Q: Campaign July 1861 vs. Feb 1861 start

[Jan H]

Q: Is it possible, to reenact the the early build-up and organization of the CSA (and USA) troops, ships, forts etc. with a campaign start in spring 61 so that one can achieve the state that is found when loading the July 1st campaign start?

I have been comparing both campaign starts and have tried different starting strategies to mimick that build-up, in particular:
- some 40 k€ troops of 12M service, quite a bit of new rifles,
- lots of small 4th, 3rd and a few 2nd rate ships build or in the building process,
- the Merrimack build at 30%,
- at least 5 new forts along Mississipi and Tennessee river completed incl. Fts Henry and Donelson

and I admit that I have not found a way to reenact the Juliy 1st setup with the pre-Sumter start.

Particularly the Forts (11 M cost per one is huge, 31 buld duration realistically starting in May means I need to form 5 armies to finish construction work until June 30th) and the Merrimack are trouble (subventions need to be build up military and policies enacted, and 30% at Juli 1st means it needs to start building on April 15th or so?).

Has anyone else tried this course?

@Devs: I think the pre-Sumter campaign start needs to get the forts and Merrimack scripted to be coming up after the fall of Sumter automatically... Otherwise it really makes no sense to use the early start when the July 1st provides that much more means to the CSA (and presumably USA will have the same).

@Devs: USA AI offensives (or are these raids?) from USA side sometimes go "too deep": e.g. an Army from Washington/Harrisburg region attacking Greensboro in June/July 1861 and then sitting there for a long while, while all its lines of communication through the shenandoah get closed by up and isolated by CSA... Perhaps this is is an aggressiness thing (mediocre/mediocre settings I have chosen)?

@Devs: AI autoresolved battles take very long and regularly result in high casualties I find (e.g. 1st Bull Run type engagement of 20 vs. 30 k green troops regularly resulting in 5000 or even 7000 casulties on one side and taking 5-10 days. In 1861 I would think most engagements last 1, or perhaps 2 days at max and casualties were low...)

[ghyslain]

Re "USA AI offensives go too deep" : absolutely agreed. In my last CSA run 1.0901 mediocre/mediocre settings, I had 4 x 15-20k US armies go deep behind my line (think North Carolina/Tennessee post Kentucky flip) and just sit there idle building supply depots. In the meantime, I casually captured Washington DC and all of Maryland almost uncontested. Safe to say that campaign was over by August 1861. Lol.

The main underlying issue with the AI causing it to behave in such a way is twofold

a) AI sets its offensive objectives way too deep behind enemy lines without securing the forefront cities/IIP.

b) AI puts way too much importance at maintaining a 100% supply status at ANY given point in time. That impacts the AI both in its offensive and in its defensive operations. That results in the AI having all of its armies constantly marching outside of a supply depot range, stopping there for 30+ days building a new supply depot, marching out of range of said supply depot only to build a new one rinse and repeat. Furthermore, said behaviour is a big waste of resource (manpower sitting idle, cost to build and supply all of those supply depots). AI should instead favour making use of existing IIP/other supply depots when considering what to do to fix its lackluster supply situation. Building a new supply depot should be REALLY low on its list of possible solution.

[biggp07]

Originally posted by Jan H:

Q: Is it possible, to reenact the the early build-up and organization of the CSA (and USA) troops, ships, forts etc. with a campaign start in spring 61 so that one can achieve the state that is found when loading the July 1st campaign start?

It is interesting that you have asked this question because I have thought and tried the same within the confines of my gameplay. I've role played as much as possible the early OOBs and to an abstracted extent what might have been the reality for the CSA during those months leading up to July 61'. (in my mind anyway, I am no ACW expert by any means). Granted, there is necessary imagination and creativity in doing that since it is a simulation game and not 100% representative of all officers leading certain regiments or brigades and many other game aspects that are abstracted and simulated. Plus, the AI is never going to follow historically, so most times it is a bust through that start.
Anyway, to my point, having tried several CSA campaigns since 1.07, I do not think reenactment of early build up to the level of Summer 61' CSA start in game is possible without the help of scripting.
And yes, the "AI" does some weird and clumsy moves sometimes, but I am entertained and have fun experimenting. I have also had the experience of the AI outsmarting me and conducting some strong tactical maneuvers, so it is not completely inept.

