Experimental mod for AACW

[Pocus]

Is this a test with the 1862 scenario or the 1861 scenario?

Anyway, outlook looks good here, any chance of a more precise layout of changes (i.e., exactly the changes to each unit, etc.).

I believe that the AI takes a lot of factors into play. I did some changing of VP areas, and it seems like the AI does react to these changes, but, it also shows that there are other factors. I believe that priority is placed on depots as well as VP (probably other factors, like city size, fortress, etc.). Yet, modding VP has improved things (like "The Valley" no longer being the primary field of battle).

The AI does love Fredricksburg. In my tests I fortified Manassas, and the Federals went straight for Fredricksburg. When I left Manassas, they went for it. It does seem that the AI does have a high priority to move into the least defended area with the maximum force avaliable.

VP can help the AI, but it cannot directly control it.

As you can expect, the value of a region is dependent of many things. VP and objective value yes, but even the number of land links and the % of loyalty is factored somehow in the evaluation.

But the bottom line is that whatever the interest of a region, the AI won't suicide units in taking it or defending it, if the outcome is too predictable. This is for examle why the AI will abandon Washington if it is not possible to hold it (and the relocate the capital). This certainly can be discussed, because in some case, a 'defence to the death' can bleed the attacked and achieve positive results (a bit like the no retreat order of Hitler in winter 41).

[Clovis]

How war evolves in the West ( yet sticking as US player to the historical plan)?

Around the Ohio army led by Buell, you may see the CSA army of the Tennessee with Polk's corps




In the mod, AS Johnston appears with a HQ ( I reduced CSA pool by one HQ) so the Army of Tennessee is always forming and deploying in the West



Grant has taken Fort Donelson and got a promotion. Now the difficulty lays in his relatively low seniority. At this time, I've yet Halleck and Butler who will be angry of Grant nomination to army command. In the contrary, nominating Halleck would spare NM points.

[Clovis]

about results in march 1862

The relatively low NM total for both sides is coming from :

- for Union, the army creation consequences. Butler costs me 8 NM points as I've no HQ to devote to him.

- CSA NM is lowered by the use of conscription as I've slighty raised the cost in NM of the option.

So far this loweringis yet reasonable as neither side is about to collapse

[Clovis]

Why I love this game ( even if AI is until now almost totally blurred in the Western theater...

As Union, I guess Beauregard army is going to Harrissonburg to spank Bank corps. So I decide to seize Fredericksburg.

But Beauregard turns clovckwise , invade Manassas and routs my division here then goes to Alexandria.

Potomac army and Washington are now under menace

[McNaughton]

I notice that the Union AI likes Fredricksburg, as it is a 'direct' route to Richmond. However, without Manassas, the Union force cannot supply itself, and usually withdraws.

The key, for the East, is Manassas. It may be good to make this city slightly more important for the AI, maybe making it a victory city (not just a VP city)?

Also, it may be good to put VP at all 'towns and depots' out east, seeing that they are all critical to the AI as the Union or Confederates.

In order to march on Richmond or Washington the following cities are critical.

Alexandria (moreso for the Union), Manassas, and Fredricksburg (more direct route, but the Western rail line is a good supply).

Also, possibly having more depots (at least supply depots), may help the AI and player keep their forces operating in the region.

[Pocus]

If you goes that way, perhaps you can try to do as they did with Elmer, the AI of Operational art of war. It advances along axis determined by a serie of location to capture.

In AACW, you would create an event, for the AI only (you can check the AI level of a given faction), which would give an Objective status to Alexandria only , if neither Alexandria, Manassas or Fredericksburg are under control. (first step)

Then a second one, which would give Manassas as an objective city (and remove Alexandria) if only Alexandria is under control.

Then a third one which would gives Fredericksburg as an objective city (and remove Manassas ) if Alexandria and Manassas are under control.

Just an example. The script engine has more than 120 commands and is underused for the most part of it.

[PBBoeye]

Just an example. The script engine has more than 120 commands and is underused for the most part of it.

And not understood by most of us at this point.

I wasn't aware the AI could be led by the nose like that. That's good to know, but it would be better to understand the scripting better so we could make good events.

[Pocus]

Yes I know. I have simply no time to write a comprehensive modding manual , which would anyway lack many of the things people would ask.

What I hope is that you the modders, you continue to ask questions on how to do things, then you add them to the Wiki once you get the picture.

Also Primasprite is fairly advanced on modding knowledge, he is event writing something to help people on that. So he can help you too.

What I ask you is to not let the info disapears. Once you get a new technique or learn how to do an even, please write something in the wiki.

