Hard Difficulty Settings

[caffran]

I know there's a combat malus for the player, and the AI gets various boosts, but is that the only effect? the AI itself doesn't do anything differently, we just get less to counter it with correct?

And if playing as UK, the AI bonus would be the same for Germany, AND America/USSR?

 

[Wraith11B]

Yes. No advanced AI changes as the difficulty changes, unfortunately.

 

[Alex Kernel]

I know there's a combat malus for the player, and the AI gets various boosts, but is that the only effect? the AI itself doesn't do anything differently, we just get less to counter it with correct?

And if playing as UK, the AI bonus would be the same for Germany, AND America/USSR?

Wiki gives this info about the Hard Difficulty:
Manpower: +0%
Revolt Risk: -0%
Industrial Capacity: -10%
Resources: -10%
Supply Throughput: -25%
Naval Base Efficiency: -25%
Combat Efficiency: -20%
AI Industrial Capacity: +25%
AI Resources: +25%
AI Supply Throughput: +25%
AI Naval Base Efficiency: +25%.

There is no mention of other penalties for the human player.

In essence, the AI reacts exactly the same, you just do things with less against more. The problem of the all the historical strategy games nowadays is that despite phenomenal computing power the AI cannot understand the world following our adaptive logic. The AI simply plays beyond its comfort zone, and that is why the developers hardcode its behaviour. In other words, the artificial intelligence does not understand what it has to do and why.

When our worlds will come closer, we'll have better AI playing partners.

 

[themousemaster]

In essence, the AI reacts exactly the same, you just do things with less against more. The problem of the all the historical strategy games nowadays is that despite phenomenal computing power the AI cannot understand the world following our adaptive logic. The AI simply plays beyond its comfort zone, and that is why the developers hardcode its behaviour. In other words, the artificial intelligence does not understand what it has to do and why.

When our worlds will come closer, we'll have better AI playing partners.

While I do not disagree with this statement, do you remember way back in the early days of HOI3? The AI was reacting to the situation of the world at large, and doing things that were totally a-historical from diplomatic standpoints, but made sense if you looked at the game-state.

The almost overwhelming response of the playerbase to the AI being able to think outside the historical box was "OMG this game is terrible! it's not a WW2 game! Germany HAS to declare war on Poland, then France, then USSR, or else I'm getting a refund!"

I believe the programming power of reactive AI is at the point where what you say is possible; but whether the people who see the HOI series as anything other than a "WW2 simulator" --are willing to accept-- historical deviation, in numbers large enough to make a game of this incredible depth financially feasible, I'm not sure.

 

[Alex Kernel]

Honestly @themousemaster , I would love to be able to play both historical, and ahistorical WWII games. Wouldn't it be possible to choose in the game's Menu via a checkbox whether to start a game closely tied to history like we have now, or one with an AI freed of historical strings (e.g. an "out of the historical box" game)?

After 2 years of playing a WWII simulation it becomes a little repetitive and annoying knowing in advance what will happen, and therefore taking always an already struggling AI on the wrong foot.

An even better option would be to have a built in AI able to learn from its mistakes every time you're beating him, and trying accordingly to change tactics in the future, or better, during the round itself. Does anyone know how far are we from such an AI in historical games?

 

[themousemaster]

You don't need to convince me; I advocated for it way back when the decision was made to "force histor-ocity".

 

[Kovax]

While I do not disagree with this statement, do you remember way back in the early days of HOI3? The AI was reacting to the situation of the world at large, and doing things that were totally a-historical from diplomatic standpoints, but made sense if you looked at the game-state.

The almost overwhelming response of the playerbase to the AI being able to think outside the historical box was "OMG this game is terrible! it's not a WW2 game! Germany HAS to declare war on Poland, then France, then USSR, or else I'm getting a refund!"

I believe the programming power of reactive AI is at the point where what you say is possible; but whether the people who see the HOI series as anything other than a "WW2 simulator" --are willing to accept-- historical deviation, in numbers large enough to make a game of this incredible depth financially feasible, I'm not sure.

I've got to disagree with this. The AI in the initial release of HOI3 did some pretty stupid stuff from ANY perspective. Japan FREQUENTLY declaring war on Finland was bad enough, but at least comprehensible (the over-the-pole distance puts them within reach). Tibet declaring war on Paraguay really crossed the line into cyber-stupidity. The AI spammed offers of alliance almost at random, which had a modified random chance of acceptance, and kept repeating them until they happened, which created insane coalitions of mismatched countries with no interests or ideology in common. The AI did use some advanced logic, but the random element was FAR too high, allowing for blatantly absurd situations. If anything can go wrong, it will, and HOI3's initial release effectively had no restraints on the AI, so when crazy events did happen, there was no "sanity check" to reject them.