[Jean]

The Spring of 61' is much more of a sandbox. I suggest for anyone that places value on immersion and a more closer to history campaign to always play the Summer of 61' as the earliest starting date.

[statboyvt]

Historically, Fort Donelson wasn't completed until early 1862 I believe, and I don't think construction began on either fort until Mid-June. I question whether any of those forts should actually be included in the June start anyway.

[Jean]

Originally posted by statboyvt:

Historically, Fort Donelson wasn't completed until early 1862 I believe, and I don't think construction began on either fort until Mid-June. I question whether any of those forts should actually be included in the June start anyway.

For gameplay purposes, it's better that they are.

[HenryPigLatin]

I used to like the spring start more (and I like it conceptually more), but I have started playing summer more, especially as the union, because of the forts that are in place and makes for a better challenge.

The AI builds too many depots but very few forts.

[Jan H]

Originally posted by statboyvt:

Historically, Fort Donelson wasn't completed until early 1862 I believe, and I don't think construction began on either fort until Mid-June. I question whether any of those forts should actually be included in the June start anyway.

Ok, I actutally don't know when construction on these Forts started and how long it took.

That would mean these 5 forts actually should ideally merely be started constructions with completion dates later that year in the July 1861 Campaign start?
I concurr for CSA AI they should definitely be present or scripted in to be build in both the July and the Feb Campaign starts.
Also I was wonderung whether "Island No 10" should be a fort in game terms and would be missing.

[statboyvt]

Originally posted by HenryPigLatin:

I used to like the spring start more (and I like it conceptually more), but I have started playing summer more, especially as the union, because of the forts that are in place and makes for a better challenge.

The AI builds too many depots but very few forts.

In my latest playthrough the CSA AI is building forts in West Virginia, especially in Charleston.

[statboyvt]

Originally posted by Jean:

Originally posted by statboyvt:

Historically, Fort Donelson wasn't completed until early 1862 I believe, and I don't think construction began on either fort until Mid-June. I question whether any of those forts should actually be included in the June start anyway.

For gameplay purposes, it's better that they are.

Yes, I agree with that. Maybe Island No. 10 as well.

[Jean]

As far as I know, Island no. 10 wasn't really a "fort" until Polk invaded Kentucky and took Paducah and Columbus.

[Jan H]

Playing a few more games, I like to suggest a few more things for consideration to capt. Mueller and the development team to improve the vanilla experience.

1) AI seems to have a strong urge to attack through the knoxville gap, while Grant often sits idle at Cairo or only moves to Island No 10. Would be great if the main axis of advance West were via one via Memphis/Vicksburg/Jackson/New Orleans and the other Nashville/Chattanoogha/Atlanta.
[Side note: Working with the AI related settings in campaignprefs.txt (the weighting of ..., the lower weighting for iips, towns and buildings that are occupied by own faction or non war participants, etc.) I seem to be able to change that, though I have no clue how exactly the imp values in the campaign debug screen are ultimately calculated in detail. A bit more detail in log or onscreen in debug mode would be nice to be able to mod those values more systematically.]

2) AI is "parking" Armies/Corps embarked in River/High seas transport fleets over weeks (sometimes in rear, sometimes near enemy). Why?

3) Is AI using trains in/through enemy territory?

4) Does AI ignore readiness for movement/defensive moves/offensive moves? I see triggers for min morale etc. for movement, but none for readiness. In debug mode I see armies/corps advancing with very low readiness...

5) When creating new armies (and corps), AI picks free new commanders. Often green, while the experienced ones remain attached to their old commands in the field. I'd suggest that AI automatically take an active field commanders instead and replace within the lower CoC new ones.

6) AI keeps marching far into rear (typically e.g., frontline near Alexandria, USA AI aims for Petersburg/Lynchburg or even further south; Frontline near St. Louis, USA AI aims for Little Rock, Warren, Camden/AR). Logs says AI army is carrying out offensive opration, and once in that deep regions say AI getting assigned new theater there. But why so deep isolated from the front?
[Side note: Could in debug mode the borders of "The size of operations theater for AI campaign units x/z" be drawn somehow?]

7) AI activity/aggressiveness generally seems high (mediocre/mediocre settings). Don't know how others see this generally? From the logs I noticed that AI quite often does not carry out offensive moves when entering my territory though. I don't quite understand which triggers/type of moves theres are. I have noticed that AI sometimes goes very close to the frontline or even in my territory to set up forts, which while not really being an offensice, can lead to engagements.

