Do the AI use tanks & AT weapons now? If not, SPG on tanks?

[foxMiD]

I know there is a LOT of posts about this on entire web. No change, AI don't make use of tanks nor anti-tank weapons. However, I just cannot find anything newer than august 2016.
One may argue against my topic, that's because it's still the same. It would just be nice with a need look; It is 6-7 months ago, and a game operates more on hamster to dog years

If the AI doesn't, could it be worth trading breakthrough & some armor for the soft attack/SPG upgrade? It's pretty damn strong, so I thought maybe it could be worth it. While loosing breakthrough, tanks are still 3 times faster than infantry.
It's from a German perspective.

Thank you!

 

[Antediluvian Monster]

They definitely build AT these days. And the major problem with tanks used to be that they forever stuck with lights.

 

[foxMiD]

So, they build medium (even heavy) now?
If so, as Germany is it worth it to get the Soviet Pact, to get that upgrade faster? Or, in any case get anti-tank guns for some of your own units?

 

[Antediluvian Monster]

They try at least. In SP absolutely go for the pact, tech buff for player is much more valuable pro than tech buff for AI is a con.

Ultimately I don't see AI using tanks in cohesive enough manner for it to be a huge threat, so you can easily skip ATGs. It just doesn't seem to realize that the armour is supposed to be concentrated on operational level too.

 

[foxMiD]

They try at least. In SP absolutely go for the pact, tech buff for player is much more valuable pro than tech buff for AI is a con.

Ultimately I don't see AI using tanks in cohesive enough manner for it to be a huge threat, so you can easily skip ATGs. It just doesn't seem to realize that the armour is supposed to be concentrated on operational level too.

Oh ye, I actually didn't realize how much bonus Doctrine research & xp it gives. Just focused on the tanks only Get those bonus researches, skip the tank it self, I guess then?
The tank cost is just massive & quite slow.

 

[Meglok]

They try at least. In SP absolutely go for the pact, tech buff for player is much more valuable pro than tech buff for AI is a con.

Ultimately I don't see AI using tanks in cohesive enough manner for it to be a huge threat, so you can easily skip ATGs. It just doesn't seem to realize that the armour is supposed to be concentrated on operational level too.

The ai is much better at trying, and sometimes succeeds depending on the size of the nation's industry. Unfortunately the new industrial changes can be too much for the ai of smaller nations to handle. The ai doesn't handle shortages well resulting in line shutdowns and the ai templates don't adjust well to this. Often divisions will be deployed with shortages in equipment the nation can't produce but the tempate expects.

 

[Dalwin]

The current problem concerning AI use of tanks is not that it doesn't try try to produce mediums. It tries, but it does it very badly. The AI is bad with its production decisions. It wastes too much efficiency by swapping lines around too often.

It also does not have the foresight to know that it will produce mediums and as a consequence makes too many lights early on when it should have been focusing on things like artillery and motorized instead. This misjudgment is exacerbated by the way it swaps templates. Lets say it has 10 light armor divisions at the point where it finishes researching mediums. There are a number of methods used by human players to gradually introduce the mediums, but the AI uses none of them. It immediately converts all of the light armor divisions to a medium armor template. It now has thousands of unused light tanks (that it never should have built) and 10 medium tank divisions with no tanks in them.

It gradually fills those tank units but at the same time will be building too many new ones. It is quite probable that when the war starts those medium tank divisions will still be nowhere near full strength and will never get to be.

(To be fair, I have seen MP streams where rookie players make the same horrible mistake regarding template swaps. They have been soundly crushed as a result.)

 

[foxMiD]

It also does not have the foresight to know that it will produce mediums and as a consequence makes too many lights early on when it should have been focusing on things like artillery and motorized instead. This misjudgment is exacerbated by the way it swaps templates.

Yep. That's always the problem with such simulators. I'm no programmer, but I do know it's hard to fix such ... Well, to be fair it's not a bug nor a glitch.

 

[Dalwin]

Yep. That's always the problem with such simulators. I'm no programmer, but I do know it's hard to fix such ... Well, to be fair it's not a bug nor a glitch.

This could be fixed, but it would require a method that Paradox has decided not to use. This is properly fixed by using scripts to make some AI decisions in advance. Since it always lives in the now and has no real concept of the future, you simulate such by doing things like having a script that tells certain AI countries not to build light tanks at all. Nations such as Germany and Russia should not build lights while places like Hungary and Rumania should.

The designers have taken the stance that a scripted AI is too predictable thus making it less challenging in the long run. While I can sympathize with their attitude, I think they need to find the middle ground and use a small amount of scripted decision making in key areas without relying overly much on such an approach.

@podcat @SteelVolt