A similar system, driven primarily by cores, political similarities, and proximity of mutual threats instead of naked opportunity, could work, but allowing purely random actions was almost a guaranteed fail. As the saying goes, "Just because you CAN doesn't mean you SHOULD." Just because the RNG says to do so doesn't make it a good idea.

I liked where the game was INTENDED to go, it just failed to get there. It wasn't just a matter of not doing what the players wanted, it utterly failed to make sense.

 

[phantomrider]

I've got to disagree with this. The AI in the initial release of HOI3 did some pretty stupid stuff from ANY perspective. Japan FREQUENTLY declaring war on Finland was bad enough, but at least comprehensible (the over-the-pole distance puts them within reach). Tibet declaring war on Paraguay really crossed the line into cyber-stupidity. The AI spammed offers of alliance almost at random, which had a modified random chance of acceptance, and kept repeating them until they happened, which created insane coalitions of mismatched countries with no interests or ideology in common. The AI did use some advanced logic, but the random element was FAR too high, allowing for blatantly absurd situations. If anything can go wrong, it will, and HOI3's initial release effectively had no restraints on the AI, so when crazy events did happen, there was no "sanity check" to reject them.

A similar system, driven primarily by cores, political similarities, and proximity of mutual threats instead of naked opportunity, could work, but allowing purely random actions was almost a guaranteed fail. As the saying goes, "Just because you CAN doesn't mean you SHOULD." Just because the RNG says to do so doesn't make it a good idea.

I liked where the game was INTENDED to go, it just failed to get there. It wasn't just a matter of not doing what the players wanted, it utterly failed to make sense.

I definitely agree with this. In the initial games I played (as Germany) New Zealand invaded East Prussia in 1939 and it was almost impossible to intercept its fleet and supply to the beachhead using the entire German navy and lots of air and a long vulnerability of getting through the North Sea, into the Baltic and then almost to the end of the Baltic. Granted the Luftwaffe hadn't spent much time worrying about anti ship action but they really did well in early 1941 against the Brits in the Med in the Cyprus campaign. The AI did not recognize the real world impossibility of this kind of adventure in the real world and in theory this is the fault of the programer. Even in the last version of HOI3 the "depth" of thought required to plan campaigns by the AI is weak -- For example if you (as Germany) don't garrison or defend ports the AI (as the Allies) will invade and then you can play whack a mole for serial invasions and knock off 8 to 10 allied divisions at a time and the Allies never will develop enough of a reserve to make a major invasion successful. (On the other hand as a non programer I can understand the complexity of the problem and in general understand the difficulty).

Also I want to defend the "historical" aspects of a World War II game. Ideally it would be an excellent simulator of WWII era land sea and air combat, it would allow the player to test different changes in strategy, tactics, equipment and unit compositions at the simulator level. In addition, at the nation level it allow each nation the player plays to see if they could "change" history by doing things differently than the "real world" nation did. In order to take this second part seriously the game need to proceed close in a close to historical way up until whatever nation you play enters into the war (which was driven historically by Germany and to a lessor extent Japan). Playing the USA for example is not much fun in the game where you spend 5 years getting ready for war and when it comes the whole strategic picture is so far from historical that it makes no sense (i.e. Germany already beaten by the Soviets, or no Japan attack). In addition, both Britain and the USA are hampered by the lack of ability to form a real alliance with common control of armies (Britain doubly so because because the commonwealth nations also behave as independent forces in parallel the to British). (Germany has the same problem with Italy and its eastern front Allies but this is less critical). I really don't want a game where in a period of 10 years a small nation can conquer Europe or knock out several major powers by doing things "right". If you want that kind of outcome you need something like Europa Universalis where you have multiple centuries to turn small nations into a major power.

 

[themousemaster]

Kovax, I do not disagree with your post, but I do wish to emphasize three things:

1) When I say "early life of HOI3", I mean up through... roughly some point around SF expansion's release, but definitely well before FtM. Yes, I do remember v1.0 myself, and the total absurdity of some things.

2) While the "just because you can" argument holds merit, I, as part of my "base" National China strategy, declare war on Haiti/Turkey, almost every time, just to "hold onto the good laws". The AI was just doing what the min-maxing player already did themselves; the AI doing things the players did is the epitome of fair, and forbidding them from doing things without blocking the player from doing the same, in my opinion, is even worse then when they DO do them.

3) I wasn't specifically saying that I agreed with the "general response" to the diplomatic system; in fact, I specifically agree that having that be a "checkbox option" where the AIs have player-level diplomatic freedom vs being forced history is a great idea. I was just pointing out the response in the official forums, other forums, chat rooms, et. al. that I experienced early in HOI3's life; that if it so much as had the GALL to not force the AI players into historical decisions, they would flat out refuse to buy the game/demand a refund.




Phantomrider:

I disagree with your second paragraph at a fundamental level; if the other countries proceed "as normal" while you do totally different things, and they don't get to react until it is too late, that's not seeing if you can do things differently, that's forcing the hand of bad AI.

