Annals of the OKW- A 1936 Germany AAR

[Henri]

Don't worry, I'm not giving up.

Me neither, but I stopped playing until it seems much safer to play. I started dozens of game, but a crash to desktop after spending over an hour setting up Barbarossa sucked the life out of me too, and another try with Germany 39 against Poland with Army Groups under AI just flattened all my tires - every unit after two months and after taking one province all around were either stationary or moving in the opposite direction.

Despite agreeing that this game has the most promise of any wargame I can think of, it is simply not playable in its present state. I don't object much to ahistorical results, but unpredictable crashes destroying hours of preparations and most units attacking in the wrong direction is - well, I just can't take it any more...

Henri

 

[unmerged(141095)]

A quick Q for you when you reach this point in the game, I've held off going to war until April 1940, and prior to this I noticed that the Molotov-Ribbentrop pact vanished from my diplomacy list... wasn't aware that this was time bound at all?

Also ran a test invasion with the AI at Korps level, and chaos ensured, have no divisions on the southern part of my western front, appears they decided to migrate north for the summer as soon as AI control went on. So I suspect you'll be manually controlling the invasion of Poland.

 

[notger.heinz]

Chilango2, glad to hear you don't give up.

Look at this, it might cheer you up (and since your german is possibly non-existent, it says "never give up"):

http://si6.mtb-news.de/fotos/data/503/38525STORCH.jpg

 

[unmerged(156111)]

A quick Q for you when you reach this point in the game, I've held off going to war until April 1940, and prior to this I noticed that the Molotov-Ribbentrop pact vanished from my diplomacy list... wasn't aware that this was time bound at all?

Also ran a test invasion with the AI at Korps level, and chaos ensured, have no divisions on the southern part of my western front, appears they decided to migrate north for the summer as soon as AI control went on. So I suspect you'll be manually controlling the invasion of Poland.

Gwynn, I have decent but far form perfect results when using the AI on Corps level. Anything higher (Army or Army Group) is totally futile as the AI will "migrate" your units to one province where you end up with 50+ divisions (and huge stacking penalties and no supply)

 

[unmerged(156111)]

Me neither, but I stopped playing until it seems much safer to play. I started dozens of game, but a crash to desktop after spending over an hour setting up Barbarossa sucked the life out of me too, and another try with Germany 39 against Poland with Army Groups under AI just flattened all my tires - every unit after two months and after taking one province all around were either stationary or moving in the opposite direction.

Despite agreeing that this game has the most promise of any wargame I can think of, it is simply not playable in its present state. I don't object much to ahistorical results, but unpredictable crashes destroying hours of preparations and most units attacking in the wrong direction is - well, I just can't take it any more...

Henri

While I wait for a patch that will fix the crash problems, I pulled out a book I read a few years ago:

"Barbarossa - The Russian German Conflict 1941-1945" by Alan Clark
http://browseinside.harpercollins.com/index.aspx?isbn13=9780688042684

This is an excellent book that I can strongly recommend reading. Hopefully the patch will come before I complete all 465 pages...

 

[unmerged(141095)]

Gwynn, I have decent but far form perfect results when using the AI on Corps level. Anything higher (Army or Army Group) is totally futile as the AI will "migrate" your units to one province where you end up with 50+ divisions (and huge stacking penalties and no supply)

How did you set your corps objectives? (e.g. for blitzing did you give panzer corps an objective several provinces in? likewise for defensive or attacking stances)

 

[unmerged(156111)]

Gwynn, what I prefer to do is:

Blitzing; (i.e your panzers or fast mechanized/motorized that you want to penetrate deep into enemy territory) set one or two provinces on the front where you want the main thrust to go through the front line and then an objective deep behind the enemy line. If this point is more than 5 provinces in you should "set a path" of objectives where you want the thrust to go. If you set only one objective too far behind the enemy line the AI will be confused and do nothing but moving units back and forth. I also have noticed that some units get "stuck" in a province after it is captured instead of pushing on. The way I solve this is to go back and un-select the province as an objective to get the unit moving again.

Attacking; set objectives along the whole front covered by the corps and one deeper in as a final objective. The final objective should not be more than a few (5?) provinces in or the AI will be confused.

