Quote:
Originally Posted by injury log
Man, just saw this now. I also don't want to go around on this again, only explain my side: If I want to see if OOTP matches up with Tango, I want to apply the Tango test in OOTP to the same players Tango looked at in real life. Otherwise, it's apples and oranges, as far as I can tell. So that's what we did in beta. Tango's studies don't include players who dropped out of MLB. If you include these dropouts when you analyze OOTP development, you're not calculating Tango curves at all. Not sure what about that 'misses the entire point of the exercise'.
And there was a decent match between OOTP and Tango at one point in beta (still with some pitcher issues), but some late tweaks may have thrown things off. Many people requested that random potential drops be replaced with injury-driven potential drops - they wanted to see a reason behind dropping potential. No one presented any data to support this change, and it should never have been made, according to my analysis of real life stats of players before and after serious injury. This needs to be fixed before development as a whole has any chance of being right.
|
Tango ignores data of players who did not have a large enough sample size of stats merely to eliminate (or reduce) the influence of random variance. If stats didn't have random variance, he wouldn't pull them out. His goal is essentially to find out how a real player's "true ratings" progress over time.
In studying OOTP, we don't need to do that because we have their true ratings.
The main value of Tango's study is to use "qualified" data to predict how other players will progress. This is, for example, why a guy who is 26 and still stuck in the minors is generally considered to be done--because the _rule of thumb_ is that players peak at 27 (regardless of level). I agree there is little data to say that "regardless of level" part with 100% faith, but my opinion is that it's the best concept available and should be used until some other model is devised that makes more sense and is easily converted into a game environment. I understand you don't agree with me. Fair enough. You win.
I obviously have no idea of whether there was a match during beta testing or not, but I can tell you that--with injuries turned off--the match is worse in v9 than it was in v2007 when played at default settings.