Ai settings

roscoe55p · 06-15-2025, 12:50 AM

What are the preferred AI settings, mine reads 60-25-10-5.

Sweed · 06-15-2025, 11:14 AM

It's an endless conversation that will come down to personal preference. I use 25/25/25/25 for reasons you can find in the previous discussions I'll link here.
Others prefer more weight on ratings, while others less. A lot of conversation on the "how's and why's" in those threads. The old threads are still relevant as nothing has really changed much version to version with regard to evaluation settings.

v25 thread... https://forums.ootpdevelopments.com/...ght=evaluation

v23.. https://forums.ootpdevelopments.com/...d.php?t=335434

v22.. https://forums.ootpdevelopments.com/...light=settings

v21.. https://forums.ootpdevelopments.com/...settings+stats

Brad K · 06-16-2025, 05:15 PM

More weight on ratings causes the computer manager to make better decisions if scouting accuracy is very high.

More weight on stats causes the computer manager to make decisions that look more like real baseball decisions.

Caution: the setting "Include stats in ratings" reduces the effect of ratings selected in the AI evaluation settings.

Sweed's comment on personal preference is on target. Experiment and do what you like best. There is no right answer.

Murcer · 06-16-2025, 07:44 PM

Quote:

Originally Posted by Brad K

Caution: the setting "Include stats in ratings" reduces the effect of ratings selected in the AI evaluation settings.

I went to double check my settings to ensure I had not selected this but can't seem to find the checkbox for it. Can this be changed after starting a league?

Brad K · 06-16-2025, 09:50 PM

Quote:

Originally Posted by Murcer

I went to double check my settings to ensure I had not selected this but can't seem to find the checkbox for it. Can this be changed after starting a league?

MLB Menu -> League Settings ->Global Settings -> last item under Scouting settings, left panel, first section.

It can be changed after save has started.

Pelican · 06-17-2025, 10:19 AM

I started off using 40-30-20-10 based on my analysis of many comments (and threads) like the above, in other words, a compromise. It works for me, at least with scouting on high accuracy, and despite high TCR settings.

Notwithstanding the above, there have been times I was tempted to reset and try a completely new setting, like the 4 x 25 many advocate. I never have, mainly because I wanted to be able to compare new sims to old sims. That variable would confuse the comparisons. I mention this because, like me, you may soon find yourself feeling bound by your original choice, for continuity's sake. Choose wisely, my friend.

Despite my reliance on the same formula above, I have made small changes in TCR, scouting accuracy, and in the aging and development process (I keep gradually increasing both the development rate for pitchers and position-players, and the aging rate for each.). And, since I play with annual recalc off, the differences accumulate over time, versus IRL.

Bill Veeck · 06-18-2025, 12:58 PM

Quote:

Originally Posted by Brad K

More weight on ratings causes the computer manager to make better decisions if scouting accuracy is very high.

I've never seen/realized this.

Guthrien · 06-18-2025, 05:51 PM

I've been playing this game for years and I'll be damned if this setting makes much of a difference at all. I like that it's there, but I've tried a lot of variation without noticeable changes in trades, etc.

NoOne · 06-18-2025, 11:02 PM

unless there's some source for statistical analysis that shows what 'realistic' is, it's usually just a matter of personal perception which is hard to trust.

regardless of what is realistic or not simply look at cause and effect.

These things will be correlated, since ratings cause stats in the video game. No ifs ands or buts. So, it shouldn't be a huge difference.

Ratings are quite accurate for mlb players at 'normal' accuracy. Those with limited service time won't be as accurate but still, by this point they are typically known quantities. Changes in ratings may take some time to catch up, so again.. only young players and really old players are a concern here.

My exxrapolation from that is the ratings weighting will be more accurate for the players with 'settled' ratings. it'll use more false information or less accurate numbers for younger mlb players, aging players etc. Stats should follow these ratings shifts, but that too takes time to present a new avarege with confidence.

stats lag behind ratings changes (under the hood, not necessarily seen). What 'self-corrects' faster? a scouting update or resulting stats finding new averages with a suitable sample size? this may be different for diffrent ratings. it doesn't take the same number of AB for all stats that directly or indirectly relate to the various ratings.

i'd wager it's fairly similar, again less reason to see major differences in each option.

one causes the other. There are contexts where that might be advantageous and when it falls behind as far as best into 'today' that is visible (scouting accuracy effect).

I usually do 40-50% ratings and i make sure "2 years ago" is very low, like 5%. I weight current year more than 1 year ago, but to avoid any weird math with few games played in "current year", it's a small difference.

i try not to worry about it too much... if the defaults are close enough i may not touch it, lol. in the end it probably won't make a huge difference except with extreme weighting.

PSUColonel · 06-23-2025, 02:03 AM

I prefer the original settings of 65/20/10/5. I think it brings the best balance..but that’s just me. I like the idea of a good mix of a competitive AI coupled with mostly realistic real world decisions. I have expemeimented with both extremes over the years. I feel a higher premium on ratings definitely results in a more competitive Ai but without always making realistic decisions…but still at times you will still se realistic baseball decisions..but also..sometimes you won’t. If you really want the most competitive AI you can get but without completely abandoning stats I’d suggest 70/15/10/5. I admit I like this years and last years being a bit closer together.