8) AI armies should not detach Corps so oft. E.g. USA AI Army of Potomac in Spring 1862 start often sends Dept. of Virigina as well as 1 to a few Corps to Norfolk, while the rest advances onto Richmond or even return overland to Washington. Besides McClellans Corps' doing a major river crossing of the James at Williamsburg a la Grant at Harrisons Landing 2 years later rather easily, which I find a bit out of place, I think the Army should have stayed closely together: no Corps regularly leaving the command radius of McClellans container.

9) Generall: AI armies. There are those that from name would seem to be assigned regional defense tasks, e.g. "Dept. of Virginia", "Dept. of Maryland", "Dept. of Indiana". However, AI doesn't care much it seems and will rather sooner than later forward them into my territory (not seldom on deep moves ending isolated).
Could AI be enhanced to have such armies/corps that remain local to their department, only transferring units in and out to others but not moving much, and especially never leaving their region?

10) Battle AI: I'd still like to suggest that in battle mode crossing hills/forrests remains way too easy. Cumberland Gap really is no terrain obstacle. Would you be able to improve this, or could the pathfinding perhaps be opened up to modding?
(e.g. https://steamcommunity.com/app/654890/discussions/4/3273562752738422349/?ctp=3#c4615641483000122709)

11) Battle AI: More objectives on the battle maps, but whole AI armies/corps please not follow them in a chain one by one as whole groups... please allow AI to use battle space and form complex, non-straigth lines of width like known from Gettysburg of Chancellorsville, or any other battle.... AI still tends to bulk together often...

12) River transport should be split up into a West and East component I think. It is way too easy to shift USA river shipping to Virginia to move huge armies there and a few days later (in mid 1862), move grants whole army along the Mississipe, all the while the latter is still largely under CSA control...

Thanks for supporting this game! I really have spent many fun hours with it now.
Curious, what the new DLC will be!

[A_P_Hill]

I almost always play CSA, and I always pick early spring 61 start. I am not really into trying to replicate 100 percent of early war history as I am trying to get a leg up on the Union build up to come.

[Jan H]

Ok, let's see whether someone can help me, or whether the devs could chime in to explain the parameters in the campaigns prefes some more. I'm trying to adjust aggressiveness of offensive macro and micro-moves (not offensive stance).

Turns out according to the log that it is often micro moves in offensive mode that lead AI containers to move deep into enemy territory/too deep offensives or that even sometimes see AI do strange things (e.g. start summer 1861: USA AI "moves" from Lex/KY via London/KY, further via noxville and Chattanoogha to aim for Nashville -- a rather long route through my CSA territory for attacking Nashville?).

These params have at least changed target selection of campaign AI so that deep offensives don't happen anymore:

-----
the weighting of the transport capacity for the calculation of ai importance values
0.01
the weighting of corporate profits for the calculation of ai importance values
0.000
the weighting of population for the calculation of ai importance values
0.000005
the additional weighting for capital (absolute)
200
the value for each soldier for the ai importance value calculation, own units
-0.002
the value for each soldier for the ai importance value calculation, enemy units
0.001
the value for each soldier for the ai importance value calculation, neutral units
0.0005
the maximum distance from ai area to ai area to search for areas to move units to
100.275
the distance up to which enemy units prevent friendly unit repositioning
150
...
...
the lower weighting for iips, towns and buildings that are occupied by own faction or non war participants
0.35
...
influence of commander initiative on offensive/defensive stances
0.6
needed zone strength relation for offensive stance (is adjusted by commander initiative)
0.8
minimum morale for offensive stance
0.9075
...
Zone importance factor if no enemy units are close [should be <1]
0.95025
Lower weighting for troops that are far away from ai zone [<1]
0.35
------

However, AI is a bit too aggressive now for 61/62 (and weakens itself in to many offensive before training and equipping up).


MY QUESTION:
I tried addressing this by varying the below parameters from very small ->0 to very large >10, either each group alone or in various combinations. However, these don't seem to do anything? Are those disabled parameters?

------
AI aggressiveness per chapter (probability of offensive operations #1-5)
0.00
0.75
1.5
2.0
2.5
AI agressiveness leverage for campaign games (min/max)
0.1
2.5
...
The AI probability for Union and CSA for evaluating offensive operations
1.0
0.5
------