Frustrated!!!

[GPounda]

I'm playing Germany on Imbecile mode with historic AI; and getting very frustrated.

I am still getting used to this version of the game; but how is it possible that in 1939 Britain and France are able to launch a combined amphibious assault within 4 days of my invasion on Poland?

I have tried this three separate times and it happens each time. . .

Within 4-7 days after I invade Poland Britain and France launch a combined amphibious invasion in the Hamburg/Bremen area of Germany.

I have put garrison units, capital ships and submarines, and Stukas and torpedo bombers in the region and they still land successfully.

There is no movement on the border with France. . .

What's going on?

 

[grandad1982]

They can invade that soon as unlike in real life they use convoys to invade, which they have loads of to start with, rather than needing to develop and build enough landing craft.

Also the rules around naval superiority and launching an invasion seem a bit screwy.

At the player it should present few problems for you (sadly) and be a great way to kill the ais manpower and equipment.

 

[GPounda]

Thank you for the feed back. It's frustrating because I selected historical AI. . . so why have they never pushed through the Western border?

 

[Davarey]

Thank you for the feed back. It's frustrating because I selected historical AI. . . so why have they never pushed through the Western border?

Historical AI only serves one purpose: To choose national focuses, nothing more. Ai is stupid no matter what you do

 

[melkor88]

Lets just hope that naval invasions get a good shake up in the future.
I'm pretty sure I remember Churchill stating in his WW2 books that he didn't have the ability to naval invade with anything more then a single division until a few years into the war.

 

[ltccone]

IMO getting rid of invasion craft was a bad design decision.

 

[Dalwin]

Another point is that the big Baltic invasion hurts your enemies badly. I have seen them commit 40+ divisions to that, if they get some port like Rostock. Just make sure to have a few extra divisions, which are not participating in the Polish campaign, in position to contain the landing. Once Poland is broken kill every unit that landed. This will often thin out the French army to the point where they leave gaps in the Maginot once you enter into Belgium.

 

[Dalwin]

IMO getting rid of invasion craft was a bad design decision.

I agree with that for the most part, though if they had not done so they'd probably be struggling to get the AI to launch invasions at all.

 

[Sourlol]

How did the AI function in previous HOI games? Was is more scripted? I suppose the design goal was to produce a dynamic (check) challenging (try challenged) AI?

 

[ltccone]

I agree with that for the most part, though if they had not done so they'd probably be struggling to get the AI to launch invasions at all.

Maybe, but Britain and France had no amphibious capability to speak of when the war started, and certainly were not able to invade northern Germany days after the war started.

In my last game, Greece invaded and captured Sardinia, and Greece had no amphibious capability at all.

 

[Nuklearius]

#

 

[Dalwin]

Maybe, but Britain and France had no amphibious capability to speak of when the war started, and certainly were not able to invade northern Germany days after the war started.

In my last game, Greece invaded and captured Sardinia, and Greece had no amphibious capability at all.

Actually there is no maybe about it at all. For years both HOI2 and 3 struggled with getting the Allies to launch invasions. They either did none at all or these poorly conceived micro invasions which mostly serve to drain the Allies manpower (and now equipment). As sad as it is, the AI poses more threat amphibiously (both to you and to itself) now than it did in previous versions.

I am not arguing the historical reality of the situation. HOI4 is more game than historical simulation. It will always be so. The only way they got it to be more historical in the past was to railroad things with scripted events. I for one am glad that they have abandoned that approach.

 

[War Emblem]

I agree with Dalwin - in past HOI games you could virtually assure the AI would fail on naval invasions, if they could even launch them. By making it much easier to conduct naval invasions the AI actually poses some threat to the unwary player. Now some players will have no problem defeating them, but some won't. For the OP I wouldn't worry too much about these 39 ones. You should be able to defeat them easily enough and it will cost the Allies more than it will cost you. Now if you take Britain and have to deal with all their partisans plus constant invasions from the UK/Free France and the USA. Well - then that becomes a game within a game.

 

[kaguravitro]

How did the AI function in previous HOI games? Was is more scripted? I suppose the design goal was to produce a dynamic (check) challenging (try challenged) AI?

Yes very scripted. There are some troubles becose this version dont allow beachs restrictition coast, something basic. Also you could set priorities, DH really hadle it well, the problem was diferent.
Playing with some mods i have decent games (BICE mainly). But somethings should be changed. the transport simplication mybe is much abstraction, but i could work if separate convoy from troop transport (same system but diferent stock, diferent stock, tranport are very easy traget). Beaches are needed. actualy you should have airsuperirity in sea, but it should need air superiority over land too.

About ai in DH, it worked in low provinces management so when we test e3 maps (similar to actual hoi iv) it didnt work in modern tactics, but very good in WW1 holding fronts. I dont thing the ai actualy is so stupid. i see the main problem is moving troops from diferent threates becose it get desbalance numbers and need reorganizate it. This should be easy fix limiting the number of unit can be transfered at same time (when anyone were doing anfibius invation or paratrooper attack. In real life you get new units to build new front.

Nuklearius says right things about supplies, other thing should be corrected.

