Can someone help me with updated stats? When I started playing this game shortly after release I got good at division design and naval comp and other elements. Now with updates and dlc I'm bad at it. I can't into design stats for divisions, tanks, ships, and planes. Also general vanilla Livers of Tungsten thread
Southeast Asia when?
Just look up any plane guide in the past 2 years. Thats all you need. The rest can be whatever basically. Just have good fighters and bombers and you win the game
How do people find this game fun? You just look up a division template and the AI can't do shit about it. Grats you won WWII, or at least made it 50% easier for you by doing this one step.Retarded game design, but the power fantasy is the purpose of these games. The more capable the AI is, the less this shit sells because people stop playing if they actually lose a war. The AI isn't even intended to be able to use the systems the player has. It would just eat up development time, too.Total War is the same, the recent AI developer blogpost confirmed it. There's no incentive to develop a good AI for games like these, quite the opposite.
>>1807119The plight of paradox becoming a "balanced for multiplayer" development studio. The AI is just there to provide a stage upon which people play against one another and is not intended to be competitive.
>>1807191are you implying other paracocks games have good AI and hoi4 is only bad cause of mp? LOL
>>1807119You have to understand that this game is incredibly complicated. HoI4 is easily one of the most complex and nuanced games ever created and since launch it has only become more and more intricate.To make an AI which would be able to effectively play HoI4 would be like making an AI for NASA. A small studio like Paradox does not have the resources to invest into such an effort.
>>1807280Nor do they have any financial incentive to do so, because it would only slow down the DLC pump.It wouldn't be that hard to make the AI use good division templates or to design a game that doesn't make the AI progress through focus trees that make it randomly declare war on countries without any other considerations.
>>1807119>Total War is the same, the recent AI developer blogpost confirmed it.I can't find a recent tw blogpost that talks about AI at all.But you're wrong. AI being better is extremely in demand across the entire genre, ai mods are the most downloaded mods for virtually every strategy game in existence. People want to be able to choose their difficulty, if the ai is too smart they can always just play on an easier setting.The issue is that these developers literally can't make a better ai and can't afford to onboard talent that could. hoi4s ai doesn't make shit divisions on purpose because the devs want the player to win. They just can't be fucked to keep changing the ais templates when their division balance is always in flux. At this point, old parts of yhe game's architecture may not even be safe to meddle with for fear of breaking something or encountering compatibility issues with new tools and an 8 year old codebase. They're still trickling on fixes for bugged tank/plane templates from BBA."Ai can't be good" is a laymen cope
>>1807284https://medium.com/@julianmckinlay/total-war-rome-ii-and-creative-assembly-my-statement-ten-years-on-d964f65b0a8f>AI being better is extremely in demand across the entire genrePlayers say they want this, but developers have no actual incentives to provide it. The average player wants a power fantasy, and is liable to even stop playing if they lose a war. There are shitty AI mods for Total War, but are these a considerable reason for the success of those games? They would be popular anyway, and so would Paradox games.Making sure the AI can actually use all the DLC bloat features they keep pumping out would just increase development times considerably, for no financial benefit. It's just not a worthwhile consideration for them.
>>1807280Only a tiny fraction of the playerbase has ever even played the intended ironman mode before. Why bother spending real effort to placate the minority hardcore players when they can just create a bunch of stupid achievements like Liberian world conquest instead to keep them busy
>>1807119I used to enjoy the game but rarely do now. The game has only gotten worse over time as more mechanics being added (but no ai tweaks to understand them)
>>1807307>my evidence is a hitpiece by a disgruntled former employeek>games have no financial incentive to provide users what they wantThis kind of attitude basically sank CA as a company, and they're treading water now only because they hard pivoted to pandering to their users and giving them exactly what they ask for. This is not the supporting evidence you seem to think it is.
>>1807329They're not making the AI any better.I don't know what exactly you're arguing against here. You said they can't afford to make better AI, which is basically what I'm saying.The average consumer also doesn't actually want a competent AI that they could lose against, even if they think that they do. Otherwise these games wouldn't be popular the way they are, and the developers would be forced to provide good AI.
>>1807370I'm telling you that your generalizations about what consumers want is completely wrong and based on nothing except a fantasy that you're denied the things you want because a mob of lesser men supercede you.The article you linked describes a former Ai developer describing the systematic mismanagement within a company with a strict hierarchy in which incompetent and ill-informed auteurs were given near complete freedom to disregard feedback both within and without their company, leading to decades of design in a vacuum that didn't reflect the reality of either what the consumers actually wanted nor what more competent developers were telling them they could and couldn't do or needed to do. The author of this article outright accuses a production lead of misunderstanding the nature of AI so completely that they genuinely believed the code could just 'figure out' how to play the game without any additional development from the AI programmers.But that aside, failure states exist in all games for a very good reason. A failure state exists to reinforce fun and engaging play patterns and encourage meaningful engagement with game mechanics. An AI in a strategy game exists for this purpose. It exists to be strong enough that the player needs to engage with the mechanics to overcome it--not to be an equal or superior substitute for a human adversary in some simulacra of multiplayer. The AI needs to be smart enough that you need to consider and plan for unpredictable eventualities and explore the game's mechanics. Failure states don't exist to make the player fail and be stuck and give up, they exist to be clear, understandable and avoidable conditions that structure play.The issue isn't that nobody wants good AI that can do these things, but that out of touch and incompetent production leads don't want to pay for it--or in some cases don't even understand what it is.
>>1807119Install a mod that gives the AI competitive templates.
