HOI4 - Together For Victory AI Update #3

[tom_jones]

That was my point. The combat units already in queue and reinforcements from war hog most of the equipment so very little is left for any garrisons.

This presumes the combat units get preferential treatment when it comes to equipment, which doesn't need to be the case, and generally isn't -- by default the game treats all units placed in the queue equally, distributing the available/produced equipment between them. Additionally, this problem affects all produced units, no matter their intended purpose, so I don't think it's an argument against dedicated garrison AI, specifically.

( and even if it did build garrisons specifically triggered by occupation needs they would be ready about right when the war ends ).

I don't believe this to be accurate -- the war, at least when performed by the AI vs AI, lasts typically couple years, sometimes more. During this time multiple states are captured and the front moves deeper in land, leaving these captured areas without garrison forces. And like discussed earlier, it takes a few months to build units, not years, so there's wide margin for them to get deployed and contribute. Even if they don't, there's more wars about to happen so having these units ready for future duties is hardly a problem? Especially when initial complaint was how the AI would be unable to prepare such units in advance...

 

[Alex_brunius]

And like discussed earlier, it takes a few months to build units, not years,

Simple task for you to prove yourself wrong: Start up a new 1939 game as Germany or 1936 game as Japan, go into console and write "human_ai". Launch game at speed 5 and wait until your at war and have your first partisans pop up (front moved far enough) then queue up say 20 cavalry divisions. Now observe how they take over a year to finish due to shortage of infantry equipment ( which is really bad for the AI in 1.2 at least ).

Even if they don't, there's more wars about to happen so having these units ready for future duties is hardly a problem? Especially when initial complaint was how the AI would be unable to prepare such units in advance...

Assuming the AI can win the wars in advance, and doesn't lose because it built useless cavalry divisions instead of more actual combat divisions

 

[joeenochs]

If the AI cannot counter resistance through garrison properly, Parodox should imho resort to cheating, in the sense that the AI countries experience much lower or no resistance. Of course in an ideal world the AI could handle all these things like a human, but whenever it cannot, the challenge for the player should be the top priority. Such a cheat would be better than others, because unlike a combat bonus you don´t feel it directly when fighting the AI. Other companies (at least some companies sometimes) manage to release strategy games, which are a challenge on day one. Of course the more open a game is and HOI4 is very open, the harder it becomes to do this. Still for me if a game does not provide a challenge in single player, it is not fun, no matter, how many other good features it has. From SteelVolts dev diary, I get the feeling, that there are welcome AI improvements, but all in all, that they won´t be enough to provide a challenge in single player when playing a big country. Before I buy the expansion, I will actually test this using the free 1.3 patch. As a customer I simple don´t care how hard it is for the AI programmer, because I think, that a proper challenge in single player should be a design priority on day one and not just an afterthought.

 

[seattle]

- AI now attempts to detect and react to many encirclement opportunities, both against itself and against enemies. Includes evacuating dangerous areas and cutting off bottlenecks.

I hope that you implement a.i. modifiers to that. For instance: The Wehrmacht encircled entire Soviet armies due to poor leadership. You could use the "Barbarossa" decision/event to give the Soviet a.i. a 6 months malus to encirclement detection.
Same for a.i. Germany once Hitler takes command of the army (historically from 1942 onwards).

Any plans on that or will the a.i. be better at encirclement evasion all the time? That'd lead to a disappointing Barbarossa experience.

 

[Alex_brunius]

I hope that you implement a.i. modifiers to that. For instance: The Wehrmacht encircled entire Soviet armies due to poor leadership. You could use the "Barbarossa" decision/event to give the Soviet a.i. a 6 months malus to encirclement detection.
Same for a.i. Germany once Hitler takes command of the army (historically from 1942 onwards).

Why not use the built in effects instead? Have the production and research weights set up so German AI are able to get massive airsuperiority during 1941-42 (which slows down Soviet forces and makes them easier to encircle), and then after 1942 the Allied and Soviet gradually wear down the Luftwaffe and the reverse is true in 1943 onwards.

Making AI Soviet or Germany intentionally weaker is IMO a bad move since I suspect they will face of against human players and be way to easy if that is the case.

 

[joeenochs]

I hope that you implement a.i. modifiers to that. For instance: The Wehrmacht encircled entire Soviet armies due to poor leadership. You could use the "Barbarossa" decision/event to give the Soviet a.i. a 6 months malus to encirclement detection.
Same for a.i. Germany once Hitler takes command of the army (historically from 1942 onwards).