[Clovis]

BTW, AI seems to get problem with units having " emplaced guns" trait. Wgen such a unit is captured, by example by an army corps, the corps seems to remain unable to move and the AI doesn't split the stack in order to get immobile unit out of the stack.

[Clovis]

About AI, as the agressiveness level can be altered, the major drawback IMHO remaining is its unability to really cope with 3 star leaders. AI is unable to replace one 3 stars commanding an army by another better ( Lee syndrom), is unable to repalce a 3 stars by a 2 stars to lead a corp ( when a 3 stars is much more useful even without an HQ to lead an indepenadant force.

[PBBoeye]

Good point. I think this follows in the area of the AI needing to be able to manipulate and use command structure better. Mentioned in another thread, but this is a good example of it. I think it will give the AI a whole new set of teeth.

That's probably not understood by many Europeans. I think it's an American phrase.

[McNaughton]

I have been thinking about this 'dilemma', about leaders not transferring command.

IMO, it would be an interesting feature to add seniority being the determining factor of who commands an army (general with the most in a region, gets the command), if General A, who commands an army has 10 seniority, while General B moves into the region, and has 4 seniority, the command goes to General B.

However, from a modding point of view, I have been thinking of a few historic transfers of command (to make things 'interesting') until if/when a change happens via AGEOD.

Albiet somewhat scripted, I think we could get the AI to follow command change based on historic plausibility.

1. A specific event based on time or situation or chance triggers the situation, checking if the 'offending player' is AI (this should not happen for human players). Can this be done?

2. The 'existing' general army general is 'removed' via event (along with their HQ).

3. The 'incumbent' general is removed from their situation (easy for someone like Lee who is locked, tougher for McClellan who may end up being a divisional commander!).

4. The 'existing' general (former commander) appears at your capitol, ready for a fresh command (or eliminated in the game based on history).

5. The 'incumbent' general appears in the former general's command along with a HQ.

6. The AI will form the new HQ with the 'incumbent' leader on its own.

--HISTORIC EXAMPLES--

I think that the following four examples of command changes are the most 'critical' in game (since out West, the command is more fluid, and more 'armies' were floating around).

1. Johnston takes command from Beauregard
i) Done by timeframe (sometime in the winter of 1861)
ii) Easy, since Johnston commands a corps, Beauregard an army (no divisional units)
iii) Beauregard and his HQ are removed from the game
iv) Beauregard appears via event in Richmond, locked for a turn, while a new HQ is placed in Johnston's stack (the AI will most likely give him command).

2. Lee takes command from Johnston
i) Done via event (Johnston's command in heated battle)
ii) Easy, since Johnston commands an army, Lee is locked
iii) Johnston is 'wounded' and inactive for a few turns
iv) Johnston appears in Richmond, locked for a few turns, Lee is unlocked and a new HQ appears in Richmond (most likely the AI will give it to Lee)

3. McClellan takes command from McDowell
i) Done via event (if certain towns in West Virginia are friendly controlled, Richomond still Confederate controlled, representing McClellan's success there, transfer sometime in winter 1861, and McDowell's failure)
ii) Tough, since McClellan could be given divisional command, but McDowell will be Potomac leader
iii) McClellan is removed from his current command (hopefully just corps commander!), along with McDowell and his HQ
iv) McClellan and a new HQ appears in the largest Eastern stack (McDowell appears back in the capitol)

4. McClellan loses command
i) Done via event (if Richmond not captured by a certain date)
ii) Very easy, since this just requires McClellan
iii) McClellan and his HQ are removed from the game (representing McClellan historically being removed from the war)
iv) a 'random' 2-star general is promoted to 3-star (could be Hooker, Burnside, Franklin, etc.)
iv) new HQ appears in the largest Eastern Stack (given to the 3-star general with most seniority)

--PROBLEM 2--

Another problem I have found is that the AI is not keen on producing Army HQs. I think that the cost is prohibative, and the AI rarely makes the choice at producing a new HQ. I think that AI events, feeding in 2 HQs in late 1861/early 1862 to represent the 'main' armies out West, and Far West, will help the 'difficulty' of the game (one army HQ means that corps can be formed, thereby increasing efficiency).

Events could be in place, specifically for the AI, to provide army HQs in the two western regions (one set of events checking for a Union AI, another for a Confederate AI).

[PBBoeye]

1. A specific event based on time or situation or chance triggers the situation, checking if the 'offending player' is AI (this should not happen for human players). Can this be done?

