Manual vs AI Development

[jslmsca]

I was surprised at how the AI set up the development sliders. I use a 10 point rating system and it seemed like the AI generally chooses to work on topping up the top skills than work on areas of weakness. The one exception I noticed is if I switch a young player's position, suddenly the AI will shoot up the defence slider.

Another thing I was surprised to see consistently is the AI choosing to go below the middle line for fastballs on many pitchers, and focusing on other pitches. I was worried that I'd lose fastball stuff.

I decided to set the development sliders myself for my major league roster and key players in the system, including top prospects, going against what the AI suggested.

It was a mistake. Come the next development report, a lot of guys lost ratings. I was shocked how many and was disappointed with the outcome. Fearing that this would happen again, I just decided to turn everything over to the AI. I haven't reached the next month in swimming but will see what happens.

What are peoples' thoughts on properly setting up the development sliders? This isn't a large sample size, I know, but I was shocked at the result.

[Syd Thrift]

Originally Posted by jslmsca

I was surprised at how the AI set up the development sliders. I use a 10 point rating system and it seemed like the AI generally chooses to work on topping up the top skills than work on areas of weakness. The one exception I noticed is if I switch a young player's position, suddenly the AI will shoot up the defence slider.

Another thing I was surprised to see consistently is the AI choosing to go below the middle line for fastballs on many pitchers, and focusing on other pitches. I was worried that I'd lose fastball stuff.

I decided to set the development sliders myself for my major league roster and key players in the system, including top prospects, going against what the AI suggested.

It was a mistake. Come the next development report, a lot of guys lost ratings. I was shocked how many and was disappointed with the outcome. Fearing that this would happen again, I just decided to turn everything over to the AI. I haven't reached the next month in swimming but will see what happens.

What are peoples' thoughts on properly setting up the development sliders? This isn't a large sample size, I know, but I was shocked at the result.

I am 99% positive that the AI concentrates on skills where the player has potential remaining (that is, where there is a gap between potential and current ratings). There's not really much to any point to trying to develop a guy's Avoid K ratings if he's at 40/80 current and 40/80 potential even if that's his biggest weakness. If you want to improve those ratings you're going to want to put him in an offseason development lab (which fail a lot, which is as things should be because otherwise interesting players with too-high Avoid Ks, as an example, very rarely learn how to cut down on the strikeouts enough to turn their other tools into usable products).

[tonnage]

Originally Posted by Syd Thrift

I am 99% positive that the AI concentrates on skills where the player has potential remaining (that is, where there is a gap between potential and current ratings). There's not really much to any point to trying to develop a guy's Avoid K ratings if he's at 40/80 current and 40/80 potential even if that's his biggest weakness. If you want to improve those ratings you're going to want to put him in an offseason development lab (which fail a lot, which is as things should be because otherwise interesting players with too-high Avoid Ks, as an example, very rarely learn how to cut down on the strikeouts enough to turn their other tools into usable products).

Can this be confirmed be a dev. If that's the case, then what's the point of being able to take control from the AI? I'd like to know 100% what the deal is here, do I now have to go back and change all my sliders to AI control? If I don't, do I risk regression as the OP suggests?

[Syd Thrift]

Originally Posted by tonnage

Can this be confirmed be a dev. If that's the case, then what's the point of being able to take control from the AI? I'd like to know 100% what the deal is here, do I now have to go back and change all my sliders to AI control? If I don't, do I risk regression as the OP suggests?

For a player who’s maxed out at everything except for a couple ratings, there isn’t much of a point, no. If you have a younger player, often times those players never reach their potential ratings (which do go down over time if they’re not fulfilled) and so you do have some leeway there to build up some ratings before other ones. I do think that yes, 99% of the time you’re better off just letting the AI handle this screen.

[Will Beh]

Can't confirm exactly how things work since there are dozens of factors that go into the slider allocation, but indeed one of the big factors for the AI is focusing more on ratings that have a big difference between the current and potential rating.
Granted, this implementation of the AI is brand new, and I am 100% sure that people will be able to better optimize it. It's possible that the crappy development report you got is just unlucky; I still recommend manually changing it for players you are invested it.