Defence; If you have enough divisions to cover the whole defensive front the AI will do a good job even on the theater level. You can do this by selecting a single point on the enemy side or all the provinces you want defended. If you are a few units short it is safer to set defensive objectives on lower levels (corps or army) or you risk the AI will do a "migration" of all your troops to one area and leave large areas undefended.


Again the AI is far from perfect but it gets the job done (in most cases) with a limited amount of manual work.

 

[unmerged(141095)]

Thanks Jan, will have an experiment with it tonight, hopefully with better success this time around.

 

[Chilango2]

A quick Q for you when you reach this point in the game, I've held off going to war until April 1940, and prior to this I noticed that the Molotov-Ribbentrop pact vanished from my diplomacy list... wasn't aware that this was time bound at all?

Also ran a test invasion with the AI at Korps level, and chaos ensured, have no divisions on the southern part of my western front, appears they decided to migrate north for the summer as soon as AI control went on. So I suspect you'll be manually controlling the invasion of Poland.

I think M-R goes away when 1940 starts, so if you're going for a spring 1940 start, you need to trigger it in 1939, then when the war starts in 1940, it will occur as intended. Mind you, this will do amusing things like trigger the "winter war" in spring, but hey, that's no big deal.

I hope to get an update to the lot of you on Thursday or Friday. And yes, I am sorry, but lifge means we've essentially moved to a once per week update schedule.

 

[notger.heinz]

I think M-R goes away when 1940 starts, so if you're going for a spring 1940 start, you need to trigger it in 1939, then when the war starts in 1940, it will occur as intended. Mind you, this will do amusing things like trigger the "winter war" in spring, but hey, that's no big deal.

I hope to get an update to the lot of you on Thursday or Friday. And yes, I am sorry, but lifge means we've essentially moved to a once per week update schedule.

Ah, life ain't what it used to be, anyway.

And Gwynn, I would not recommend signing the M-R-pact, so be glad you missed the opportunity. But then again ... why would you want to wait with the war until 1940? I usually go for it as fast as possible, to minimize the risk of a well-prepared opponent.

 

[loki100]

Ah, life ain't what it used to be, anyway.

And Gwynn, I would not recommend signing the M-R-pact, so be glad you missed the opportunity. But then again ... why would you want to wait with the war until 1940? I usually go for it as fast as possible, to minimize the risk of a well-prepared opponent.

actually sitting on the other side of that particular fence at the moment, I'd agree. My USSR would have been pushed to the limit by a 1940 attack - worse if it had been earlier.

 

[unmerged(144305)]

Holy mother of all threads.

I've finally gotten to the end of it.

Chilango, everyone's praised your work here, so I hope mine doesn't fall on deaf ears. This is the most insightful glimpse into game mechanics that I've yet seen, and you've done higher quality, more thorough work than I have ever seen. THANK YOU SO MUCH!

I've spent the last three days trying to bite off pages of this thread when I should really be studying for my economics test tomorrow, but really, when the decision is between econ study and HOI3 study, I think my choice is understandable. Good luck in school, dude. It ain't easy when you've taken a long break (mine was 14 years).

I'm very interested in seeing how your AI conquest of Poland and later France goes. One thing I would ask you to do, assuming you find time, is to run the battle of Poland (and maybe France) more than once using a save & reload to see what variations might play out if you do different things with objectives and/or setting AI control to Korps/Armee/Theater. If the time isn't there, well, I'll just try it on my own.

I've really appreciated how you have taken steps to avoid gamey manipulation of the system, by the way, and your gestures toward historical progression. Thanks for that!

Now I have some questions for anyone who has capability to answer them, Chilango or otherwise.

Spying/Intel Missions:
To "pull" a sympathetic country like Romania toward you, I understand you want to keep your threat low. But is there any other value you want to keep track of (like neutrality)? Does the target country's political situation, in terms of the red/black/blue internal party composition, matter to "pull?"

Speaking of threat:
How do you annex Chekoslovakia 'peacefully'? I've only played the 1938 game to this point, and I always invade. Seems that's a really bad idea given the threat consequence.

Practicals & Production Efficiency Bonuses:
Does anyone have any data on whether it's better, in the 1936-39 time frame, to build small parallel builds of 1 or 2 units, over and over, in order to gain production efficiency bonuses from practicals in a short term? Or is it better to build parallels in 5 or 10 or some other size just to crank it all out at once? I have a very hard time understanding the significance/contextual implication of practical bonuses to both research and production. Sure could use some help there.

Chilango: No CAS? Warum?

What message settings do you prefer? You seem to have a very controlled feedback mechanism that I'd like to emulate.

Gluck auf, Herr Chilango
und vielen danken!

 

[unmerged(141095)]

Ah, life ain't what it used to be, anyway.

And Gwynn, I would not recommend signing the M-R-pact, so be glad you missed the opportunity. But then again ... why would you want to wait with the war until 1940? I usually go for it as fast as possible, to minimize the risk of a well-prepared opponent.

I've had Republican Spain, Belgium and the Netherlands all side with the Allies prior to war breaking out, so I wanted to ensure my western front would be secure while I dealt to Poland, so I gave myself 6 months to prepare for this. Haven't had a chance to fire off the war yet and keep on experimenting with AI control, hopefully should have a crack at it tonight.

Schedule is fairly tight though, as I'm aiming to launch Barbarossa in May 1941, so 13 months to take out Netherlands, Belgium, Denmark, Norway, France, Spain, and Yugoslavia (Italy should have hopefully of dealt with Greece by then).

 

[notger.heinz]

Gwynn: Norway, Denmark, Yugoslawia, Spain ... why? It's a waste of resources, if you ask me. Ok, Denmark possibly not, but there you would only need a small army and could do it simultaneously to Barbarossa.
Don't waste troops when there is a real (red russian rampage-ready) threat on your eastern border, that does not like alliterations.

heliodorus4:
Threat is in 1.2 less important and only plays a role if you are a neighbour or they come close to your faction alignment-wise, as far as I understood it.
Spying/Intel: Check the tool-tip in the diplo-window when you have threat (start 1941 ). It matters, but influence trumps everything.
Practicals: If they decay faster than you can build them up, then you make no progress. So you would want to crank it up to maximum bonus as fast as possible, thus I suggest building the troops exclusively one type at a time and research their techs afterwards (practicals go a long way in terms of research time!). I do not know the equilibrium state or the impact of upgrade costs, though.
In short: Build first, build lots, research later for max speed (which seems gamey to me).
CAS: I tend to skip them, too. TAC are more versatile and have longer range, which you want to have against GB and Russia.

 

[unmerged(141095)]

Well I'm at war with Spain, so once I've knocked France out I may as well carry on and deny the Med to the British. Yugoslavia and Greece as part of securing my flank (Yugoslavia is drifting to the allies awfully fast).

Have just tested out theatre AI versus Poland. And it worked surprisingly well, victory over Poland in 22 days, a couple of multi division encirclements... and a Polish surrender (not a GiE!). Suspect that last part was partially because of not having signed Molotov-Ribbentrop.

More interesting was the Western Theatre AI. Being at war with France, Belgium and Netherlands, the defensive AI has done exactly what Paradox has said it would do. It's assessed that it's divisions are stronger than the Belgian's and Netherland's and it's advanced against both of them. It's worried me that it's advanced so far, especially in Belgium, as if it hits the French border I expect it'll bring reinforcements north before I can redeploy from Poland. However it fits in with what I'd read earlier in the Dev Diaries about the AI, if it can improve its defensive condition then it will... the fun thing is that it has taken that several provinces beyond the objectives I gave it (all in German territory).

Will hopefully get to run it further this weekend and see how it handles invasions of France and Spain. Given positive results in Poland, I'm hoping to be wrapped up in the west by the end of July 1940.

 

[notger.heinz]

Starting in '39 for fun, I gave objectives for the army group AI to defend the border against France (defensive stance, german provinces marked, one province in France marked), which was the sole enemy in that set-up.
The army group immediately redeployed and left Todtmoos and Baden undefended. -> Fail.

In the same scenario in the east, the army group north was very reluctant to move one after the initial capture of Danzig (set as objective). It seems, the group is really listening to your axis of advance.
Army group south did only slow progress against weak polish troops. -> Weak fail.

Any hints, Gwynn/Chilango2/potski?

Some advocate using corps AI, but that seems to defeat the purpose of AI command somehow, since the step to controlling everything manually then only would be a comparatively small one and I would lose my supreme-commander-feeling.

In 1.1c I liked to use the army AI, but she does not seem to work too good, anymore.

