Hearts of Iron IV

[Pete Dero]

Is it perfect ?  I hope not (I already paid for the 2 next expansions  ;))

Is it fun ?  I spend more time on it last month than I ever did on HOI3, so I would say yes.

While you are playing this is a good book (part 1 of 3) to read about the strategy and tactics but also the economic, political, and social aspects of WWII :  The War in the West - A New History - Part I : 1939-1941
https://www.amazon.com/War-West-History-Ascendant-1939-1941-ebook/dp/B00XGX9GIU/ref=sr_1_1?s=digital-text&ie=UTF8&qid=1467826892&sr=1-1&keywords=war+in+the+west#navbar

[SirAndrewD]

I know I "complain" a lot about the AI, but I hope it's understood that my criticism is meant to be constructive, and is in part seeing if the things I have happen are replicated often for other players.  My genuine hope is that bringing attention to the places where the AI is not working as designed so that they can be addressed in the (hopefully near) future. 

I wouldn't waste my time mentioning anything if I didn't like the game.  And I've already eaten my crow about my initial opinion that the game was "dumbed down". 

The AI is actually making logical decisions as far as it knows.  When it sees an undefended costal provinces, it goes for it.  When the fascists run low on trade options and WT gets high enough, the AI declares war on any place that has resources it needs, it just doesn't get it's unable or impossible to get to them.  It does its best to defend every front it's fighting on, and goes for the front that has the most weight of potentially hostile units, even if that front is Ethopia or Burma.  It prioritizes going to the conscription law that gives the country the most manpower, and then uses that manpower to spam as many divisions as it can. 

All the above things are issues that are going on under the hood, and they're fixable and I have confidence they will be fixed.  What's great is that the game is actually fun, even with some of these problems.  I've also found that most of these issues are less in the '39 scenario, primarily because the AI has less time at peace to make setup mistakes.  Keeping these things in the light are going to make them get resolved faster for the people that have the issue.

The Paradox forums constantly haranguing the devs over not having these things fixed at release is tiring, annoying and one reason I very rarely post there.   I am by trade a historian, not a game designer, but I can only imagine how hard putting a game like this together must be.  Heck, I did Historicity Beta on the first two games, so I absolutely had some insight into how these games looked prior to their release and what PD had to do just to get them ready for release. 

I think the discussion over here has been pretty solid all around.  And as for in game generalship, I make von Manstein look like Fredenhall.   8)

[sandman2575]

Some AI moves in games are a result of bad AI or bugs and whatnot. All AI decisions are results of algorithms; but to me, it's not immersion killing when an AI makes a stupid move since REAL people do it...so long as it is "reasonably stupid" and coherent. 

"Reasonably stupid" is brilliant, by the way. I think you're absolutely right there's an important distinction to make between 'dumb AI moves' that are plausible, and ones that are completely implausible.

The reality that even good generals sometimes made stupid mistakes is, to me, one of the strong arguments in favor of using the Battle Planner, and for taking as 'hands off' approach as you can to HOI's battles. Draw up the operation, then leave it to your generals - i.e. the AI - to achieve the objectives. This has one benefit of leveling the playing field and making for a more challenging game. It's your plan, but in another sense it's AI vs. AI, both handicapped by the same shortcomings.

I don't get a whole lot of joy from micromanaging. This is the main reason I used HOI3's 'army AI control' feature, which was the prototype for the more advanced Battle planner in HOI4. HOI3's AI army control was definitely clumsier and harder to use well than HOI4's. But I like the idea that not everything is in my control -- I think it makes for a far more interesting game.

I overall love HOI4's Battle Planner but do think it needs to be refined. I find myself sometimes intensely micromanaging how an army's front contracts or expands depending on what territories are being seized and when. Sometimes an army's front expands in rather ridiculous fashion, which then completely throws off how the AI positions the units and how it approaches reaching the goals you set out initially. Not sure how Pdox is going to fix that -- it seems like a 'baked in' problem to how they designed the battle planner -- but I do hope they improve it.

[sandman2575]

I've also found that most of these issues are less in the '39 scenario, primarily because the AI has less time at peace to make setup mistakes.

This is exactly my thinking in starting up a USA '39 campaign last night. I always started in '36 in HOI3, always. Am curious to see if things play out in a little more historical fashion.

