So, I've been spending pretty much all my gaming time the past couple weeks with HOI4. Thought I might share some observations about where things currently stand in the game's evolution.
Couple things right off the bat: overall, I love this game, warts and all. I think it's the best in the series, and I loved HOI3 (and HOI2 as well). I agree with most of the design decisions Podcat and his team made. I'm not convinced by arguments that HOI4 represents a 'dumbing down' of the series, at all. I say that as someone who loved HOI3's OOB system and still do miss that in HOI4. Nevertheless, I find 4 superior to 3 in most ways.
That said, HOI4 still suffers from some serious shortcoming.
1. Right now, AI division spam is HOI4's single biggest problem. More than any other factor, the AI's knack for churning out hundreds upon hundreds of crappy, unsupported infantry divisions is the biggest obstacle to having a good, challenging enjoyable game. It makes the game merely frustrating. First image below shows the German army toward the end of my campaign playing as USA, mid-1945, after years of being ground down and strat. bombed into oblivion. Yep, it's still fielding 612 divisions. That's down from estimates of over 1,000 divisions when the Germans were still strong. Second image shows the blob of divisions Germany is just fielding along its southern Italian front. These divisions are like a swarm of ants. At this stage they have almost no manpower, but they constantly swarm my much better but much fewer divisions, making progress against them tedious at best.
Pdox has got to solve this problem. Like I say, for me, this is thing that more than anything else ruins the HOI4 experience.
2. AI is still terrible at managing production -- tanks in particular
Images show that Germany has a grand total of a single Panzer IIc -- one -- in its entire army of over 6M men. I also never saw the German AI build anything other than light tanks -- no mediums and definitely no heavies. Still, this is better than the mighty Soviet army, which has exactly 0 tanks in its arsenal. In 1945.
I'd say this is a bit... immersion killing.
3. Japan and Pacific theater remain largely inert
Playing as USA, I had far more interesting naval battles against the mighty Italian navy than I did against Japan. Early in the war, Japan made some good headway against China, but never captured key goals like Hong Kong. And this with boosting the Japan AI in the game settings by +50%. By 1945, the Pacific has been a static theater for years. Japan managed to take some territory in New Guinea, but that's it. Never invaded Philippines nor Dutch East Indies.
These are my three biggest complaints at the moment. So, takeaway -- HOI4 remains a very flawed gem. I still have hopes that Pdox can solve some of these issues, but they really need to focus on the shoddy AI issues and worry less about offering underwhelming DLC like "Together for Freedom."
Agree with most of that. I've had more than my money worth in terms of hours out of Hoi4- but it needs some serious AI love.
Have you tried Expert AI 2.0 mod. That goes some way to giving AI better division templates and capping their spamming. Makes for a better game.
Is this with the latest DLC included or base game just patched up?
What I find both incomprehensible and sad is that this is the fourth iteration of HOI and the AI is as bad as the bugged disaster of version 1.
They have solved AI problems multiple times, but every new version is like some reincarnation parable - no sentience, knowledge of the last version, or avoiding the previous one's mistakes.
Maybe HOI 5 will come back as a fruitbat.
Quote from: fabius on January 21, 2017, 03:46:37 PM
Have you tried Expert AI 2.0 mod. That goes some way to giving AI better division templates and capping their spamming. Makes for a better game.
I tried Expert AI a while back but wasn't too thrilled with the experience. AI-controlled majors seemed to be churning out aircraft in huge numbers, almost to the exclusion of all else. But maybe it's time to give it another try, because the vanilla AI simply isn't able to deliver an enjoyable experience at the moment. Capping the spamming and giving better div. templates would alone be worth the price of admission.
I got the results I mentioned above playing the latest version -- 1.3.2 -- but without the Together for Victory DLC, which I still have not bought because it frankly seems to offer very little for its $15 pricetag. Still, I don't think TfV would have changed the results I got in my current campaign in any meaningful way.
Yeah I would say Expert 2.0 is worth another look. I've had my most intense 3 months (game time) against the Russians with it.
Once Russia broke though- it was broke completely with a huge encirclement. But was fun while it lasted. I even lost provinces and nearly a whole state, and Finland before grinding counter attacks could push the AI back.
Together for Victory- agreed. I held off but picked it up on impulse. Not missing game changing stuff, but what's in there is better flavour for puppets than I expected. Easily a wait for sale dlc though.
Oddly, the game, for me, was highly enjoyable with the beta-1.3.1 release.
The Historical Divisions mod really helped, but it's not been recently updated.
I think all your points are valid and I think the game itself isn't broke, but it needs a real fine-tuning in regards to Historical accuracy.