Any plans on that or will the a.i. be better at encirclement evasion all the time? That'd lead to a disappointing Barbarossa experience.

As it stands, Paradox has their hand full of creating one half decent AI, let alone different versions with different "personalities". It would be like trying to run before you can walk. So I am pretty sure that it is not in although I cannot be 100% sure of course.

 

[Xuanzue]

no fix yet for the stamping that suffers the AI doing bad disembarks? repeating the process till millions of manpower has been wasted

 

[SensitiveDannyBoi]

Can the AI decide to fabricate on other countries outside of focuses? It would be nice to see a brain behind that element of the game. My Understanding is that the ai just works through focuses, historical, in a certain order, and ahistorical in a random order.

Edit: Also can the AI build intelligently now? Like infrastructure in Africa if it has alot of units there.

 

[Axe99]

Cheers for all the responses to questions Steelvolt - top-notch communication and always interesting to read .

Why not use the built in effects instead? Have the production and research weights set up so German AI are able to get massive airsuperiority during 1941-42 (which slows down Soviet forces and makes them easier to encircle), and then after 1942 the Allied and Soviet gradually wear down the Luftwaffe and the reverse is true in 1943 onwards.

Making AI Soviet or Germany intentionally weaker is IMO a bad move since I suspect they will face of against human players and be way to easy if that is the case.

It'd depend a fair bit on how the encirclement code is set up, but I like the point @seattle makes here - Germany did very well early because both the Polish and French armies were encircled far more easily than expected. Some kind of AI malus for detecting encirclements could potentially go well. Might work even better if linked to doctrine/combat experience? Just thoughts, as always could be rubbish.

 

[princeseth]

@SteelVolt

Thanks for the DD and your work ... dont you think that the naval invasion target system should change ? i think not all coastal provinces should be suitable for invasion but only some of them (just like DH) ... imo that would be better for the ai and the human player ...

 

[Zog I of Albania]

Awesome! Can't wait to see if King Zog will accept the Ultimatum even if he has superior firepower.

 

[him_15]

Great Job Steelvolt, AI improvement is always welcome, can't wait to try out the AI. I have one question though, would the AI now occasionally halt advance when there are actually no enemy standing in front of? This is a problem that brother me a lot at the moment.

 

[tom_jones]

Simple task for you to prove yourself wrong: Start up a new 1939 game as Germany or 1936 game as Japan, go into console and write "human_ai". Launch game at speed 5 and wait until your at war and have your first partisans pop up (front moved far enough) then queue up say 20 cavalry divisions. Now observe how they take over a year to finish due to shortage of infantry equipment ( which is really bad for the AI in 1.2 at least ).

... Why would you want it to build 20 cavalry divisions at once? Conquered territory expands rather slowly, and you don't need such large numbers for suppression from the day 1. It's far more likely for the AI to request some 3-5 units every few months, which is quite easier to fulfill.

Shortage of infantry equipment is the result of current silly approach to production, with the AI never caring about running a deficit and ordering more and more units no matter if it needs them, or not (which only makes problem worse, as more units = even more equipment needed to keep them supplied) A simple change to production script, "don't make new stuff if you don't have spare gear" helps to smooth things out a lot, and this is without even touching the actual production code to make it borderline competent, something the game programmers are in position to do.

Assuming the AI can win the wars in advance, and doesn't lose because it built useless cavalry divisions instead of more actual combat divisions

If anything, short of the gamey cases where resistance disappears overnight, the garrison units are more likely to help win the wars than harm the effort -- with resistance activity suppressed the AI has more operating factories to produce stuff, and can expand its industrial base further on top of that with the civilian factories, vs the current game of whack-a-moledamaged-shit it gets permanently stuck in.

 

[Dragnar]

But to kick it off I thought I would tell you yet another story from the trenches.[...]

I love stories like these.
They go to show that with a system as complicated as AI, even statements and functions that, when looked at in isolation, appear totally correct and smart can still cause really silly bugs somewhere down the road.

 

[Alex_brunius]

... Why would you want it to build 20 cavalry divisions at once? Conquered territory expands rather slowly, and you don't need such large numbers for suppression from the day 1. It's far more likely for the AI to request some 3-5 units every few months, which is quite easier to fulfill.