[RyanE]

"Reasonably stupid"...I agree this is a brilliant term.  And there is the opposite for cheating AIs...unreasonably intelligent.  The definition of that would be SkyNet.

[glen55]

Yes, there's much brilliance in this game.  Especially as somebody who likes to play at the U.S., the pre-war part of building up the U.S. is just absolutely brilliant, way better than it was done before.

And yes, criticisms of the AI should rightfully take into account how incredibly difficult it is to pull of such a multi-pronged AI . . . but . . . still . . . in the long run, this game MUST have better AI or people just aren't going to keep playing.  As has always been true with HOI, the AI just isn't good enough.  That's where they have to concentrate on the patches, not on adding new packages of options.  It's the AI.

[MetalDog]

Speaking as a total grand strategy noob and Paradox noob, I find this game intolerably infuriating and I've shelved it possibly forever

But that is more on me than the game 

But, once again, I'd like to thank everyone who posted on this thread. I'd like to say it helped me, but it didn't (on me, not all of you) 

If you'll take it, I have two pieces of advice:

1)  If you think you like grand strategy games, keep trying them until one clicks

and

2)  Go into the game trying not to focus on everything.  There is a LOT of stuff to do in any grand strategy game.  Pick a few aspects you want to focus on and press Start and let the chips fall where they may.  And when you get tired or frustrated, start a new game and incorporate what you've learned in to the new play.

[RyanE]

Yeah, but people need to be realistic about it.  I really don't want patching resources spent on trying to make a perfect AI.  A reasonable one?...sure.  But listening to people complain about the every aspect of the AI, especially claiming its doing unrealistic things, when its not necessarily unreasonable based on history, will not help.

The problem I see, even on this board, is general claims of the AI doing this or that, without any context whatsoever.  This is a very complex game and throwing out a general statement about too many amphibious landings without some details won't help get anything fixed.  How about year, alliances, enemies, national focus, aggressiveness, production, research, etc.  I am still not familiar with HOI4 enough to even know...can a save be posted so others can see settings?

I also think a big factor is people who played a lot of HOI2/3 coming and knowing just enough to be dangerous in HOI4.  They are assuming HOI4 is less of a game than its predecessors or the AI is broken.

And I, now and then, pull out an old Command HQ example from my days playing in Windows 3.1.  I used to play it to death and beat the AI a lot, but got killed a couple times when the AI loaded up a huge armada and attacked me through Africa up into Europe.  It was the first MP game I had ever owned and I got so pissed off at the game and AI that I convinced my buddy at work to hook up our PCs through serial ports and play against me.  In our first game, guess what he did?  Built a huge armada and attacked me through Africa.  He cleaned my clock.  When I asked him why he chose that strategy, he said he wanted to do something unconventional.  I gained a little more respect for the AI on that simple game.

My point is that when people demand an AI be fixed, I think they sometimes lose sight of the fact that you wouldn't want an AI that is perfect.  There are times that an AI is broken and needs fixing, but sometimes its a game mechanic that is actually broken.  For example, the marching through another country's territory bug.  That to me is actually a game play bug that manifested itself as a perceived AI bug.  In HOI4, I have yet to run into anything that makes me throw my hands up in the air and declare, "I can't believe the AI did that, no human would ever do that."  I have seen human players in Combat Mission and other players do some really stupid things.

Maybe as people play it more they become jaded because they can see patterns in AI play.  But I think if people had to play scenarios clean and not be allowed to replay them the AI would appear to be a much better opponent.  The one all encompassing fault with AIs right now is that they don't learn from replaying a scenario.

[SirAndrewD]

  In HOI4, I have yet to run into anything that makes me throw my hands up in the air and declare, "I can't believe the AI did that, no human would ever do that."  I have seen human players in Combat Mission and other players do some really stupid things.

To each his own.  But I really, honestly, and not thinking myself a superior player at all find this so brutally hard to believe that it borders on hyperbole or makes me think that somehow you are playing a state of the game, so absolutely and fundamentally different from the one that I have, that it has to be something akin to a different game. 

But, then again who knows?  I'm just honestly and genuinely stunned at that feeling.  It really does make me wonder if there's more going on and wrong under the hood, perhaps operating in different ways for different people that some of the issues aren't as transparent as they seem. 

