View Single Post
Old 12-04-2005, 02:22 PM   #21
Jayzone
All Star Reserve
 
Jayzone's Avatar
 
Join Date: Apr 2004
Posts: 706
I, too, want to make the game to have in-game rating translation ability and I have already make a few threads/posts about it. The main benefit of having in-game rating translation is actually on AI improvement. As computer AI knows (more or less) what kind of production will the player have based on ratings, computer can then value the said player correctly. However, this is not the disscusion intent of this thread (see my sig thread for the disccusion on this).

The one main issue about rating translation is that rating translation is 'unique' as it is the result of the interaction between specific league rating distribution and league total. In other words, the translation 'formula' will be different for every league. One solution to this is to investigate the data after the said interaction between rating distribution and league total (after-season data). From there, a mapping of ratings to stats could then be constructed via regression analysis. The downside of this is that it's done 'after the facts' so OOTP would have to trial simulate ahead first in order to obtain the mapping result.

Fixed rating translation formula won't work as it would fail to take neither specific league rating distribution or the influence of league total into account. So, you would possibly have a very inaccurate translation result based on fixed formula for a league with abnormal rating distribution and/or abnormal league total.
__________________
Rating Translator v0.93b (for OOTP 6/6.5) - released/updated on 03/13/05.
- The function of this utility is to translate both the player's ability and talent ratings into stats (expected performance).

Note: Rating Translator is now also hosted at jazzrack's CLB - OOTP Download.

Click here for the discussion on "How computer GM may be made smarter by having built-in rating translation ability"
Jayzone is offline   Reply With Quote