>>1807119>>1807370I don't know what you guys are talking about. I lose against the AI every single time. That's why I made the thread. I used to be able to figure it out and be just fine. Now I can't figure out what the magic combination of stuff is to win. I'm afraid I'll just have to watch 90 hours of youtube but I was hoping someone here had a good guide.
HOI4 with good AI sound like a WW1 nightmareI'd be nice if they used the designers, but them being intelligent would make them never push
>>1807454"good" ai in the game design sense doesn't just mean the Ai is smart and makes all the correct decisions, but rather that it's designed in such a way that it encourages the player to use all the mechanics available to overcome it, and is always beatable given that those mechanics are used well.The problem with the AI in HOI4 is not that it should be making the perfect decisions every time and playing like a meta multiplayer tournyfag or something. It's that it makes so many wrong choices, or fails so hard to encourage interacting with mechanics that all you need to do to beat it is use meta templates and battleplan forward. If the game can basically play itself then the opposition is failing to perform the necessary role of encouraging you to engage with the 'strategic' elements of a strategy game.To put forward an example:Imagine if the AI made a small number of templates with extremely high hard attack/piercing used specifically to counter tanks, and the AI was programmed to position these divisions to respond to attacks by the player's tank divisions. Now concealing the movement of your tank divisions to achieve surprise would be a way to overcome this and achieve localized success, and your tank offensives would have a functional time limit before the AI's counter could respond and contain it. You may even seek to outmaneuver these divisions, baiting them to engage your tanks and then pinning and encircling them with infantry to neutralize the threat. This would be the AI working not to make the game more difficult, but to encourage you to engage with the strategic decisionmaking the game is designed around and rewarding you for doing so.
>>1807449Infantry divisions are simple. They just exist to hold the line. Poverty divisions are just doing to be ~20 width with support arty and AA and nothing else. If you can spare the industry you can add line artillery for more killing power, but these really just exist to sit entrenched on the frontline and not die. Your #1 industrial priority is getting enough infantry equipment that you can actually mobilize your population to hold a front. Support companies and extra arty are luxuries that you add if you can spare them or have nothing else to put industry towardsAir is your second priority.Most of the work is done by air. Fighters want the light frame and as many HMGs as you can fit. If you have a good source of rubber, give them self-sealing fueltanks. Otherwise armour plates. Then a bit of extra fuel for range. This will outclass the AI's plane designs so hard you'll have like a 5-1 kdCAS is also simple. Light frame with as many bomb locks as you can fit (or heavy bomb locks, once researched) and fuel for range.CAS does all the work in ground battles, it will be 90% of your damage. CAS is king. Fighters exist to enable CAS and protect you from enemy CAS. Your divisions can literally be worthless dogshit and you'll still win wars just by having enough CAS. The only reason air isn't priority 1 is that you need SOMETHING on the ground to put into combat for your air to do its work, but that something on the ground doesn't need to be good.The goal should be to research the 1940 light air frame and 1940 engine BEFORE 1940. This will be your main aiframe for 90% of the game so rush it and build a tonne of it. Focuses can provide research bonuses to help you get them.3rd priority are tanks. Mediums are the most cost-effective. The AI can't really do tanks so you don't need to worry about hard attack or countering enemy vehicles. Stack Soft Attack and Breakthrough while keeping costs low and reliability high.
>>1807494Overflow post:Tank divisions should be around 50/50 tanks and mot or mech, around 35 combat width and roughly 8km/h speed (double the movement speed of inf). You want around 35 org give or take so you don't run out too fast while moving but you can fudge around with the ratio a bit to get there. More tanks = more stats but less org, more mech = more org but less stats. 35ish width will ensure you fill plains and hill tiles with 2 divisions, which is the only place tanks should be fighting. Use pairs of them to surgically break through where there's good terrain, and then use their speed to encircle enemy divisions on the frontline and/or quickly capture supply hubs while your shitter infantry flow in to hold the new tiles you take and kill the encirclement. Once again, your CAS will do most of the real work, and you don't need many tank divisions.Navy can basically be ignored. Naval powers like the UK, US, Japan start with enough that they can win the naval war without any extra ships or player designs, because the AI is super shit at using navy. If you don't start with a navy the only reason to build one is to invade the US, UK or Japan but you can often just bypass needing to fight the naval war entirely because the AI is bad at protecting their home shores. For example, you can paradrop the UK from France, capture a port and just stream units in while the AI's fleet is too busy elsewhere to raid, and then once the UK capitulates you can just claim Canada in the peace conference and invade the US from there without ever fighting their navy.
>>1807449Literally just make infantry divisions with artillery and even if you're playing as fucking Uruguay it should not be possible to actually lose a war to the AI.
>>1807494>>1807500Thanks this is exactly what I was looking for. You're a true hero
>>1807459>"good" ai in the game design sense doesn't just mean the Ai is smart and makes all the correct decisions, but rather that it's designed in such a way that it encourages the player to use all the mechanics available to overcome it100%. I wish more people would play battlebrothers, the AI in that game is extremely good but also themathic
I NEED HELP WITH PARANIGGERS!
>>1807119>the power fantasy is the purpose of these gamesI disagree. WA is much more fun than vanilla, and if power fantasy was what people cared about they'd play with cheats which is just boring.>>1807500>because the AI is super shit at using navyUnironically how would you fix navy?
>>1807273How do you get that out of what I said? The whole balanced for multiplayer meme started with stellaris afterall and spread out to their other games.