I don't, having an only barely rudimentary understanding of how the coding works understand how that works.  Especially considering Podcat has admitted that the AI doesn't understand concepts of encirclement and swiftness yet, and has very deep issues with front priority.  I just don't know.

I accept that you haven't had issues.  I would just ask you to accept that perhaps the issues that others have might have been real too.

[Swatter]

Yeah, but people need to be realistic about it.  I really don't want patching resources spent on trying to make a perfect AI.  A reasonable one?...sure.  But listening to people complain about the every aspect of the AI, especially claiming its doing unrealistic things, when its not necessarily unreasonable based on history, will not help.

The problem I see, even on this board, is general claims of the AI doing this or that, without any context whatsoever.  This is a very complex game and throwing out a general statement about too many amphibious landings without some details won't help get anything fixed.  How about year, alliances, enemies, national focus, aggressiveness, production, research, etc.  I am still not familiar with HOI4 enough to even know...can a save be posted so others can see settings?

I also think a big factor is people who played a lot of HOI2/3 coming and knowing just enough to be dangerous in HOI4.  They are assuming HOI4 is less of a game than its predecessors or the AI is broken.

And I, now and then, pull out an old Command HQ example from my days playing in Windows 3.1.  I used to play it to death and beat the AI a lot, but got killed a couple times when the AI loaded up a huge armada and attacked me through Africa up into Europe.  It was the first MP game I had ever owned and I got so pissed off at the game and AI that I convinced my buddy at work to hook up our PCs through serial ports and play against me.  In our first game, guess what he did?  Built a huge armada and attacked me through Africa.  He cleaned my clock.  When I asked him why he chose that strategy, he said he wanted to do something unconventional.  I gained a little more respect for the AI on that simple game.

My point is that when people demand an AI be fixed, I think they sometimes lose sight of the fact that you wouldn't want an AI that is perfect.  There are times that an AI is broken and needs fixing, but sometimes its a game mechanic that is actually broken.  For example, the marching through another country's territory bug.  That to me is actually a game play bug that manifested itself as a perceived AI bug.  In HOI4, I have yet to run into anything that makes me throw my hands up in the air and declare, "I can't believe the AI did that, no human would ever do that."  I have seen human players in Combat Mission and other players do some really stupid things.

Maybe as people play it more they become jaded because they can see patterns in AI play.  But I think if people had to play scenarios clean and not be allowed to replay them the AI would appear to be a much better opponent.  The one all encompassing fault with AIs right now is that they don't learn from replaying a scenario.

I think the AI is fairly decent, but it has some serious problems. From my own experience it has problems deciding which front is the most serious threat. Playing as the Soviets, I saw the German AI not leave a single division on its front with me for over a year. I could have driven to Berlin in a week, but I help myself back. It eventually saw the threat, but this was no designed behavior. Later in the game, the AI left one division to defend Japan.

It will be fixed, I am confident.

[mikeck]

Yeah, that's a pretty legitimate AI concern. Leaving a land border with a potential adversary completely undefended. I her having to alter my plans to "help" the AI but sometimes you have to. In my current Stellaris game, I'm in the "steamroll" phase. No AI empire could keep up from the start...so, I helped them
By gifting them "research agreements and ridiculous anounts of minerals.

[Swatter]

What I want to see is human players being able to provide detailed advice to allies via the battle planner tools. That would probably be the easiest way to fix the AI in the near and medium term.

[Ian C]

If anyone wants to see what the AI does exactly, bring up the console window and type observe to see the AI run your nation. Also checking the 'see allies warplans' will provide greater insight.


http://www.hoi4wiki.com/Console_commands

[RyanE]

I have been using that method for quite a while to learn the game.  For a week, I let the AI go at by itself and watched what it did.  It is actually very fascinating.

[SirAndrewD]

If anyone wants to see what the AI does exactly, bring up the console window and type observe to see the AI run your nation. Also checking the 'see allies warplans' will provide greater insight.


http://www.hoi4wiki.com/Console_commands

Very much so. 

Try switching around to different countries with this feature enabled as well.  This is how I saw Germany constantly keeping over 2/3 of its army in constant confused back and forth strategic redeployment between the active western and inactive/at peace Eastern front.