For example, a common argument I hear is that if Germany hadn't "wasted it's effort" on surface aships and instead tripled it's Sub usage with those resources, then they would have won the war before the US could save Britain.

The thing is, the US and britain weren't sitting around waiting for Germany to attack. UK's entire naval strategy involved hemming in the German surface fleet (along with protecting it's interests in the far east, though we know how that went). If Germany didn't HAVE a surface fleet, and if the UK (and USA) had seen Germany dismantling large ports and building tons of sub pens in the pre-war years, they would have slashed their capital ships budget, and redirected HEAVILY into anti-sub tech and tactics.

But "forcing historical" stops the UK/USA from doing that; and I'm sure I don't need to tell you, in the game as it is right now, If a GER player DOES go full-bore into subs, he very much can destroy the Allies (sans USA) with ease.

that doesn't show you can have done better than history, it shows that you can abuse someone with a forced handicap.

 

[phantomrider]

Kovax, I do not disagree with your post, but I do wish to emphasize three things:

1) When I say "early life of HOI3", I mean up through... roughly some point around SF expansion's release, but definitely well before FtM. Yes, I do remember v1.0 myself, and the total absurdity of some things.

2) While the "just because you can" argument holds merit, I, as part of my "base" National China strategy, declare war on Haiti/Turkey, almost every time, just to "hold onto the good laws". The AI was just doing what the min-maxing player already did themselves; the AI doing things the players did is the epitome of fair, and forbidding them from doing things without blocking the player from doing the same, in my opinion, is even worse then when they DO do them.

3) I wasn't specifically saying that I agreed with the "general response" to the diplomatic system; in fact, I specifically agree that having that be a "checkbox option" where the AIs have player-level diplomatic freedom vs being forced history is a great idea. I was just pointing out the response in the official forums, other forums, chat rooms, et. al. that I experienced early in HOI3's life; that if it so much as had the GALL to not force the AI players into historical decisions, they would flat out refuse to buy the game/demand a refund.




Phantomrider:

I disagree with your second paragraph at a fundamental level; if the other countries proceed "as normal" while you do totally different things, and they don't get to react until it is too late, that's not seeing if you can do things differently, that's forcing the hand of bad AI.

For example, a common argument I hear is that if Germany hadn't "wasted it's effort" on surface aships and instead tripled it's Sub usage with those resources, then they would have won the war before the US could save Britain.

The thing is, the US and britain weren't sitting around waiting for Germany to attack. UK's entire naval strategy involved hemming in the German surface fleet (along with protecting it's interests in the far east, though we know how that went). If Germany didn't HAVE a surface fleet, and if the UK (and USA) had seen Germany dismantling large ports and building tons of sub pens in the pre-war years, they would have slashed their capital ships budget, and redirected HEAVILY into anti-sub tech and tactics.

But "forcing historical" stops the UK/USA from doing that; and I'm sure I don't need to tell you, in the game as it is right now, If a GER player DOES go full-bore into subs, he very much can destroy the Allies (sans USA) with ease.

that doesn't show you can have done better than history, it shows that you can abuse someone with a forced handicap.

This is a reasonable criticism. Yes everyone reacts to what their potential enemies does in terms of strategy. In addition, everyone in the pre WWII world placed too much emphasis on the importance of big gun surface fleets (when aircraft carriers/anti-naval air and subs were to be the decisive weapons of war) with minor deviations in Japan and the US. Ideally if you and are were smart enough with AI design we could also build in a way of allowing the AI to "react" to developments around it as people do in the real world but understanding why people do what we do and whether what we do is the "most appropriate" thing to do in that situation is quite complicated as is writing a program for a computer that reflects it. Chaos theory (things really are too complicated to be fully understood) has a certain attraction in situations like this.

 

[Wraith11B]

My concern is that it shouldn't be terribly difficult to program. We identify "competitors" (ie, the "peer" and "near peer" of yesteryear; in order to keep the UK from caring about--for example--Romania) and pass in certain arguments of a ratio between their various ship types. Another routine would be to work on updating the fleet (taking into account the age of the vessels in each type class) for replacements, and then the combination of the two with an assumed doctrine path yields how the AI (and the player for a bit) should produce ships.

 

[Alex Kernel]

My concern is that it shouldn't be terribly difficult to program. We identify "competitors" (ie, the "peer" and "near peer" of yesteryear; in order to keep the UK from caring about--for example--Romania) and pass in certain arguments of a ratio between their various ship types. Another routine would be to work on updating the fleet (taking into account the age of the vessels in each type class) for replacements, and then the combination of the two with an assumed doctrine path yields how the AI (and the player for a bit) should produce ships.

