Home Trending Latest threads New posts Developer diaries

Ai

Jump to latest Follow Reply
  • 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.
The good news is, Germany doesn't get crushed instantly like it did before.
Bad news is it comes to a halt in France and is slowly pushed back to Russia.
Also Austria-Hungary seems content with destroying Italy while it's eastern borders are overrun by Russia.
October 1st now and I think the troops will be home by Christmas as was promised.
Will now try AI drafts.rar.
Biggest problem with Germany is how much they garrisson their Northern Coast.
 
Last edited:
Really? I've only had them garrison Stralsund. Did you let the events fire on July 28th?

I've had the Eastern Front collapse in A-H, only to have it recover to their original borders by June, 1915.

The Western Front is still messed up, but it HAS to do with the invasion orders for Germany in Belgium. I'd just copied the Schlieffen orders and crossed it with the DAIM orders for ordering the invasion of the Benelux without the Netherlands.

Oddly, if the Germans do the full Schlieffen plan including the Netherlands, they'll run over Holland and stop at the border with Belgium. Perhaps the invasion orders for Belgium need to be scrapped and the basic German invasion AI just take over?

Edit: I see you haven't tried AIdrafts yet, never mind about the first question. But it is important to remember that right now it WILL NOT fire until June 28th with the current setup. If you declare war beforehand it will be the default AI.
 
Ok, tweaked Austria a bit (they used to get overrun by Russia while keeping 50 divisions garrisonned in Serbia) and will have to tweak down recklessness (Austrian armies fight in modern day Russia while separated from Austria by several provinces) but no matter what I do I can't get Germany to stop garrisoning the northern coast.
All their divisions in the west are 1 each at two provinces bordering France.
I tried setting beach priority to -1 but it seems to have no effect.
my.php

As you can see, Austria-Hungary will begin to reshuffle (and quite well) while Germany will stay the same.
A month later:
my.php

Germany's the same.
Austria then fights quite well against Serbia (reduced front distribution to favor Russia so Austria doesn't get overrun by the Russians) and too recklessly against Russia (next game will lower recklesness to 2), but goes too far and gets cut off as you can see in the next pic.
my.php

Meanwhile, for Germany it's a mad dash to the front from the beach (far more unproffesional than Germans in real life) to save what you can.
What you can see, isn't the worst case scenario.
Normaly, all of Southern Germany is French.
And if the Netherlands join in, Germany is lucky to stop the entente at the Lubeck-Dresden line.
Germany does not garrisson it's borders properly.
Why it barely puts anything on the French-Belgian-Dutch border while garrisoning the northern beaches is beyond me. I set beach priority to -1 but to no effect.
Do you know what can be done?
Oh, and these last two posts HAVE been about AIdrafts (btw your AI switch made Germany take the 1914 AI, I made them take the Schlieffen AI and put an event that triggers the 1914 AI mid september for them).
Really, Germans garrisoning the beach is the only problem I see for now.
Because Austria-Hungary coming to the aid of a collapsing Germany is just ridiculous.
Apparently all my predictions have been broken as the central powers have (albeit barely) lasted into march 1916.
Serious AI issue: Britain has managed to raise 135 divisions (biggest army in the world) and approximately 1 of them is in Europe at the moment.
Serious AI discussion ends:
 
Last edited:
TurnerBenton said:
The AI files, along with a crude event that will load the AI after the assassination of the Archduke, is located here .
Click to expand...

Awesome.

Seylanov said:
Do you know what can be done?
Click to expand...

I know that that wasn't directed towards me, but with a starting pack of ai files, I was thinking of trying a player country-centered approach. If I start with a series of ais that are activated by a human Germany, I can ignore all of the flaws of a German AI and roleplay to test the non-player countries' AIs. Eg. Response to historical Schlieffen variant, Russia-first, etc. It takes an important variable out of the mix and allows a starting point.

First off is the French defence. I'll be taking a look at the French defensive capability against a human opponent and see if there's anything I can tweak. Next, Austria vs Russia, then Austria vs Italy (all AI). Build priorities might be something to play with for Britain, but just seeing the Axis ally able to stumble slowly rather than rapidly would be nice.

With Austria, Russia, Italy, and France AIs' early-war behavior somewhat competent, things might be in better shape for developing a German AI to hold the bag.
 
Last edited:
Seylanov said:
Serious AI issue: Britain has managed to raise 135 divisions (biggest army in the world) and approximately 1 of them is in Europe at the moment.
Serious AI discussion ends:
Click to expand...

