AI released #27 minor league prospect

[TeeClev]

AI released #27 minor league prospect instead of retaining

[Lukas Berger]

Originally Posted by TeeClev

AI released #27 minor league prospect instead of retaining

That's not the AI releasing them, it's simply them becoming a free agent, which is their right after a certain amount of pro service years.

[pfholden]

Originally Posted by Lukas Berger

That's not the AI releasing them, it's simply them becoming a free agent, which is their right after a certain amount of pro service years.

Semantics aside of if this player was released or just became a free agent, this should be considered a bug. The AI can retain them by just putting them on the 40-man roster. If the choice on the 27th best prospect in baseball is to lose them for nothing, or retain them on a league min deal by placing on the 40-man, it should be the latter every time.

[Cwalker]

Originally Posted by pfholden

Semantics aside of if this player was released or just became a free agent, this should be considered a bug. The AI can retain them by just putting them on the 40-man roster. If the choice on the 27th best prospect in baseball is to lose them for nothing, or retain them on a league min deal by placing on the 40-man, it should be the latter every time.

I agree with pfholden on this. Carson Williams is already on the RaysÂ’ 40-man roster in real life anyway. I understand that when you start a new game, there are quite a few empty spots on the 40-man specifically to prevent a ton of AI roster moves before Opening Day. Still, players like him should be included on the 40-man in a new save, whether theyÂ’re projected to make the 26-man roster or not, so situations like this donÂ’t happen.

[TeeClev]

Originally Posted by Lukas Berger

That's not the AI releasing them, it's simply them becoming a free agent, which is their right after a certain amount of pro service years.

Yeah, sorry I uploaded quickly and release wasn’t the proper wording. The main issue is the AI letting a top 50 prospect just walk for free rather than throwing them on the 40 to retain him as pfholden mentioned. Would be immersion breaking to see top prospects regularly hit free agency.

[Lukas Berger]

I think this is actually a pretty complicated issue, much more than it might appear on the surface.

There's a lot that can be going on here.

1. With scouting on, there can potentially be a gap between how the default OSA rating see Williams and how the Rays see him. The Rays could potentially see him as like the 230th prospect instead of the 23rd. Now in theory you still wouldn't let even the 230th prospect go, except that

2. The Rays are likely having a major 40-man roster crunch here. They have the last couple years in real-life and will likely have that in-game the next couple years as well. Likely whoever they leave exposed to the rule 5 draft or let go here, some of them are going to look bad.

I'm pretty sure we have not made the AI smart enough yet to do what the real Rays do, and trade guys who need to be exposed to the rule 5 for slightly worse and younger guys who have a year or so left before hitting rule 5 eligibility.

So I'm not 100 sure this is quite as clear as just 'the AI is letting too good a prospect go for nothing'.

I mean, sure they are, and ideally this wouldn't happen, but this is also probably almost as tailor-made a difficult situation for the AI as possible, where almost anything they do to deal with their 40-man crunch is going to look sketchy if you look at it in isolation.

[cojo82]

i had problems with minor league roster limits. once i removed the limits the AI wouldn't release top 100 prospects.

[Rain King]

I think that 5-6 month injury probably adds quite a bit to the AI weirdness here also. In all of the tests I've run he gets called up (thus added to the 40-man) during the season at some point. Obviously the injury prevents that here.

[TeeClev]

Originally Posted by Rain King

I think that 5-6 month injury probably adds quite a bit to the AI weirdness here also. In all of the tests I've run he gets called up (thus added to the 40-man) during the season at some point. Obviously the injury prevents that here.

Yeah, I'd get that if he wasn't fully healed by free agency filing date and they added him to their active roster after I edited him back into their minors and on the 40 man.

[pfholden]

Originally Posted by Lukas Berger

I think this is actually a pretty complicated issue, much more than it might appear on the surface.

There's a lot that can be going on here.

1. With scouting on, there can potentially be a gap between how the default OSA rating see Williams and how the Rays see him. The Rays could potentially see him as like the 230th prospect instead of the 23rd. Now in theory you still wouldn't let even the 230th prospect go, except that

2. The Rays are likely having a major 40-man roster crunch here. They have the last couple years in real-life and will likely have that in-game the next couple years as well. Likely whoever they leave exposed to the rule 5 draft or let go here, some of them are going to look bad.

I'm pretty sure we have not made the AI smart enough yet to do what the real Rays do, and trade guys who need to be exposed to the rule 5 for slightly worse and younger guys who have a year or so left before hitting rule 5 eligibility.

So I'm not 100 sure this is quite as clear as just 'the AI is letting too good a prospect go for nothing'.

I mean, sure they are, and ideally this wouldn't happen, but this is also probably almost as tailor-made a difficult situation for the AI as possible, where almost anything they do to deal with their 40-man crunch is going to look sketchy if you look at it in isolation.

I appreciate the nuance in your explanation here. I am sure this is pretty complicated on the backend. But I am seeing this issue on more than just this prospect. For example, Brady house put up 3.3 WAR and a 130 wRC+ in AAA in 2026 and then the Nats just let him go. (San Antonio was an expansion team) I can post more examples if helpful, but it does seem something is off with top or MLB-ready prospects just being allowed to walk instead of being kept on a league-min deal.