Yes, that is indeed the question. I would add that if not, then it would behoove development to create this scripting capacity. The reason is simple - it allows us to do for the AI via event, what the AI is not currently doing via coding. That is crucial, and to me that was what allowed EU/EU2/HOI/VIC to succeed in that the balance could be struck between scripting and AI behavior.

So I am really pushing for this if it does not exist currently. Pretty important.

Another problem I have found is that the AI is not keen on producing Army HQs. I think that the cost is prohibative, and the AI rarely makes the choice at producing a new HQ. I think that AI events, feeding in 2 HQs in late 1861/early 1862 to represent the 'main' armies out West, and Far West, will help the 'difficulty' of the game (one army HQ means that corps can be formed, thereby increasing efficiency).

Events could be in place, specifically for the AI, to provide army HQs in the two western regions (one set of events checking for a Union AI, another for a Confederate AI).

Yea. Count me as a Yea on that. It will be interesting to see how much better the AI performs with adequate HQ allotment.

The other option would be to get the cost for an AI HQ down, but that would require.... I don't know. Maybe just best to do the event trick.

I always like to see if the AI can be cajoled into doing something if you make it attractive enough - in this case, a lowered HQ purchase value. But who knows if the AI would then do it properly?

[McNaughton]

I guess, in regards to the planned objectives, are there event commands for the following?

1. A switch to check for if the faction is AI controlled or Player controlled?

2. A switch to check for specific control over a region? (most likely, something like the "Papers Push for Offensive" event)

3. A way to check if a unit or commander was recently in combat?

4. A way to check if a commander is in command of a division? (useful to determine if a leader should be removed via event or not)

5. A way to remove something from a stack (say a leader) as well as for the event to track down that same stack and add something (say a leader and a HQ)?

6. A way to check if a commander is in command of a particular army?

[PBBoeye]

Oh, I would add in that 'Grant in the East' is crucial. But perhaps the game already does that automatically?

[McNaughton]

Yes, that is indeed the question. I would add that if not, then it would behoove development to create this scripting capacity. The reason is simple - it allows us to do for the AI via event, what the AI is not currently doing via coding. That is crucial, and to me that was what allowed EU/EU2/HOI/VIC to succeed in that the balance could be struck between scripting and AI behavior.

So I am really pushing for this if it does not exist currently. Pretty important.

That's what I was thinking, basing things on EU2/HoI/ViC modding experience that there were many possibilities (along with leaders, getting 'rolling' vp events, to help the AI strategize their 'attacks').

Yea. Count me as a Yea on that. It will be interesting to see how much better the AI performs with adequate HQ allotment.

The other option would be to get the cost for an AI HQ down, but that would require.... I don't know. Maybe just best to do the event trick.

I always like to see if the AI can be cajoled into doing something if you make it attractive enough - in this case, a lowered HQ purchase value. But who knows if the AI would then do it properly?

You can bump up the priority the AI looks at army HQs when it comes to purchase time, but the cost is still very high and I don't know if it 'retains' enough cash by the time it comes to army purchase time to effecitively utilize it (another 'problem' could be the AI over-buying HQs, getting every HQ that appears when they most likely will work at peak efficiency with 3-4 HQs).

[Clovis]

Yes, that is indeed the question. I would add that if not, then it would behoove development to create this scripting capacity. The reason is simple - it allows us to do for the AI via event, what the AI is not currently doing via coding. That is crucial, and to me that was what allowed EU/EU2/HOI/VIC to succeed in that the balance could be struck between scripting and AI behavior.

So I am really pushing for this if it does not exist currently. Pretty important.



Yea. Count me as a Yea on that. It will be interesting to see how much better the AI performs with adequate HQ allotment.

The other option would be to get the cost for an AI HQ down, but that would require.... I don't know. Maybe just best to do the event trick.

I always like to see if the AI can be cajoled into doing something if you make it attractive enough - in this case, a lowered HQ purchase value. But who knows if the AI would then do it properly?

That 's I've done with AS Johnston who is appearing with an HQ.

The AI has built another HQ but has decided to build a new army under Bragg into Virginia...the lesat needed place.

[Pocus]

I needed the relocate order for a proper and efficient relocation of army HQ and high level leaders, and it is done at last. Now, it will be easier to implement improved behaviors dealing with the Command Chain. The next patch already feature the redeployment of Army HQ across the map with that.

My last test was one where Grant (under AI command) in KY received by redeployment an HQ just produced in Washington, so things bode well.

I will investigate for this HQ buying problem. Some special units should follow special rules...

[Pocus]

I guess, in regards to the planned objectives, are there event commands for the following?

1. A switch to check for if the faction is AI controlled or Player controlled?

2. A switch to check for specific control over a region? (most likely, something like the "Papers Push for Offensive" event)

3. A way to check if a unit or commander was recently in combat?

4. A way to check if a commander is in command of a division? (useful to determine if a leader should be removed via event or not)

5. A way to remove something from a stack (say a leader) as well as for the event to track down that same stack and add something (say a leader and a HQ)?

6. A way to check if a commander is in command of a particular army?

1. yes
2. yes
3. no
4. no
5. yes, but this will remove the army or corps status. You would need a replace command
6 no

[PBBoeye]

Check the WIKI in a few days as I hope to have some definitions up for these soon.

[Clovis]

Here's the result of a full 1862 battle between the 2 main armies in Eastern theater ( in game equivalency of Antietam)



No one regiment destroyed.

[PBBoeye]

Is there some way you can document what you've done? The best thing would be to have the units and models XLS files adjusted. I really, really like this mod but putting the changes in XLS format makes it easy to reproduce the files when changes to the major files are made.

I've done a test run with the CSV Splitter program, and it's very easy to create your own files. So I hope you will consider doing an XLS version of your mod. Otherwise, I can tell you the devs won't look at it, and I think eventually it will get lost in the patching progress.

Too good a mod for that!!

If you don't have Excel, you can use OpenOffice.org (freeware). Or I can help with the data entry, but I don't want to have to open every damn file!

[Clovis]

Is there some way you can document what you've done? The best thing would be to have the units and models XLS files adjusted. I really, really like this mod but putting the changes in XLS format makes it easy to reproduce the files when changes to the major files are made.

I've done a test run with the CSV Splitter program, and it's very easy to create your own files. So I hope you will consider doing an XLS version of your mod. Otherwise, I can tell you the devs won't look at it, and I think eventually it will get lost in the patching progress.

Too good a mod for that!!

If you don't have Excel, you can use OpenOffice.org (freeware). Or I can help with the data entry, but I don't want to have to open every damn file!

Thanks.

I will do that, I'm using Openoffice (and generally the most possible opensource programs.).

My real constraint is time. I must admit I'm doing this mod to play with as from the start I feel the 1861 situation isn't rightly addressed by the game engine ( that's not to say AACW is a bad wargame- I'm playing wargames (board or computer) since 25 years and AACW is without contest one of the five best I've ever played).

So I started by modifying game files directly, because the mod files weren't on board.

So now I would need time to transfer all the modifications in the csv files, but I prefer for now continue to experiment until I get a satisfactory results.

This objective being of course delayed each time a new patch is made

The good news is I've currently a mod version working with the 1.07 version.

I suppose it will be yet another big patch after NCP will be completed, as AI is supposed to be enhanced in the new game whose mechanics seem very close to AACW.

[Clovis]

Now I've a running version under 1.07, I'm going to introduce a new change for recruitment in order to get more historical feeling.

First, draft option will be delayed until 1862.

Two, I will give both sides some free conscript points in the first turns of the game, simulating the afflux of volunteers at the eve of the war ( before the real side of war revealed itself...)

Last, I've created a great difference between USA and CSA volunteer number results. Basically, USA will get 1.5 volunteer for 1 CSA at the same bounty level. But USA drafting will be slighty less proficient than CSA one...

So USA should rely more on volunteers and less on draft but will need a lot of money, or choose an alternative path with more drafting proportion... On the contrary, CSA should have more incentive for drafting.

Now the grat unknow is as always... AI reaction. Is the AI able to compute the disadvantage or it will be hardwired to go after any recruitment source...

[McNaughton]

One thing that I did for my mod was a hefty re-design of the July 1861 scenario (started in July so testing game aspects would be easier, as more units are on the map).

Pretty much removed most of the 'militia brigades' and added more 'volunteer' (what the game calles regular) brigades, representing the 'federalizing' of state forces by both the CSA and USA.

Here are some additional starting forces... (in case you are interested, and don't want to have to do all of the research)

--CONFEDERATE--

Added units to Army of the Shenendoah

Smith's Brigade (the brigade currently in game is Ezley's Brigade, Smith commanded a different force that did not reach 1st Bull Run)


Added Army of the Northwest

Beverly WV (in Grafton in April)
Brig.Gen William W. Loring
Jackson's VA Bde.
Garnett's VA Bde.
Rust's AR Bde.


Added Army of the Kanawha

Beverly WV (in Clarks, WV in April)
Brig.Gen John B. Floyd
Floyd's VA Bde. (later moved, with Floyd to Fort Donelson)
Wise's VA Bde.


Added Independent

Harrisonburg VA
Hatton's TN Bde.


Changed Army of the Peninsula (formerly MacGruder's Command)

D.H. Hill's NC Bde.
Pryor's AL Bde.


Changed Huger's Command

Mahone's VA Bde.
Wright's GA Bde.


Changed East Tennessee Department (Zollicoffer's Command)

Rain's TN Bde.
Zoilicoffer's Bde.


Changed Department #2 Cavalry (CSA Cavalry)

Forrest's Cav. Bde. (these regiments were quickly organized into a brigade)


Added (to Hardee when he appears)

Shaver's AR Bde. (this was initially Hardee's Bde., of federalized pre-war militia)


Added Arkansas Department (McCullough's Force, were all pre-war militia, some federalized, some still state troops, wich fought in 1861 battles in Missouri)

Pearce's AR Bde. (Militia, were state troops)
McCullough's AR Cav. Bde.


I am still revamping the Missouri State Guard (trying to come up with a combination that represents their size, at around 6000, but under strength and ill equipped, but with the possibility of increasing in size if allowed to do so by the USA).

Let me know if you are interested, and I can post the USA changes.

[PBBoeye]

Pretty interesting stuff all around.

Yeah, I will be interested to see how the AI reacts. I'm hoping... maybe against hope, but still hoping.

Anything to get a more realistic portrayal of manpower issues and cut down on these uber-armies, especially early on.

[PBBoeye]

Any release date coming up?

And how about a synopsis of your changes, if you don't mind.

[Clovis]

One thing that I did for my mod was a hefty re-design of the July 1861 scenario (started in July so testing game aspects would be easier, as more units are on the map).

Pretty much removed most of the 'militia brigades' and added more 'volunteer' (what the game calles regular) brigades, representing the 'federalizing' of state forces by both the CSA and USA.

Here are some additional starting forces... (in case you are interested, and don't want to have to do all of the research)

--CONFEDERATE--

Added units to Army of the Shenendoah

Smith's Brigade (the brigade currently in game is Ezley's Brigade, Smith commanded a different force that did not reach 1st Bull Run)


Added Army of the Northwest

Beverly WV (in Grafton in April)
Brig.Gen William W. Loring
Jackson's VA Bde.
Garnett's VA Bde.
Rust's AR Bde.


Added Army of the Kanawha

Beverly WV (in Clarks, WV in April)
Brig.Gen John B. Floyd
Floyd's VA Bde. (later moved, with Floyd to Fort Donelson)
Wise's VA Bde.


Added Independent

Harrisonburg VA
Hatton's TN Bde.


Changed Army of the Peninsula (formerly MacGruder's Command)

D.H. Hill's NC Bde.
Pryor's AL Bde.


Changed Huger's Command

Mahone's VA Bde.
Wright's GA Bde.


Changed East Tennessee Department (Zollicoffer's Command)

Rain's TN Bde.
Zoilicoffer's Bde.


Changed Department #2 Cavalry (CSA Cavalry)

Forrest's Cav. Bde. (these regiments were quickly organized into a brigade)


Added (to Hardee when he appears)

Shaver's AR Bde. (this was initially Hardee's Bde., of federalized pre-war militia)


Added Arkansas Department (McCullough's Force, were all pre-war militia, some federalized, some still state troops, wich fought in 1861 battles in Missouri)

Pearce's AR Bde. (Militia, were state troops)
McCullough's AR Cav. Bde.


I am still revamping the Missouri State Guard (trying to come up with a combination that represents their size, at around 6000, but under strength and ill equipped, but with the possibility of increasing in size if allowed to do so by the USA).

Let me know if you are interested, and I can post the USA changes.

Of course. I'm starting to investigate too OOB for both sides at start and units model.

many thanks

[Clovis]

Any release date coming up?

And how about a synopsis of your changes, if you don't mind.

Waiting for the final 1.07 patch, taking my time to implement this in a less "quick and dirty" way

I hope before the end of october to put all change in XLS files too

[PBBoeye]

I really hope so because the work you are doing on CSA manpower via drafts and so forth could really change the official model if we can get a great effect. I say 'could' because I don't know how they'll ultimately view things. But ideally we'd like to get CSA numbers down to historic levels, and then address troop quality afterwards. Right now the CSA stands strong by means of it's historic leader quality and ahistoric troop numbers.

So your work is valuable to the community as a whole, is what I am saying.