The Holy Grail (July 29, 2003)
The first component of the holy grail search is... ...to reduce the margin of error as much as possible, given your sample size, to establish a player's set of true talent tools levels (power, speed, strike zone judgement, picking up the third base coach, etc, etc). You certainly do NOT want one unifying number.
The second component of the holy grail search is to model how all these types of players, with varying quality of tools, leverage their context (be it opposing players who have their own quality of tools, or the environment like park, weather, men on base, etc), so that we can establish, again with the smallest margin of error with the highest confidence level, the probability of the various outcomes of that PA (or even more precise, that pitch event). As you can see, the closer we get, the more dizzying numbers we have to contend with.
Once you've got that, once you know how every player will interact with all the other players, then what you've just created is "the perfect sim". That's your holy grail. Once you've got that, once you know exactly how things interact with each other, then you can start adjusting your environment, so that you can figure out the effect of your variables.
To further your holy grail, the third component, since we are talking about human beings, is to figure out how they learn/improve/degenerate based on age, conditioning, experience, and other conditions. And how the economics of the game affect the existence/movement of the players in the league. Once you've got this, you can now extend "the perfect sim" so that it can be played out for the next 20 years.
Because we are talking about people, you will always be left with a certain margin of error. This means that you can never get that holy grail, but only come closer to it.
Now, we've all got lives, but we're also all baseball nuts. Just trying to answer some bits and pieces of that is fun.
--posted by TangoTiger at 02:27 PM EDT