-
We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Yeah but that's a really false comparison. The AI in Rome has a much wider field to roam in because the history isn't as well known. The equivalent of "Finland invading Japan" or "Canada joins the Axis" happen all the time but we don't care because they're not obvious. For an AI to be challenging, at least to me, it has to be unpredictable. In HOI3, that same unpredictability is generally labeled as ahistorical and people don't like it (I wouldn't). I think that's part of the problem. The AI of HOI3 has to be challenging but still stay within a much smaller confine of historical reality.
It seems you have misinterpreted my comparison. I was not comparing EU:R's AI to HOI3, I was comparing it to EU:3 which has a very similar battle system and movement system. I was comparing HOI3 to HOI2, because that is the relevant comparison. The point is the AI in EU:R was not a step backwards from EU:3 whereas the AI in HOI3 was a step back from HOI2. Moreover, the same type of historical inaccuracies and accuracies are present in both HOI2 and HOI3--they both have (or should have) smaller confines of historical reality than EU:3 or R, but that isn't relevant to the discussion as I wasn't comparing the two different franchises against each other.
In other words, the AI in EU:R (at release) was worthy because it was competent in comparison to the AI in EU:3 (at the time of EU:R's release). On the other hand, HOI3 AI (at release and currently) is not nearly as competent as the HOI2 AI (at HOI3's release). Game AI needs to evolve (or at least not devolve) in subsequent iterations and not just features.
It is as certain as day that 90% of those who reviewed Empire: total war as good played it for 3 hours (after 3 hours it starts to be clear how stupid everything really is). The rest were bought by CA/SEGA.
I dunno about Hoi3 though but I wouldn't be surprised if the same as above happened to Hoi3.
What AI are you talking about?
If you're talking about diplomatic AI, there really isn't anything broken in the HOI3 model. Sure, it could be encouraged to act more reasonably or try to plan ahead a little more, but it's on par with EU3's diplomacy for the most part.
If you mean combat AI, well HOI3's AI would probably beat the pulp out of the EU3 AI in a game of EU3.
It seems every war I read about in the EU3 section deals with a single human country battling 2 or 3 larger nations at once and managing to win all of those wars due to the ineptness of the AI when dealing with its armies and coordinating attacks.
What needs tweaking in HOI3 is research, production, resource management, and manpower management.
Sure, the AI is not perfect either, but with improvements in those areas, it will at least have the proper tools necessary to fight a war.
Old enough to remember black and white, no patches and NO sequels either.
I bought Windows 286 1.0 just so I could play Rise and Decline of the Third Reich. Yep, the Avalon Hill game on the PC.
Dang! It sure was purty.
You could set up your beginning OOB and run through Poland and France like poop through a goose . . . but there was no AI (no AI, none) and it was single player and the opposing side never moved a chicklet.
To clarify I was talking about combat AI. And again, i'm not comparing HOI3 AI to EU:3, see my post above. The battle mechanics are very very different between the games and such a comparison is irrelevant.
Well the Rome AI is a very different animal for the HoI3 AI. Doomdark considers it to be one of its best works. It set ups regions, builds armies to defend them fights in these regions. If a region is out numbered it gets reiforced. With HoI3 you have front lines, which makes things ever so slightly more difficult. However from a testing perspective the HoI3 AI was one of the easiest. Set up your theatres and observe. Oh look the AI is not transporting units, better get that fixed. There were a lot of bugs with the AI that were quickly spotted and fixed thanks to the theatre system. Making HoI3's AI in some ways in the best shape of any release AI. However the complexity of World War II made it seem inferior to the EU:R's AI or EU3's 1.0 AI.
Is the creation of new theathre being looked into for 1.4? Currently it's very tricky to use and I think this also botches up the AI a lot. As an example, if you play USSR and take over Denmark, Germany AI will move disportionate amount of troops to Denmark region, leaving the Eastern Front very poorly defended.
Because I think currently the AI does excellent job when it can face you with right number of units, but it has trouble facing you in the first place.
Only administrators can do it, but they never do it because it is alot of work and it can screw up the forum-database.
Theatre creation is a tricky one, we ditched player defined theatres pretty early because of interface issues. There are rather a lot of provinces out there and it was going to be so much work for a player to define areas. So that kind of thing will require a lot of work on interface which isn't patch material I am afraid.
However the AI problem you are seeing it not 100% a problem. The AI has two fronts against the SU, Denmark and Poland. Denmark is the weaker of the two so the AI is trying to eliminate the weaker front to allow it to concentrate its forces. Now the idea isn't bad it just needs to work a bit more on the execution here.
Yes I can totally see how it is tricky. Maybe preset theatres that unlock as you occupy any province that is part of the preset theatre? Like Germany that started in 1936 campaign can unlock the Norwegen theatre by occupying any part of Denmark/Norway/Sweden/Finland?
And well after reading your reply I guess I forgot an important detail... which is that Germany made this extreme redeployment (by extreme I mean it left about 2 ~ 3 divisions max per province, while sending vast army to mere 2 province that share border with Denmark) while we were at peace. It sort of "overreacts" to new potential fronts by applying disportionate force (which is needed elsewhere). Perhaps this problem was addressed by solution to the superstack issue, which would be nice.
Preset theatres work for historical situations. The problem comes when there is an ahistorical situation. There is also the issue of well we know Germany needs a Nrowegian theatre, but what about other countries? We personally thought that static theatres could cause as many problems as they solved and also give the player who did not use theatres an advantage over the AI.
There is the superstack issue there, there is also the nasty AI habit of trying to win quickly by deploying overwhelming force. Not a bad idea in general, just needs fine tuning.
The AI really just needs to focus on solidifying its frontlines up to a certain degree, I mean each area has a specific frontage, so it's not like it really needs every unit it has at a single point, just enough to maximize the frontage it can cover.
Once it has achieved this, it should use extra units as reserves or contribute them to a point where it intends to make a "breakout".
That should help alleviate some problems as well as make it harder for the AI to be encircled if there is a break in its lines.
Once it has achieved this, it should use extra units as reserves or contribute them to a point where it intends to make a "breakout".
That should help alleviate some problems as well as make it harder for the AI to be encircled if there is a break in its lines.
The AI does try to do that, but also tries to take the initiative as well. It is trying to find the right balance. Make the AI too cautious and you end up with a dumb AI that sits around and does nothing because things are never quite right for it. At the moment it is too agressive and needs to asses things a bit better. However if we are going to err here I would prefer to do so on the side of agression.
I agree with this philosophy, but adequate testing on a variety of platforms means you need to keep the RC cooking for a sufficient amount of time. Everyone was complaining about graphical problems, for example, in the RCs, and they were never resolved because there was a rush to get something, anything out the door. As a result, I've had to tell people who ask about HOI3 that they shouldn't buy it yet, because its graphics are still in beta (flags, counters cut off, not being able to click on units, flashing bars in the corner when you click on a province, selection boxes slanting, etc). This is the type of problems people noticed right away, but because no one was responding to their cries, they just gave up trying to help you make a better product.
Naturally, the RC testing process would work better in a forum if there were a moderated first post listing all known bugs in the current RC (and likely the graphics processor type/windows version) so people could see that their complaint was listed on there and see the official response in the next RC. I personally felt rather insulted when I stood up to say that there were still many known problems, and I was told by Paradox that they weren't going to be fixed, despite the claim that you guys would "fix everything we can."
I'm not demanding perfection, but if you won't even fix the problems we take the time to tell you about, you'll never hear about the rest of the problems.