The ratings determine the stats..so to me it makes sense to have higher ratings weight so that the AI doesn’t “over react to a small sample size. I believe the most competitive while still at least attempting to stay within reality is the 70/15/10/5..just my opinion though.

06-15-2025, 12:50 AM	#1
roscoe55p Bat Boy Join Date: Jan 2018 Posts: 19	Ai settings What are the preferred AI settings, mine reads 60-25-10-5.

06-17-2025, 10:19 AM	#6
Pelican Hall Of Famer Join Date: Mar 2021 Location: Wilmington, Delaware Posts: 2,939	I started off using 40-30-20-10 based on my analysis of many comments (and threads) like the above, in other words, a compromise. It works for me, at least with scouting on high accuracy, and despite high TCR settings. Notwithstanding the above, there have been times I was tempted to reset and try a completely new setting, like the 4 x 25 many advocate. I never have, mainly because I wanted to be able to compare new sims to old sims. That variable would confuse the comparisons. I mention this because, like me, you may soon find yourself feeling bound by your original choice, for continuity's sake. Choose wisely, my friend. Despite my reliance on the same formula above, I have made small changes in TCR, scouting accuracy, and in the aging and development process (I keep gradually increasing both the development rate for pitchers and position-players, and the aging rate for each.). And, since I play with annual recalc off, the differences accumulate over time, versus IRL. __________________ Pelican OOTP 2020-? ”Hard to believe, Harry.”

06-23-2025, 02:03 AM	#10
PSUColonel Hall Of Famer Join Date: Feb 2002 Posts: 13,104	I prefer the original settings of 65/20/10/5. I think it brings the best balance..but that’s just me. I like the idea of a good mix of a competitive AI coupled with mostly realistic real world decisions. I have expemeimented with both extremes over the years. I feel a higher premium on ratings definitely results in a more competitive Ai but without always making realistic decisions…but still at times you will still se realistic baseball decisions..but also..sometimes you won’t. If you really want the most competitive AI you can get but without completely abandoning stats I’d suggest 70/15/10/5. I admit I like this years and last years being a bit closer together. The ratings determine the stats..so to me it makes sense to have higher ratings weight so that the AI doesn’t “over react to a small sample size. I believe the most competitive while still at least attempting to stay within reality is the 70/15/10/5..just my opinion though. Last edited by PSUColonel; 06-23-2025 at 02:07 AM.

06-16-2025, 05:15 PM	#3
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	More weight on ratings causes the computer manager to make better decisions if scouting accuracy is very high. More weight on stats causes the computer manager to make decisions that look more like real baseball decisions. Caution: the setting "Include stats in ratings" reduces the effect of ratings selected in the AI evaluation settings. Sweed's comment on personal preference is on target. Experiment and do what you like best. There is no right answer.

06-18-2025, 05:51 PM	#8
Guthrien Minors (Double A) Join Date: Mar 2011 Posts: 193	I've been playing this game for years and I'll be damned if this setting makes much of a difference at all. I like that it's there, but I've tried a lot of variation without noticeable changes in trades, etc.

06-18-2025, 11:02 PM	#9
NoOne Banned Join Date: Apr 2015 Posts: 7,273 Infractions: 0/2 (4)	unless there's some source for statistical analysis that shows what 'realistic' is, it's usually just a matter of personal perception which is hard to trust. regardless of what is realistic or not simply look at cause and effect. These things will be correlated, since ratings cause stats in the video game. No ifs ands or buts. So, it shouldn't be a huge difference. Ratings are quite accurate for mlb players at 'normal' accuracy. Those with limited service time won't be as accurate but still, by this point they are typically known quantities. Changes in ratings may take some time to catch up, so again.. only young players and really old players are a concern here. My exxrapolation from that is the ratings weighting will be more accurate for the players with 'settled' ratings. it'll use more false information or less accurate numbers for younger mlb players, aging players etc. Stats should follow these ratings shifts, but that too takes time to present a new avarege with confidence. stats lag behind ratings changes (under the hood, not necessarily seen). What 'self-corrects' faster? a scouting update or resulting stats finding new averages with a suitable sample size? this may be different for diffrent ratings. it doesn't take the same number of AB for all stats that directly or indirectly relate to the various ratings. i'd wager it's fairly similar, again less reason to see major differences in each option. one causes the other. There are contexts where that might be advantageous and when it falls behind as far as best into 'today' that is visible (scouting accuracy effect). I usually do 40-50% ratings and i make sure "2 years ago" is very low, like 5%. I weight current year more than 1 year ago, but to avoid any weird math with few games played in "current year", it's a small difference. i try not to worry about it too much... if the defaults are close enough i may not touch it, lol. in the end it probably won't make a huge difference except with extreme weighting.