Do you think there is any chance Paradox will return to Hearts of Iron 3, or that they would sell it to other developers willing to introduce new DLCs? HOI4 lacks the complexity and depth of HOI3, and there is clearly a niche of players who want more of a historical simulation (like HOI3) than a sandbox for casual players (HOI4). Otherwise, there wouldn't be here a continuous community keeping after more than 10 years the HOI3 threads alive and kicking.

For the foreseeable future, Paradox seems poised to continue with the HOI4 project. No HOI5 is in sight either. It would be nice to have a deep HOI3 (further developed with new DLCs) or HOI5 game for armchairs generals ;-) loving an immersive historical experience, and the sandbox version HOI4 for the casual strategy gamer, who abhors spending hours into micromanagement matters, at the same time.

 

[Wraith11B]

No, they won't. I have pointedly asked about it, and Paradox refuses to license the engine upon which HoI3 ran, even just to fix bugs.

 

[GarfunkeL]

No, they won't. I have pointedly asked about it, and Paradox refuses to license the engine upon which HoI3 ran, even just to fix bugs.

While unfortunate and sad, I can understand them - they've been burned in the past with modders-turned-developers (remember East vs West?) and all their engines share significant similarities so it's not like the Hoi3 version of it would be anything even close to abandonware.

Still, it would be fantastic if it happened! There are modders still working on Civilization 4 for example and look how Doom modding bounced back from the grave after its source code was made public.

 

[Wraith11B]

I even offered as a strictly "in-house" style (ie, not a license, but just an honest-to-God bug hunt), but no joy.

 

[caffran]

Thanks for all the replies. we do seem to have gone off at a slight tangent however, so to pull it back on topic, the harder settings have no effect on AI behaviour.

Does that include the 'judgement' calls it makes? for instance, the German AI compares its strength to the SU, and then when it thinks its strong enough, 'decides' when to launch Barbarossa correct?
in the above example, player=UK, would this affect how the German AI behaves in launching Sealion? if it had more tanks, trucks, planes etc, would it tip the AI to thinking it has a chance and actually go for it? if i left a few south east ports ungarrisoned, would it be tempted?

In fact, has anyone actually witnessed the AI doing sealion against a player controlled UK?

following on from that, if Japan and US gets extra production boost, logically that should mean Japan will be stronger, as it will have 4 years or so with the best laws and more materials so can produce much more yes? or is my logic flawed?

There are also two higher settings as well, hard and very hard i believe. does anyone know if this has been tested? does having the extra equipment, troops, etc actually make the AI 'think' more bolder?

 

[Kovax]

There are also two higher settings as well, hard and very hard i believe. does anyone know if this has been tested? does having the extra equipment, troops, etc actually make the AI 'think' more bolder?

The only cases I can see where it might change the AI behavior is where it can't break a Non-Aggression pact unless it has a highly favorable force ratio on the border. Not sure about TFH, but in previous expansions the AI would often build itself out of Manpower on higher difficulty settings. Hard sometimes turned out to be more challenging than Very Hard, as a result.

 

[Crispin]

What damage to you think east vs west did? And also that other eu mod game which the name eludes me. There's still a few players and modders fiddling with hoi2/aod/dh and hoi3 and one would think there could be money to be made. It would also indicate goodwill and an unprecedented way to make preserve and maintain the first and loyal paradox customers. I may be insanely naive but I fail to see the big downsides so anyone - please enlighten me. Also where is iron cross? Why not make it available again? It may seem so small and insignificant, but what could they possible have to lose ? A gain is a gain after all, no matter how small. And their reputation would increase I assume. By licensing and or cooperating with some small studio they could still retain their former selves? I mean where is the old paradox that gave us the unoffical victoria 2 patch way down the line? I just cannot imagine it costing very much to do so again - except all the corporation bs.

Alas a naive and romaticizing view I fear.

 

[Kovax]

Iron Cross Edition (ICE) was updated and renamed "Black Iron Cross Edition" (BICE). That's still being supported by the mod team, with a sub-forum section here on the Paradox forums.

The "other eu game" may have been "Crown of the North", which was based on an early release of EU1.

[ BTW - I'm really starting to hate "spell checking", when it changes what you're typing to something you REALLY don't want. ]

 

[Wraith11B]

According to the gouge--rumors--East Vs West really poisoned Paradox about their stance towards licensing their products. Unlike what they expected--a sort of Armageddon-style follow-on to HoI3 in the style similar to how they received Darkest Hour for HoI2, they got an absolute mess that never resulted in even the barest hints of a project. So now they don't license their engines to anyone.

I've offered (granted, I don't exactly have any bona fides here like the BICE or HPP teams) to help with others to just help correct the few remaining bugs (think the air and naval war, ministers not giving the correct results, etc) in the game for a final patch (basically the DH-Lite treatment) but to no avail because they don't license the engine and they don't want to be distracted by yet another project that would compete fully with HoI4 at this point.