Personally, I like the development model (at least for nonpitchers) a LOT already. Here's my conceptual defense of it.
1. First of all, recognize that the model WORKS, statistically: the right number of prospects pan out in the right range of ways. Thus, when I rationalize it below, it's because the results impress the heck out of me. And yes, Statfreak, there absolutely are real-life Angel Berroas and Zach Greinkes and Calvin Pickerings who establish that they're major-league calibre players at 20 or 23, then dive below that surface never to recover. (Um, I think you can figure out who my favorite real-life baseball team is...)
2. "Potential" or "Talent", in this game, is never meant as God's Absolute Truth to me - it can't be, given all the changes that follow. It represents, clearly, What The Ideal Human Scout Would Perceive. But even very smart humans are flawed.
3. Why would those estimates go down, without injury? Lots of reasons:
* The estimate hoped the player might "fill out", physically, but he didn't.
* The estimate hoped the player would make adjustments to his stride or swing that either it no longer looks like he can make, or that, when he makes them, don't actually help as expected.
* The estimate hoped he could learn to hit a good curveball (for example), and, um, no.
* The estimate hoped he could adjust well to the rather brutal schedule of traveling and intense daily workouts - and, if he makes it to the majors, the overwhelming media attention - and he didn't.
* The estimate hoped that we'd know about any injury right away, but there could be a tiny, persistent, nagging injury that no one catches. (I used to do long-distance running, but I got slower and slower over time - I think I was getting some barely-noticeable cousin of shin splints).
4. Why would those estimates go up? Because the player figured out adjustments in ways even a brilliant scout wouldn't have expected.
*********************
I'm not sure why the pitcher model, unlike the nonpitcher model, generates such low potentials and relies on pitchers to exceed those potentials over time. To me it would make more sense to generate high pitcher potentials and just have a lower rate of pitchers fulfilling them. However, even if I'm right -- and I might not be -- don't anyone think it would be less frustrating to play.
