Research does not equal data driven results. If that was the case, we wouldn't be having this argument and there would be some quantitative effect of a "disruptive player" or "good clubhouse chemistry" that you could point to in terms of wins and losses. You seem to think that I'm making the assertion that personality and team dynamics don't matter. I am not. I am only making the assertion that no one had managed to prove anything specific with baseball (or really most pro sports) and that as such OOTP's model does not seem to have any basis on reality, is entirely made up, and is woefully inadequate to address factors that might actually matter.RonCo wrote: ↑Tue Jan 19, 2021 12:30 pmThere are literally thousands of research papers that attempt to quantify various elements of team dynamics. As I recall, Bill James has done some half-handed attempts at quantifying them as to how they apply to wins and losses without really finding anything.Ted wrote: ↑Tue Jan 19, 2021 12:19 pmRon, I'd appreciate it if you'd stop comparing things like park effects that we have numerical data to model upon with personality and baseball team chemistry that we do not have any hard data for.
Doing so is at worst intellectually dishonest and at best shows a shocking lack of understanding on your part on what a good model is.
Either way it cheapens this discussion with distractions.
You have made a lot of good points, but the way you choose to hold on to completely invalid comparisons makes it hard to wade through your argument and overall makes it difficult tot take seriously.
I have literally spent a large part of my professional career working in this field. I also have both academic and professional experience in the act of building models of complex systems. I have literally created models of cultures. I understand what I'm talking about. Or at least I have enough of an understanding to discuss the things you're bringing up with a analytical frame of reference that is valid.
I'm sorry that you feel I'm being intellectually dishonest, but I can't help that I suppose.
If one wants to model the baseball environment on the field, that model will be enhanced by modeling park effects--even if that model is "wrong." If one wants to model the baseball environment off the field, that model will be enhanced by modeling personalities--even if that model is "wrong." There is an equation there that is valid.
I think most of us would be fine with a model that made any kind of sense, or had some sort of hard data driven foundation. This one simply doesn't seem to. The labels it puts on players more or less demonstrates that it doesn't.
The game would not be enhanced by an entirely incorrect modeling of park factors. That is incorrect. It would be better off not having them than having ones that made no sense and had no basis in reality, unless you just want complexity for the sake of complexity.
There is, of course, some subjectivity in "What is better", but I think me coming from the standpoint that most people want a model to be based on something real and represent it fairly is a rather defensible one.