I want to start a thread specifically dedicated toward problems in HoI in regards to invasions and the AI.
There are many problems, such as the AI not attacking certain places, the AI not backing itself up, the AI attacking with the wrong units.
HoI 1.5c has improved invasions, but they still are not perfect.
There are three areas/regions where I think the AI suffers invasions, each differently.
#1. Western Europe. (Allied AI attacks, sometimes, but never in force)
#2. Southern Europe. (Allied AI rarely attacks)
#3. The Pacific. (Japanese AI never goes on the offensive into S.E.A.)
I think that we can find a way to improve how the AI attacks, and consolidates, where it is deficient.
The main way I think we can learn how to make the AI invade in these areas is to discover how the AI succeeds in invading in other regions, or how we managed to get the AI to invade successfully. I was inspired by Steel's modification of the Italian AI and its successful invasion of Albania, when in most cases it would falter.
What are the ways we can help the AI?
Direct modification of the AI files. Creating specific AI's dedicated to fulfilling these goals. Having technologies dedicated toward sucessful invasions is important, and adding these tech trees to AI files could aid in their amphibious assaults.
-One problem that I noticed with the Allied AI's and their amphibious attacks is that they were ill timed and with different goals. Possibly we need to create specific Dday.ai files for the major western Allies that limit the assault beaches so they are more coordinated (instead of Canada landing in Cherbourg, England in Antwerp, and the US in Hamburg).
Events. A while back I was toying with ways in making a successful Western Europe invasion. One thing we could do is to create events that aid the invaders, or hinder the defenders.
-A hindering event could be something like the rise up of local resistance groups, who sabotage infrastructure, detect enemy units, etc...
This basically is fine tuning the offensive AI to want to attack the right places, and the defensive AI will not defeat the attacks quite as easy, each within the realm of reality using historic events (such as partisans).
---------------------
Here is something a bit more contraversal, which in many ways is AI cheating.
Triggers will be based on Allied occupation of bases X, Y, or Z + France occupied by Germany and western nations being controlled by the AI (Basically when the AI lands a force on Cherbourg after the Fall of France). There should probably be something about control of the English Channel, but should the AI actually invade in the first place, then chances are it is not a German lake.
This event fires for all Western Allies that take part in the liberation of France (USA, UK, Canada). The USA Event will check to see if they are at war with Germany before it fires.
The event...
Part A: Supply
-A series of large convoys are immediately formed carrying supplies and oil from England, Canada and the US (if US at war) to the specific base.
Part B: Shock
-Partisan resistance events (such as those discussed above) are triggered limiting German response.
Part C: Support (this is the big cheat)
-MANY troops from all three nations (two nations if US not at war) appear at this base using the {add division province } command. This will have to be Motorized Infantry units, perferably with Brigades attached.
-Airborne harassment assault attacks, using the {add division province}, would drop down 2-3 airborne divisions in provinces around the newly liberated province (most likely these will get into combat if German occupied, or take empty bases). This will #1. tie up German forces, #2. expand the bridgehead. (problem is, they are at 0 organization, and will be defeated if base is occupied)
Part D: Consequences
-To represent the sudden appearance of troops to actually be the deployment of troops already in existence, sufficient manpower and supplies will be deleted from the pools to represent the forces added.
Now, some will say, and I agree, that THIS IS CHEATING! One of the greatest complaints about HoI is that the Allies do not attack Western Europe in large enough numbers to be a threat to Germany. Tinkered with enough, this will 'probably' do it. I would perfer to do the stuff in the above section, but it might not be possible given that a lot of the AI and invasions appear to be somewhere not in the AI code.
There are many problems, such as the AI not attacking certain places, the AI not backing itself up, the AI attacking with the wrong units.
HoI 1.5c has improved invasions, but they still are not perfect.
There are three areas/regions where I think the AI suffers invasions, each differently.
#1. Western Europe. (Allied AI attacks, sometimes, but never in force)
#2. Southern Europe. (Allied AI rarely attacks)
#3. The Pacific. (Japanese AI never goes on the offensive into S.E.A.)
I think that we can find a way to improve how the AI attacks, and consolidates, where it is deficient.
The main way I think we can learn how to make the AI invade in these areas is to discover how the AI succeeds in invading in other regions, or how we managed to get the AI to invade successfully. I was inspired by Steel's modification of the Italian AI and its successful invasion of Albania, when in most cases it would falter.
What are the ways we can help the AI?
Direct modification of the AI files. Creating specific AI's dedicated to fulfilling these goals. Having technologies dedicated toward sucessful invasions is important, and adding these tech trees to AI files could aid in their amphibious assaults.
-One problem that I noticed with the Allied AI's and their amphibious attacks is that they were ill timed and with different goals. Possibly we need to create specific Dday.ai files for the major western Allies that limit the assault beaches so they are more coordinated (instead of Canada landing in Cherbourg, England in Antwerp, and the US in Hamburg).
Events. A while back I was toying with ways in making a successful Western Europe invasion. One thing we could do is to create events that aid the invaders, or hinder the defenders.
-A hindering event could be something like the rise up of local resistance groups, who sabotage infrastructure, detect enemy units, etc...
This basically is fine tuning the offensive AI to want to attack the right places, and the defensive AI will not defeat the attacks quite as easy, each within the realm of reality using historic events (such as partisans).
---------------------
Here is something a bit more contraversal, which in many ways is AI cheating.
Triggers will be based on Allied occupation of bases X, Y, or Z + France occupied by Germany and western nations being controlled by the AI (Basically when the AI lands a force on Cherbourg after the Fall of France). There should probably be something about control of the English Channel, but should the AI actually invade in the first place, then chances are it is not a German lake.
This event fires for all Western Allies that take part in the liberation of France (USA, UK, Canada). The USA Event will check to see if they are at war with Germany before it fires.
The event...
Part A: Supply
-A series of large convoys are immediately formed carrying supplies and oil from England, Canada and the US (if US at war) to the specific base.
Part B: Shock
-Partisan resistance events (such as those discussed above) are triggered limiting German response.
Part C: Support (this is the big cheat)
-MANY troops from all three nations (two nations if US not at war) appear at this base using the {add division province } command. This will have to be Motorized Infantry units, perferably with Brigades attached.
-Airborne harassment assault attacks, using the {add division province}, would drop down 2-3 airborne divisions in provinces around the newly liberated province (most likely these will get into combat if German occupied, or take empty bases). This will #1. tie up German forces, #2. expand the bridgehead. (problem is, they are at 0 organization, and will be defeated if base is occupied)
Part D: Consequences
-To represent the sudden appearance of troops to actually be the deployment of troops already in existence, sufficient manpower and supplies will be deleted from the pools to represent the forces added.
Now, some will say, and I agree, that THIS IS CHEATING! One of the greatest complaints about HoI is that the Allies do not attack Western Europe in large enough numbers to be a threat to Germany. Tinkered with enough, this will 'probably' do it. I would perfer to do the stuff in the above section, but it might not be possible given that a lot of the AI and invasions appear to be somewhere not in the AI code.
Last edited:
