What are you doing with your air wings? at what level have you attached them? I find that when I assign an army or army group a air group, they automatically detach and spread out my carefully constructed air groups.
-
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.
I have noticed a bit of that. More so at the Army Group level than at the Army level. And it hasn't happened all the time. I will keep a more watchful eye on it in the next chapter this weekend.
That said, I haven't overly worried about the splitting up of air groups though, as I have seen the AI direct as many as 7 air units into bomber interceptions, so it is doing its job so some extent.
Are your Panzer Armee at an Army Level or at a Corp Level ?
I noticed they have 5 divisions in them.
I noticed they have 5 divisions in them.
Corps. I just like to call them Panzer Armees because it feels right.
Defensive AI Assessment - Poland: 1 Sept 1939
For the review of the Defensive AI, I've gone back to the Blitzkrieg campaign, but this time I'm Polska!!
Now that I know what I am doing, reorganising the army is much easier. It's helped by the Polish Army being pretty well organised in the default scenario. So just a bit of tidying up was required. I decided to play at the Army Level and gave all the front line units a Defend order to watch the AI at work.
The default game starts with most of the air units seperated. I decided to created larger air groups to give my bi-planes some sort of chance. 3 Groups were created and formed up in the early hours of 1 Sept: Krakow, Warsaw and Lodz Air Groups, each assigned to their respective Army Groups.
I sent the Polish Navy to Dover
My real involvement would be in the creation of the 1st Reserve Army Group from the units away from the front line. 4 Corps averaging 4 Divisions each. Just 16 Divisions of Infantry/Militia for counter-attack and support purposes. Here's where they started:
As you can see, as soon as I hit the Run button, the Germans launched their usual attacks including a very heavy one towards Warsaw from East Prussia. So I decided to move to Corps level orders for the Reserve Army, and sent the Warsaw Reserve Corps and the Skwarczynski Corps into counterattack. The East Reserve Corps were ordered to Warsaw to act as the new reserve there, leaving the Kruszewski Corps as my only interior line of defense against the onslaught from the West . Not how I had seen it going!!
One of the comments earlier in the thread has been the AI's tendnecy to split up Air Groups. Well I didn't see any initially as this fight on Day one testifies. The seven air units are from the Lodz and Krakow Air Groups:
It did split up the Warsaw Air Group: splitting away the 2 Tactical Bomber Wings from the Fighters. Fair enough, I let the AI do that.
So far so good.
I've got to do some chores now (well under the thumb). Will talk though the campaign this evening...
For the review of the Defensive AI, I've gone back to the Blitzkrieg campaign, but this time I'm Polska!!
Now that I know what I am doing, reorganising the army is much easier. It's helped by the Polish Army being pretty well organised in the default scenario. So just a bit of tidying up was required. I decided to play at the Army Level and gave all the front line units a Defend order to watch the AI at work.
The default game starts with most of the air units seperated. I decided to created larger air groups to give my bi-planes some sort of chance. 3 Groups were created and formed up in the early hours of 1 Sept: Krakow, Warsaw and Lodz Air Groups, each assigned to their respective Army Groups.
I sent the Polish Navy to Dover
My real involvement would be in the creation of the 1st Reserve Army Group from the units away from the front line. 4 Corps averaging 4 Divisions each. Just 16 Divisions of Infantry/Militia for counter-attack and support purposes. Here's where they started:
As you can see, as soon as I hit the Run button, the Germans launched their usual attacks including a very heavy one towards Warsaw from East Prussia. So I decided to move to Corps level orders for the Reserve Army, and sent the Warsaw Reserve Corps and the Skwarczynski Corps into counterattack. The East Reserve Corps were ordered to Warsaw to act as the new reserve there, leaving the Kruszewski Corps as my only interior line of defense against the onslaught from the West
. Not how I had seen it going!!One of the comments earlier in the thread has been the AI's tendnecy to split up Air Groups. Well I didn't see any initially as this fight on Day one testifies. The seven air units are from the Lodz and Krakow Air Groups:
It did split up the Warsaw Air Group: splitting away the 2 Tactical Bomber Wings from the Fighters. Fair enough, I let the AI do that.
So far so good.
I've got to do some chores now (well under the thumb
). Will talk though the campaign this evening...
Last edited:
Battle for Poland - The Trials and Tribulations of the 1st Reserve Army Group
The first thing to say straight away, is that the standard AI defense is perfectly fine. Tell units on the front line to defend (even without specific objecties), and they will do just that without going on long country marches. They'll defend reasonably well, perhaps not as well as micromanaged armies, but we've been through that.