Besides: Let Yugoslawia become an allies member, declare war on you (they will not necessarily) and then call in the Italians to deal with them. I strongly favour an early Barbarossa, though that borders on being gamey. Plus: You can go to Egypt from Sevastopol.

 

[unmerged(102275)]

Very nice! Well done!

 

[notger.heinz]

Ok, 1.2 seems to have given the theatre AI an enormous boost.

'39 scenario, Germany overran Poland in five weeks, the only slow-downs being Warsaw and Grodno. Minor encirclements and a constant push made a very good show.
While blitzing, it still held reserve troops to protect my ports in northern Germany.

Re-deploying army group north to the western theatre in two ways, delivered the following:

1. Re-deploying w/o assigning to OB West, giving the army group the attack command on Netherlands/Belgium, then declaring war. Nothing happens. Literally.

2. Re-deploying then assigning to OB West, declaring war and changing the defense-stance to attacking, while giving all of the southern border as objectives as well as some on the dutch border and Eupen. Hell breaks lose. Apart from stacking airplanes, the AI rushes like hounds of hell unleashed and pushes, where it can, even if not given as objectives. It sees its chance and goes in. Great, I love it.


Conclusion: The AI seems to be highly aware of all of the theatre and seems to take all the troops into account. On a theatre level, it seems to be very fast in using its advantage and I did not see any retreat-like behaviour.
If you give provinces bordering yours as objectives, or a clear path of objectives connected together, it will push along this line.
The axis of advance is important, too, but not so, when you go on all-out-offense. Haven't had a chance to work against a strong opponent and see, what the axis does there.
On army group level, it still seems to factor the whole theater in, but only as far as the enemy goes, not as far as the own troos go, and tends to be very cautious, thereof.
AI on army level e.g. rates the polish enemy as equally strong or stronger ... for every of your armies! The theatre rates it weaker and is therefore more aggressive. The army and army group AI seems to see ALL of the polish army as appropriate estimate of the enemy strenght, while only taking the army itself as own strength, ignoring nearby comrades.

For me, it is theatre AI all the way, for now.
Special duty task forces (all the navy and parts of the air force I do not want to have split up) will still be controlled manually, though.


(Chilango2, I hope you do not mind me posting some minor AI tests in your thread to keep mouths wet until you can do an update. Your thread has become the major source for AI control theory, so I feel this is the right place to do so. If you object, drop me a note.)

 

[unmerged(141095)]

Likewise I hope you don't mind our musing on the subject.

For my theatre AI I removed all naval vessels and my Port garrison forces from it, if only because it suddenly started wanting to use these as troops to hold the line against France... and I was envisaged a nightmare scenario where the garrison army for France would suddenly be lined up to hold the Western Front while the rest of the army went through Belgium.

I also suspect that had I put the theatre AI on prepare they would have held their ground rather than taking advantage of the Dutch and Belgian weakness, but all in all it was impressive to see them take the initiative.

I suspect some of the issues with Corps/Army/Army Group AI is that it's not entirely aware of the other formations stationed around it that aren't in its control. It doesn't know what you're going to do with them, therefore it must play havoc with it trying to decide what to do.

Theatre AI has impressed me in such that it's done all I've asked of it so far. It will be interest to see how it copes as I redeploy 2 Panzer Armies and 2 Army Groups from Poland to the west. I'm intending to line these up in the low countries, set the objectives as victory points throughout the Netherlands and Belgium initially, before then switching them to northern france, and gradually moving them down towards Paris. I'm hoping I can perform a Schlieffen without jepodising my defensive line to the south, but this might be a bit too much for the AI.

The alternative is to set the Theatre AI with overall goals such as the capitals, then to set more localised objectives for Army Groups... I'll run a couple of experiments over the weekend and see.

 

[notger.heinz]

Gwynn, that is precisely what I wrote, but I guess it was lost in my long post.

When redeploying, I first did it with objectives on army-group level and prepare stance, and when everyone was in place, I gave the army group over to the theatre. This way, the theatre hadn't to move around troops from Todtmoos to Emden.

The conjunction of separate theatre and army group AIs does not seem to work well, for me. As we both observed ... the army group seems to be unaware of the other friendly troops in the theatre while being aware of all enemy troops in the theatre.

I will try an approach with a specialised army, next time, too. I do not know wether the theatre AI knows how to effeciently use special troop types, though I have seen it to attack hills with mountain troops and holding back the tanks. We will see.