But worst wrong error in all series game (partialy correted sometimes) armored units cant do anfibius assault (exept anfibius armor). In some mods of DH this was emilated with an insane restriction for those units but this even were used by ai... and players.

 

[ltccone]

Actually there is no maybe about it at all. For years both HOI2 and 3 struggled with getting the Allies to launch invasions. They either did none at all or these poorly conceived micro invasions which mostly serve to drain the Allies manpower (and now equipment). As sad as it is, the AI poses more threat amphibiously (both to you and to itself) now than it did in previous versions.

I am not arguing the historical reality of the situation. HOI4 is more game than historical simulation. It will always be so. The only way they got it to be more historical in the past was to railroad things with scripted events. I for one am glad that they have abandoned that approach.

I understand where you are coming from, but the ability to invade at will adds a new dynamic to the game.

And while players don't have much trouble defeating an invasion, the AI does. It is disheartening to play years as Japan to see an AI England launching a successful naval invasion of Germany in the Fall of '39. Before the latest patches England would launch an almost always successful invasion of Italy as soon as Italy joined the war, because Italy was totally incapable of defending itself.

 

[Dalwin]

I understand where you are coming from, but the ability to invade at will adds a new dynamic to the game.

And while players don't have much trouble defeating an invasion, the AI does. It is disheartening to play years as Japan to see an AI England launching a successful naval invasion of Germany in the Fall of '39. Before the latest patches England would launch an almost always successful invasion of Italy as soon as Italy joined the war, because Italy was totally incapable of defending itself.

What would you propose as a solution? Some are calling for bringing back amphibious craft as a separate item.

Would it be better to move the amphibious techs to later years and then not have most countries begin the game with them, perhaps even reducing the first one to 4 divisions instead of 10?

If you make this function too restrictive, you run the risk of curtailing Japan's historical push in the Pacific, though I suppose one could always give Japan an advantage either in the form of beginning with a given tech researched or through some NF or national spirit.

 

[ltccone]

What would you propose as a solution? Some are calling for bringing back amphibious craft as a separate item.

Would it be better to move the amphibious techs to later years and then not have most countries begin the game with them, perhaps even reducing the first one to 4 divisions instead of 10?

If you make this function too restrictive, you run the risk of curtailing Japan's historical push in the Pacific, though I suppose one could always give Japan an advantage either in the form of beginning with a given tech researched or through some NF or national spirit.

IMO it is too late to go back to having separate invasion craft. To me the best solution is to make the first amphib tech a 1940 tech. This would stop all of the silly invasions at the beginning of the war, and would prevent AI minors from launching invasions at all (In addition to me seeing Greece launch an invasion of Sardinia I've seen The Netherlands invade Ethiopia).

Both Japan (for China) and Germany (for Norway) would get those techs from NFs.

I don't think that the number of divisions (10) needs to be reduced for the first tech.

 

[browd]

I don't think that the number of divisions (10) needs to be reduced for the first tech.

I disagree. I like the approach taken in Vanilla+, where the initial invasion tech only permits 6 divisions, and (IIRC) +12 at the second tech level. Won't stop all early invasions, but it makes the invasions that do occur more realistic in size.

 

[adam_grif]

The AI will definitely need to be tweaked to be unwilling to commit to an invasion unless it believes it could plausibly achieve something with it. That's a sort of high level judgement that the AI is just not capable of doing currently. The British "could" have landed a division in northern Germany but it was suicide, so they would never do it. That is sort of true in game except that they can land too many, too quickly, and they KEEP DOING IT even though it's suicide and they keep getting repulsed.

I would love for landing craft to be a thing that needs to be built actually. But you should still be able to amphib invade without them, but if you are opposed in your landing it needs to be even harsher than it is now.

 

[Dalwin]

The AI will definitely need to be tweaked to be unwilling to commit to an invasion unless it believes it could plausibly achieve something with it. That's a sort of high level judgement that the AI is just not capable of doing currently.

I agree that it is not capable of making such judgments. This is one area where I disagree with the policy of purism that is being taken in the area of AI design. They have said a number of times that they want to program the AI to make its own decisions rather than scripting it along certain paths. They support that stance with the idea that a scripted AI is more predictable and can be defeated more easily by knowing in advance what its scripted decisions or reactions to certain triggers will be.

That all sounds noble as part of a mission statement. However, at some point they need to realize that they are struggling to meet those lofty goals, perhaps outright failing. I see nothing wrong with scripting a few tendencies. Play the odds and do things that will usually be better while not being so easily abused. For example, a script that told the UK AI to never launch an invasion before '43 would be a good decision most of the time, but would then allow the German player to ignore port defense prior to '43 and instead commit those units to the Eastern Front.

Better might be to allow the AI to cheat very slightly. By that I mean let it know how many divisions are on port and coastal defense bordering a given sea zone (but not exact locations.) Then include triggers that say in year X if the garrison is below Y, I am allowed to begin planning invasions in that region. A player could still look at that code (and some would) and be careful to always stay above the minimum. Most players would not do this and even those that did would be kept honest to a degree by the escalating nature of the decision process.