[Rain King]

Originally Posted by TeeClev

Yeah, I'd get that if he wasn't fully healed by free agency filing date and they added him to their active roster after I edited him back into their minors and on the 40 man.

I mean, it comes down to "when" then checks for that 40-man addition happen and for whom.

Basically, it seems to me like that injury has probably exposed an additional time where that should occur (right before minor league free agency).

The issue seems likely to be a combination of those factors.

- A player that would normally already be on the 40-man except for how OOTP handles starting rosters.
- A player that gets injured very early and is injured for most of the season.
- A player whose minor league free agency has come up

I think this helps show how difficult the AI is and how easily circumstances can occur that cause something to slip through.


*Edit* - Yeah, the House example seems to show something bigger potentially. Letting him go to see another team pay that type of contract seems odd...although I'd also like to maybe see his ratings and understand how you handled expansion (i.e. what kind of roster/budget situation was San Antonio looking at).

[Lukas Berger]

Originally Posted by cojo82

i had problems with minor league roster limits. once i removed the limits the AI wouldn't release top 100 prospects.

Yes, with the changes we've added to improve the roster AI this year and stop or slow down the mass releases/mass movements, you really can't use minor league roster limits any more or they will cause issues.

Now the AI really depends on having roster flexibility throughout the minors.

[pfholden]

Originally Posted by Lukas Berger

Yes, with the changes we've added to improve the roster AI this year and stop or slow down the mass releases/mass movements, you really can't use minor league roster limits any more or they will cause issues.

Now the AI really depends on having roster flexibility throughout the minors.

This is good to know, thank you. FWIW, I didn't add any roster limits in the sim with the House example, everything is default.

[TeeClev]

Originally Posted by Lukas Berger

Yes, with the changes we've added to improve the roster AI this year and stop or slow down the mass releases/mass movements, you really can't use minor league roster limits any more or they will cause issues.

Now the AI really depends on having roster flexibility throughout the minors.

Would you recommend not using roster size limits this year? Previous years I would set all levels to a max of 32 except rookie ball I'd put it as 50 max. I just get worried about the AI hoarding. lol

[malor]

Originally Posted by Lukas Berger

Yes, with the changes we've added to improve the roster AI this year and stop or slow down the mass releases/mass movements, you really can't use minor league roster limits any more or they will cause issues.

Now the AI really depends on having roster flexibility throughout the minors.

So that explains the AI minor league player hoarding of 40+ players per roster that I was seeing, with the lesser ones just sitting the bench. not getting playing time. I was only fast simming and did not dig into the details, planning to revisit once I could start a new game with more bug fixes applied.

[pfholden]

FWIW, I am still regularly finding top 100 prospects being allowed to leave as free agents. I recognize this is a complicated issue, but this is an immersion-breaker for me and I hope there's something that can help address it

[Sweed]

Originally Posted by TeeClev

Would you recommend not using roster size limits this year? Previous years I would set all levels to a max of 32 except rookie ball I'd put it as 50 max. I just get worried about the AI hoarding. lol

I use similar numbers for my minors along with age and service time limits. And yes, hoarding is the first thing that comes to mind. Career minor leaguers wouldn't be a big deal unless you as the human ran a "tight ship" and needed injury replacements. Perhaps there'd be a shortage?

Originally Posted by Lukas Berger

Yes, with the changes we've added to improve the roster AI this year and stop or slow down the mass releases/mass movements, you really can't use minor league roster limits any more or they will cause issues.

Now the AI really depends on having roster flexibility throughout the minors.

Lukas you are so busy with all going on during the release phase of v27 I hate to trouble you, but if you have time...

Would you say age/service limits are still ok for the AI as long as roster number limits are set to none? I like the way players will "age out" and retire because they aren't good enough to say, make the jump from A+ to AA. Without age limits a guy could sit at A+ until he was in his 30's. I'd hope if no limits includes age, these guys are still "smart enough" to retire, or be cut and not signed by a different AI team, rather than sit at A+ until they reach their 30's?

[Rain King]

I also wonder if it might be Ok to set a roster limit that is a few above the real life versions. Particularly if we set the Number of Position Players and Number of Relievers numbers to conform to the lower (real life) roster numbers.

[Cwalker]

Originally Posted by pfholden

FWIW, I am still regularly finding top 100 prospects being allowed to leave as free agents. I recognize this is a complicated issue, but this is an immersion-breaker for me and I hope there's something that can help address it

Agreed. It’s clearly a complex issue, but it appears to be a recurring problem rather than an isolated incident in saves right now.

[Sweed]

Originally Posted by Rain King

I also wonder if it might be Ok to set a roster limit that is a few above the real life versions. Particularly if we set the Number of Position Players and Number of Relievers numbers to conform to the lower (real life) roster numbers.

I was having the same thoughts about what might or might not work. I was thinking if one did 40 to 50 rather than unlimited it would mean a good sized reserve. Then if you use the "play players with higher potentials" option these "reserve" players would only see playing time in emergency situations. In my game I've had limits around 35 and it's not unusual to see low rated players getting very few PA's or innings pitched.