Science is neat. In many scientific experiments, you can run trials, generate results, slightly change the conditions, run trials, and generate other results. Then you can compare those results to measure the effect of the change that you made. I used to work in a neuroscience lab with fruit flies, and one of the first projects to which I was assigned attempted to measure the draw of potential mates against the draw of fresh food. Without going into detail, we were constantly futzing with the method and seeing what happened to the numbers in the end. It was not a very good experiment and it never came close to getting published. At least there were usually donuts.
Baseball isn’t like science. In baseball, there is but one trial, and it’s always going on. We can speculate about the effects of certain things, and we can feel pretty confident about our speculations, but we can never know for sure. We can never know for sure how many wins above replacement a player is or was worth. We can never know for sure the significance of a borderline pitch call. And we might never know the meaning of a play that Alex Gordon made in Detroit Wednesday evening.