Quote:
Originally Posted by MizzouRah
Here is another one using my current eval settings.. (haven't changed yet)
San Francisco is leading the NL West by 5 games with 25 games to go. There is no reason to release this player. He was an All Star this year to boot. Now Washington signs him to help them stay atop the NL East.
The news article even mentions this signing as Washington winning the lottery for his services - BUT does mention his age as a concern. So, I believe you guys are right when you say the AI does these moves due to a players age.
I'm just not sure changing eval settings will help fix this... but maybe changing to 30/50/15/5 will help eliminate these types of releases. I'm currently using 50/30/15/5, so essentially I'll be switching the 50 and 30 (ratings and stats around).
|
it would only help if those types of players consistently perform better than their ratings... so, probably not going to matter.
but, in the context of a player that does do this, they are probably valued more highly due to the higher overall ratings... that would be one tick for the "improve" column, but doesn't guarantee a decisions... it might be simply at that age if a viable destination is found, they just do it, so that they get something, rather than nothing.
if these types of player rotatate through more than 2 teams a season then i'd definitely bring that to their attention. e.g. if the destination teams takes the trade, they should be more willing to retain longer, since that's the only reason they'd go for it.
if you want the AI to make decisions based on data that isn't necessarily sound, increase stats portion of the ai evaluation (it's using small sample sizes). this only affects overall, so it's quite inconsequential to most human beings -> it's a common suggestion to ignore overall rating when making a decision - e.g. i only use it for filtering lists. i currently use 50-60% ratings and the rest stats, but i'm thinking of testing even less stats the more i think about it. i wish i could set the AI to have a better scouting accuracy, too...
i like being hamstrung, but the AI needs all the help it can get. giving it bad info can't be helpful. if i recall offensive numbers increase 5-10% with 100% ratings... but i could be confusing that with the accuracy settings. it's been a few years at this point.