Because 20 cavalry divisions is a minimum of how many your going to need... Especially if you want the most Industry out of the occupied territory and run harshest occupation. As Germany I end up with around 150-200 cavalry divisions when I want to suppress partisans all over occupied Europe and a bit into Russia in 1941 with the somewhat inefficient garrison order, and building them 3 at a time @ 120 days would take me over 16 years to get enough suppression. Hopefully you agree with me that this is way way to slow rate of training?

I'm pretty much always forced to deploy with minimum training to push out enough garrisons (because it doesn't matter for suppression), but the AI is not so fortunate.

If anything, short of the gamey cases where resistance disappears overnight, the garrison units are more likely to help win the wars than harm the effort

My point was in the example where the AI makes zero progress and don't occupy anything. In these cases the garrisons are wasted.

Trying to estimate garrison needs based on future conquests it will pull off is way above what the AI can do.

 

[SteelVolt]

Can the AI decide to fabricate on other countries outside of focuses? It would be nice to see a brain behind that element of the game. My Understanding is that the ai just works through focuses, historical, in a certain order, and ahistorical in a random order.

Edit: Also can the AI build intelligently now? Like infrastructure in Africa if it has alot of units there.

Yes, it can fabricate outside of focuses. It has been able to to so for quite a while, but it will not do so just because more territory is nice, but because it thinks it will help its other war efforts. An AI the freely fabricates makes for very chaotic games.

The construction AI has gotten a couple of minor fixes, but will get more love in the future. More work has been put into production and deployment.

 

[SteelVolt]

Great Job Steelvolt, AI improvement is always welcome, can't wait to try out the AI. I have one question though, would the AI now occasionally halt advance when there are actually no enemy standing in front of? This is a problem that brother me a lot at the moment.

The AI can have a tendency to be a bit cautious about its planning bonus. It has not been given a proper overhaul, but I have tweaked it and it should be somewhat better now.

 

[SteelVolt]

@SteelVolt

Thanks for the DD and your work ... dont you think that the naval invasion target system should change ? i think not all coastal provinces should be suitable for invasion but only some of them (just like DH) ... imo that would be better for the ai and the human player ...

I see no reason in blocking the player from being able to invade badly. And even a seemingly poor invasion location can turn out to be an important step in the grander scheme of things.

When it comes to invasion AI, check back in tomorrow, when I intend to include some points about that

 

[seattle]

Cheers for all the responses to questions Steelvolt - top-notch communication and always interesting to read .

It'd depend a fair bit on how the encirclement code is set up, but I like the point @seattle makes here - Germany did very well early because both the Polish and French armies were encircled far more easily than expected. Some kind of AI malus for detecting encirclements could potentially go well. Might work even better if linked to doctrine/combat experience? Just thoughts, as always could be rubbish.

IIRC mods usually give the player an incentive to pick an event/decision like "Barbarossa" or "Pearl Harbour".
For instance, why would the Japanese player ever attack the US? Three possible gameplay-solutions:
1. Force him by firing the event with no choice to reject it. --> poor game design
2. At one point let the US auto-join the Allies and dow Japan --> poor game design
3. Give the Japanese player an incentive like "surprise bonus: for the next 6 months gain bonus xyz while the US gets malus abc" --> the player has an interesting choice to make which can go both ways

As for Barbarossa:
Why would the German player start the war with Russia via NF, event, decision? Why not simply dow Russia precisely when the player wants to?
1. The NF/event/decision could be the only possible way to dow Russia --> questionable game design
2. The NF/event/decision provides the player an incentive to go that route. Some examples of previous mods: 6 months bonus to movement, logistics and combat abilities while the Soviets get a malus. Some events fire like "surprise air attacks with random number of Soviet planes destroyed on the ground" (historical and makes perfect sense, think also of Israel's wars).

I'm a huge fan of choices and no railroading. Choices however (save for house rules) only work as intended when you give the player incentives to pick options that are not favourable without the incentives. From then on it's only a balancing act, one that can be gradually improved via customer feedback.

@SteelVolt
Anything planned in that direction? All those popular mods from HoI1-4 can't be completely wrong.

 

[Moser]

2. The NF/event/decision provides the player an incentive to go that route. Some examples of previous mods: 6 months bonus to movement, logistics and combat abilities while the Soviets get a malus. Some events fire like "surprise air attacks with random number of Soviet planes destroyed on the ground" (historical and makes perfect sense, think also of Israel's wars).

I think this would be really cool. I always wanted to figure out how to script the mass destruction of the Red Air Force in the opening weeks of Barbarossa.