By Sep 3. It was clear that the Germans in East Prussia were continuing a major attack. I decided to throw the large Pomorze Army, which straddled the Danzig Corridor into an attack German east Prussian East wing. This may divert some of the German attacking forces, and also start to extract my forces from potential envelopment in the Danzig Corridor.
Using the experience gained in my 2 earlier campaigns, I gave the Pomorze Army one objective and they attacked accordingly...
Still, looking fine. Obviously, my troops all over the front are being beaten comprehensively. But they are putting up a fight, and slowing the German advance. No holes anywhere.
AI Commanders going against orders!
Unfortunately, the AI started to do silly things.
Sep 5, and my East Reserve Army Corps have started heading for the southern front rather than covering Warsaw as I had ordered. Even after I directly ordered the HQ to the right area, it's Divisions continued on their way south
Sep 7 and the Kruszewski Corps which had faithfully set up along the defensive line around Lodz and Tomaszow, were seen going off in all directions, presumably where the AI thought there would be holes in the front. My Army Group is falling to pieces
At least the attack AI is being obedient and the Warsaw Reserve Corps' counter attack goes in as planned on Sep 12. They are assisted by the Modlin Army, despite not being ordered to do so (good initiative, give that General a promotion)...
Here is the big picture by Sep 15: 2 weeks into the war. Overall, I am fairly pleased. While my Pomoroze Army attack on East Prussia is running out of steam, it seems that it may have done its job in stopping the attack on Warsaw. The counterattack by the Warsaw Reserve Corps is just about to succeed. The rest of the front is falling back, but the either the German attack AI is failing to envelop or our defensive AI is doing a good job at avoiding envelopments.
Look at the middle of Poland. I have no reserve!!!!. So it comes as no surprise that on Sep 18, the Germans breakthrough on the Western front, and there is a clear road to Warsawo. My Reserve Army is embroiled in the counterattacks north of Warsaw and cannot help. Did I commit them too early? Possibly. It's the beginning of the end.
Air AI works fine
Time for a brief respite from the ground war. The question as to whether the air wings get involved and stay together is answered in this shot. These boys have been fighting sure enough!
Last Stand
I won't give a blow by blow of the last stand, other than the AI dug it's heels in. On Sep 26 I gave orders for all my northern armies to Defend Warsaw. The southern and western armies are too embroiled and too far away to help...
Oct 1 and the airforce continues to fight against the odds...
Oct 3 and my Reserve Army Group calls it quits...
Defeat - Oct 6th. History pretty much repeats itself!
Next time... Learnings
The first thing to say straight away, is that the standard AI defense is perfectly fine. Tell units on the front line to defend (even without specific objecties), and they will do just that without going on long country marches. They'll defend reasonably well, perhaps not as well as micromanaged armies, but we've been through that.
By Sep 3. It was clear that the Germans in East Prussia were continuing a major attack. I decided to throw the large Pomorze Army, which straddled the Danzig Corridor into an attack German east Prussian East wing. This may divert some of the German attacking forces, and also start to extract my forces from potential envelopment in the Danzig Corridor.
Using the experience gained in my 2 earlier campaigns, I gave the Pomorze Army one objective and they attacked accordingly...
Still, looking fine. Obviously, my troops all over the front are being beaten comprehensively. But they are putting up a fight, and slowing the German advance. No holes anywhere.
AI Commanders going against orders!
Unfortunately, the AI started to do silly things.
Sep 5, and my East Reserve Army Corps have started heading for the southern front rather than covering Warsaw as I had ordered. Even after I directly ordered the HQ to the right area, it's Divisions continued on their way south
Sep 7 and the Kruszewski Corps which had faithfully set up along the defensive line around Lodz and Tomaszow, were seen going off in all directions, presumably where the AI thought there would be holes in the front. My Army Group is falling to pieces
At least the attack AI is being obedient and the Warsaw Reserve Corps' counter attack goes in as planned on Sep 12. They are assisted by the Modlin Army, despite not being ordered to do so (good initiative, give that General a promotion
)...
Here is the big picture by Sep 15: 2 weeks into the war. Overall, I am fairly pleased. While my Pomoroze Army attack on East Prussia is running out of steam, it seems that it may have done its job in stopping the attack on Warsaw. The counterattack by the Warsaw Reserve Corps is just about to succeed. The rest of the front is falling back, but the either the German attack AI is failing to envelop or our defensive AI is doing a good job at avoiding envelopments.
Look at the middle of Poland. I have no reserve!!!!. So it comes as no surprise that on Sep 18, the Germans breakthrough on the Western front, and there is a clear road to Warsaw
o. My Reserve Army is embroiled in the counterattacks north of Warsaw and cannot help. Did I commit them too early? Possibly. It's the beginning of the end.