However, from the evidence on their forums, including from Podcat who replied about the Division Spam, he thinks the AI spamming is not broke.
Paradox have no desire to make the game perform in the historical ballpark and have said so.
This is the eternal gripe of all HOI hardcore fans since HOI 1. "We want more realism". Paradox: "No".
Supply and demand. Maybe the penny will drop one day.
If the Division spam is the way for the AI to beat a human then that's good right? It adds to the challenge. Maybe the human player should do it too. ;D
While I'm not in the camp of folks who think HOI should faithfully model ww2 'as it actually happened' -- my preference is certainly for the game to be more historically plausible than not. I mean, HOI already goes to terrific lengths to do exactly that -- giving you realistic (for game purposes, realistic enough) equipment, vehicles, technologies, militaries, political leadership, etc. etc. I'm all for a game that says 'here's roughly what the geopolitical state of affairs was in 1936 or '39 -- let's see how it plays out'. That's HOI's strength -- if I want truer fidelity to historical realism, I play a Grigsby title or equivalent.
But it's hard to escape the conclusion that division spam is simply a cover for inadequate AI. Give the AI many, many more pieces than historical so that it can cover its bases through a simple brute-force approach. I can accept that up to an extent -- but right now I think HOI is abusing this mechanic. It's simply ludicrous for nations to be pumping out 1000+ divisions. Even minor powers like Hungary or Spain churn out hundreds upon hundreds of divisions. It also has the unfortunate side-effect of creating noticeable lag -- so many divisions on the map, even powerful PCs start to labor under the load.
I'm OK with HOI presenting us with a WW2 that fields more - even many more - divisions than historical. But right now it's out of hand. It needs to be dialed back to numbers not quite so ridiculous.
For me early HoI3 was more towards the 'Historical Accuracy' side of things, and that was just a rail-road. Events were used as a way of shoving history as-was down the player's throats regardless of actions taken.
I much prefer the sandbox nature of it, although perhaps they should give up trying to offer the middle ground - don't think it really works and it pleases no-one.
They're probably being glib regards to the AI - "broken" is a word that's easy to throw around... it's probably not broken, it's doing everything it's supposed to within the parameters that's been given to it, it's just those parameters need better tuning.
Division spam is probably the major reasons I've held off on purchasing this title just yet.
As others have alluded, it may be an appropriate response for the AI, given what it's got to work with. But it destroys the immersion for me. In doing so, it severs the historical affinity that WW II games rely on to draw in their core audience.
The notion that you can conduct a war by breaking your army up into battalion- (or regiment-)sized units and throwing them all into the blender just seems kind of silly. Devs who defend that too seriously are spending too much time in the bowels of their code and not enough time focused on the user experience.
I'm still pulling for this one. I haven't wanted a title to succeed so badly since Eador.
I think the issue for me is that the game appears to model the correct Manpower, yet produces ahistorical amounts of Divisions. I think partly this is down to the incomplete Division design of the base templates the AI uses.
I don't like a railroad ahead of me and I don't want to play through a re-run of the war, and I'm overjoyed when the AI throws me a different path, because it opens new strategic and tactical opportunities BUT, and this is the crux of the matter for me, I'd like the game to use reasonable authentic WW2 behaviour, because when it doesn't, immersion breaks.
That's all.
I see division spam and infantry focused production as a response to the game mechanics instead of poor AI. If you let a machine learning bot develop a strategy, it might think of doing division spam too. Game mechanics that doesn't encourage division spam might be the answer (like less micromanagement and formation building) but that will be a totally different game. There's no turning back.
So yeah, getting historical accuracy right for a game like HOI4 would be tough. Encouraging a historical AI may make it easy for the player since the player doesn't need to follow history and perform these game mechanics exploits. Sad to say, I guess we just have to ignore this title. But who knows, Paradox might be able to fix the situation and find a balance.
Grigsby's War In The East Soviet AI plays like that as well - the only way it has to slow down a human player is to carpet pretty much the front up to a depth of three or more hexes, all the way from the Baltic to the Caucasus. Some referred to it as the "Red Meat Carpet" and rightly so.
Sounds like Paradox is borrowing some moves from that playbook.
HOI has been for me a complete waste of money: no small scenarios, and no chance to setup an MP game with humans playing the majors. It is a game I want to like, but things go so off rails that The Man on the High Castle sounds like a very conservative counterfactual.
I prefer to play Stellaris (or EU 4) instead, very much as I do prefer TW Warhammer to pretty much every TW other game.
EDIT: It was the "Red Meat Carpet"