Yes, the home garrison number needs to be changed extensively. I don't know how to (a) keep additional troops in Britain and (b) send them to Europe as well as the Suez and not to every island in the Caribbean.

Edit: Alternatively, we can just accept it and make some event about Douglas Haig declaring "The Bahamas and the Andaman/Nicobar Islands must be free of German aggression". I'll make an event image if needed.




By the way, that's sarcasm.
 
Last edited:
Why would it if they're controlled by the British AI?
One of the things I'm fairly confident about is that Britain never did raise the largest army in WWI.
Why it should do so in game (especialy considering what a tough time Germany has with France as it is) is beyond me.
 
Seylanov said:
Why would it if they're controlled by the British AI?
One of the things I'm fairly confident about is that Britain never did raise the largest army in WWI.
Why it should do so in game (especialy considering what a tough time Germany has with France as it is) is beyond me.
Click to expand...

The British AI needs the most help. It's actually got one of the lowest land force production rates of any nation as a ratio to naval and air production. Obviously, it needs to be lower - but it might be due to their high IC. I guess the consumer goods penalty needs to be higher to reduce their capacity to create units. After all, the lack of rationing led to the near collapse of the British economy when unrestricted submarine warfare was recommenced - so their is a precedent for it.
 
Seylanov said:
Why would it if they're controlled by the British AI?
Click to expand...

Why would an expforced unit be exclusively controlled by the true owner's country? Its been my impression that the receiving country has control over it.

One of the things I'm fairly confident about is that Britain never did raise the largest army in WWI.
Why it should do so in game (especialy considering what a tough time Germany has with France as it is) is beyond me.
Click to expand...

Historical manpower research stats for the mod suggested a manpower pool of well over 5 million, iirc. But that pool wasn't fielded as 5 million infantrymen. It should be possible to maintain a large manpower pool without a huge number of standing infantry by changing production priorities and playing with CG.
 
I've had a look at the revised files, concentrating on the ENG one, and here are a few comments:

Radar stations: in 1914, these actually represent coast-watchers, observers and radio stations. Therefore, I suggest that countries should be permitted to build them up to level 1, especially in coastal provinces.

I noticed that ENG has relative_build_scheme=no instead of 'yes' like GER has. Was this deliberate? (As I understand it, setting this to 'no' means that ENG will build units strictly according to the percentages in the following section. Setting it to 'yes' means they will look at the units they already have and build whatever new units are needed to balance out their OOB according to those percentages)

Regarding the ENG build scheme - no wonder ENG acquires such a huge army, you've got them spending 77% of their IC on land units! In the HoI2 1939 ai file, ENG only spends 34% of IC on land units and 53% on naval... I suggest changing the balance to much more in line with eng_1939.ai. Also, ENG should spend slightly more than GER on air units, not less. Perhaps 36% land, 12% air, 52% naval? Also, I suggest allowing ENG to build garrison units (to defend islands and colonies) and perhaps even a very, very small amount of militia (native levies) - not sure about that last one, though.

As for the garrisoning the UK problem, try this as a suggestion: decrease overseas_multiplier to 0.05. I think that means that the UK will only put one garrison unit per 20 provinces in non-core territories. At the moment you've got 0.5, which means one per two provinces (and places like Canada and India have lots of provinces, hence lots of units...). Incidentally, for you've changed some of the values as compared to my draft but left the comment unchanged, so it's now wrong (for example, I had war_zone_odds = 2.0 # Will attempt to transport units to a warzone up to twice enemy strength - you've changed the 2.0 to something else, but left in the comment about "twice enemy strength")

Secondly, you've put exp_force_ratio = 0.10, which as I understand it means that if the UK has 60 divisions they will only ever send a maximum of 6 to fight in France (or anywhere else belonging to an ally). I suggest changing this to the default value of 0.75 (or possibly even higher).
 
That may in fact be the problem.
Will look into all of this.
But since Britain originaly fielded something like 4 or 6 divisions, shouldn't the exp force ration stay large and then change (currently I'm working on implementing changes for AI mid september to reflect the new reality of war).
 
I don't think there's any real need to change the percentage. If Britain has 8 divisions they'll send up to 6 to France, when they get 80 divisions they'll sned 60 of them to France. We can always tweak the exact value later.

If you are devising extra AI files for later in the war, one thing I suggest is to change the build schemes for air units. Instead of 10% or so, I'd suggest reducing this to 5% or so in 1914, then back to 10% for 1915/16, and perhaps as high as 15% for 1917/18. (Take the extra percentage away from naval units).
 
You must log in or register to reply here.
Top Bottom
Back