Quote:
Originally Posted by RonCo
It is, to me, unreasonable to say: I have a broken process, but am unwilling to move to the only model I've seen put forward that relies on real data merely because there exists some unquantifiable level of statistical uncertainty in applying the data across the board.
|
Again, I find this rhetorically misleading. It's not that there exists some 'statistcal uncertainty' that makes Tango inapplicable to minor league players; it's that there is (rather extreme) bias in the sampling. The data simply isn't relevant to the minor league population.
Quote:
Originally Posted by RonCo
If you have a model that is proven to represent human baseball players better, bring it on. If you point to OOTP's model as one of those, I'll respectfully disagree. You, of course, will win that one. This is like watching Corporate America at its finest, really. I live this every day. "Since we don't have 100% certainty in going the way that some data says is good, we're now justified in doing whatever the heck we want and calling it great!"
Then six months later you're in another meeting trying to figure out why the process is still messed up.
 |
I've consistently pointed to flaws in OOTP's development system. I am not an apologist for something that has obvious flaws- every post I've made acknowledges the problems with OOTP's development engine. How you've interpreted my arguments as being some kind of microcosm of public debate in American politics I don't think I'll ever be able to fathom. All I've argued is that statistics shouldn't be misused, and I've merely been regurgitating what anyone would learn in a Stats 101 course- nothing more. Biased samples are not random samples- and that's all I've said. In American politics, statistics are consistently
misused to support courses of action; I'm arguing against doing that.