Air AI works fine
Time for a brief respite from the ground war. The question as to whether the air wings get involved and stay together is answered in this shot. These boys have been fighting sure enough!
Last Stand
I won't give a blow by blow of the last stand, other than the AI dug it's heels in. On Sep 26 I gave orders for all my northern armies to Defend Warsaw. The southern and western armies are too embroiled and too far away to help...
Oct 1 and the airforce continues to fight against the odds...
Oct 3 and my Reserve Army Group calls it quits...
Defeat - Oct 6th. History pretty much repeats itself!
Next time... Learnings
Learnings (or how to run an HQ AI Military Game and Survive)
Well, for the past three weekends, I have been on a bit of a journey to find out if v1.3's HQ AI cuts it. My conclusion is YES, but you have to approach with both a different mindset and take due care in certain aspects in order for it to work out right.
Here are my steps to AI Happiness:
Wishlist from PI.
Now this has not been a scientific assessment, but an empiical one. Consequently, it is quite feasible that some of my recommendations are flawed. If I find different outcomes in my games ahead, I'll make sure I'll update the list. New ones will be highlighted in Yellow
So long for now. Happy to hear your thoughts.
Well, for the past three weekends, I have been on a bit of a journey to find out if v1.3's HQ AI cuts it. My conclusion is YES, but you have to approach with both a different mindset and take due care in certain aspects in order for it to work out right.
Here are my steps to AI Happiness:
- Organise your entire Armed Forces using the Theatre/Army Group/Army/Corp/Division/Brigade hierarchy/OOB.
- Don't leave any spare units reporting outside the hierarchy or they won't get used effectively.
- Forget the micromanagement / clickfest approach. Wean yourself off it. Give orders at the Army level. Temporarily detach units or set Corps level objectives for special tasks or objectives which are in a different direction to your main thrust.
- Plan your War beforehand. Give your Armies objectives.
- Get to enjoy building your OOB. Give your armies memorable names where possible, to help you remember what their objectives are.
- Ensure you have Reserves.
- Ensure your HQ is close to its subordinates. Manually move it towards them if not.
- Attach your Air Wings and Groups to the Army Level. Any higher, and they won't fight well. Ensure they are close to their Army HQ.
- Create Air Groups of the same aircraft type, or else the AI will split it up.
- To ensure that your armies follow your orders, place both attack and defend orders INSIDE enemy territory if possible.
- AI armies are quite likely to follow their own "initiative" than your orders if you give orders WITHIN your territory.
- AI armies are quite likely to follow their own "initiative" than your orders if there is a HOLE in the Front Line near to it.
- At the Front, only give your army 1 or 2 objectives no more than 2 or 3 provinces away. More objectives and greater distances mean that they are likely to stray easier.
- Because of the fair chance of one or more of your armies straying from their orders, keep an eye on all of them, not just focussing on the key battle areas.
Wishlist from PI.
- Reduced levels of AI initiative and straying from orders (based on Leadership % perhaps). So a strong leadership would ensure greater acceptance of orders.
- Get Prepare working properly. I see no difference between it and Defend.
Now this has not been a scientific assessment, but an empiical one. Consequently, it is quite feasible that some of my recommendations are flawed. If I find different outcomes in my games ahead, I'll make sure I'll update the list. New ones will be highlighted in Yellow
So long for now. Happy to hear your thoughts.
Last edited:
Well, you nearly quoted von Manstein here.
Thx for your AAR.
Some guys (Chilango2, Gladiator, Myth, Henri and me) did some analysis on the 1.2 AI back then and we can support your findings: The AI works decent and it is fun to use it.
Did you see any influence of the axis of advance you set? Did you set them at all?
Do you plan doing the Battle for Britain?
Thanks for the AAR. I love people who take the time out to figure stuff out for the rest of the players. So now we know what to do and what not to do. Thanks again!
Well, you nearly quoted von Manstein here.
Thx for your AAR.
Some guys (Chilango2, Gladiator, Myth, Henri and me) did some analysis on the 1.2 AI back then and we can support your findings: The AI works decent and it is fun to use it.
Thanks - Good to know we are on the same page. I see this AAR gets a mention in the 1.3 strategy guide update which is encouraging.
No. That is remiss of me. Have you found that shift / right click increases the chance of the AI following your orders?
Seelowe? There's a thought. AI led amphibious and air assaults. I can see a new section to my AAR coming
Did you experienced some of your mobile troops (such as Arm or Cav divisions) to be drag outside the range of their HQ corps ? I did in my 36' japan game last night. I landed all my cav and mot units around Shanghai in an absolute mess within a couple of province. After a month of blitzing stance I found some units to be distant from their HQ about 1000km away.
The main problem was those units completely lose initiative and stood still far beyond the front line. I guess they were waiting for some orders that would not come.
