own.

The key thing is that you did your thing with the data directly. It was not a “meta-analysis” of DIPS studies (study of studies). You did not need to, I think, because you have the full population of relevant data to work with. You did not need to combine the power of other people’s samples to strengthen your case.

I don’t write academic papers, but I do read them. And while a lot of people in my field (finance) fling around untested hypothesis, I do try to get some data, open up R, and test such stuff formally.