View Single Post
Old 11-21-2016, 03:42 PM   #4
NoOne
Banned
 
Join Date: Apr 2015
Posts: 7,273
Infractions: 0/1 (3)
i wouldn't worry about how this setting affects you, a human player, nearly as much as how it affects the AI. it only affects overall rating, which is most likely not heavily relied upon when making a decision as it is with the AI.

what i mean is the ai cannot look at a pitcher that is well-rated and looks great untill you see he throws fb-sinker and nothing else. (maybe the ai is sophisticated enough for this example, but you get the idea... it doesn't understand nuance as well as we can)

So, this setting has profound influence on what your ai competition is doing and very little to do with what you see... in fact nothing oustide of a bloated or shrunken overall rating. if you see ai decisions that you don't like, this is something you should look into adjusting.

as long as your game's settings provides ratings that are more predictive than resulting statistics as the simulation runs, the AI is better off with less weight going to stats of the previous 3 years.

ratings better?
100%? yes, obviously
normal? most likely
low? in my experience with a maxed scouting budget and legendary scout, yes still beter.

haven't played on other settings, so i cannot attest... even so, the human eye will not be able to tell if it is anywhere near the breakeven point. i could be wrong about low accuacy setting, but quite certain about normal.

this setting incorporates potentially insufficcient sample sizes... the impact it has depends on the base it is affecting. in some cases it may actually improve ai perception and in others it could make it worse... i think this setting, and some others, are strongly correlated to the threads i see complaining about what the AI is doing in transactions and decisions.... this is one of the first things you should be willing to change your thoughts about.

there is no one-size-fits-all answer unless we all use the same settings that affect or are affected by ai evaluation. there are times when it's a good idea and times when it is a bad idea.... i'm not intimating that it should be one or the other, or i would be specific about it.

instead of looking at it from a top down approach... use resulting AI behavior as your compass, even if your idea of how things worked could be proven to be wrong

Last edited by NoOne; 11-21-2016 at 03:45 PM.
NoOne is offline   Reply With Quote