I found quite annoying to reorganise my OOB in the midst of a blitzkrieg. I stated (and not yet tested) that I should create "reserve" HQ corps with no division attached in order to rally those divisions. Those reserve HQ would manually follow the front and rally any out of range units. HQ seems to follow the axis of progression while fighting units are dragged away with this battle frenzy you talked about in this AAR.
I doubt you'd encounter such obvious and grotesque behavior before late (and victorious) Barbarossa. China has few units and large territory, the blitz never ends there.
Regarding Seelöwe, I'm looking forward your conclusion since amphibious warfare is a key feature of Japan expansion. I personally conclude that the IA is not working great with this. Ai would launch an invasion only with ridiculous favorable odds with only short term view.
For instance it won't try to encircle a port and try a frontal assault. There would be no combined invasion on large coast sections (say 4/5 provinces). This tend to a 1 province invasion with large stacks that are easily driven out by few defending divisions.
The main problem was those units completely lose initiative and stood still far beyond the front line. I guess they were waiting for some orders that would not come.
I found quite annoying to reorganise my OOB in the midst of a blitzkrieg. I stated (and not yet tested) that I should create "reserve" HQ corps with no division attached in order to rally those divisions. Those reserve HQ would manually follow the front and rally any out of range units. HQ seems to follow the axis of progression while fighting units are dragged away with this battle frenzy you talked about in this AAR.
I doubt you'd encounter such obvious and grotesque behavior before late (and victorious) Barbarossa. China has few units and large territory, the blitz never ends there.
Regarding Seelöwe, I'm looking forward your conclusion since amphibious warfare is a key feature of Japan expansion. I personally conclude that the IA is not working great with this. Ai would launch an invasion only with ridiculous favorable odds with only short term view.
For instance it won't try to encircle a port and try a frontal assault. There would be no combined invasion on large coast sections (say 4/5 provinces). This tend to a 1 province invasion with large stacks that are easily driven out by few defending divisions.
I did not encounter this behaviour. As you say, the Western European war zone does not allow much room for running very far away from the HQ assuming it was in contact in the first place. Did you keep your objectives to within a few provinces, and your troops just kept going? As I have said, the AI tends to give units too much initiative for my tastes. There are many examples in WW2 where leaders have been frustrated by being held back by HQ, when there is a clear road ahead, but more often than not, those leaders have obeyed orders. Only the undisciplined ones would have disobeyed, which is is why I would like the chances for AI going against orders to be based on the leadership %.
Hmm, the D Day landings were at the Corps level for each beach, and that is how I think will approach it for Seelowe.
From some much more casual observation, I think that in 1.3 the AI has started to try and keep the division-corps relationship relatively intact (in terms of HQ range). Its still pretty random above that point though and operates with no feel for Army/A grp relationships.
Kirth,
This is a magnificent AAR, and a great project -- well done!
Rather than reprise your hard work, I figured I'd point people over here, so I made reference to this AAR in my HOI 3 v1.3 Update Guide.
I hope you don't mind.
Thanks!
Rensslaer
This is a magnificent AAR, and a great project -- well done!
Rather than reprise your hard work, I figured I'd point people over here, so I made reference to this AAR in my HOI 3 v1.3 Update Guide.
I hope you don't mind.
Thanks!
Rensslaer
More about air wings. It looks like the AI doesn't like different types of air units in the same group. I have a tendency to attach at least one interceptor to my tactical and CAS bomber wings, and the first thing the AI seems to do when I give it control is split off the interceptors.
In your scenario here, It looks like when you kept the Polish interceptors together, they stayed that way. That's something to keep in mind, when playing the game using this approach, at least from my experience with the AI.
In your scenario here, It looks like when you kept the Polish interceptors together, they stayed that way. That's something to keep in mind, when playing the game using this approach, at least from my experience with the AI.
Agreed. The mixed Group was split up, the single type groups were not. It makes a mess of your OOB. I think this is worth adding to my list.
Thank you. I had spotted it earlier, and it was flattering! I can't help but think that we are still on a little journey here which needs more analysis, and I am sure I will change my conclusions as I go along.
Last edited:
No. That is remiss of me. Have you found that shift / right click increases the chance of the AI following your orders?
Seelowe? There's a thought. AI led amphibious and air assaults. I can see a new section to my AAR coming
Yes, the axis of advance helped tremendously on army level.
In one test run, I chose to set one final objective plus the axis of advance. This objective was chosen quite deep into the enemy territory, e.g. Torun, Lodz or Warsaw.
In this run, Poland fell after 15 days, just as fast as the theatre AI could do it and only one day short of my (not too good) manual control.
Concnering the invasion AI: Looking very much forward to it!