Full Campaign Report (vs AI)

[Paul Roberts]

I recently finished a full campaign against the AI, and I thought I would record the results (with a few comments) here.

This was the Full Campaign starting July 1861. I played CSA with almost all options set to default/normal (including normal difficulty, low AI detect bonus, normal aggression, and "give AI more time"). I started the game under 1.08 and moved to 1.08b and 1.08c as these became available.

In brief, I won the game after a genuine struggle. Score was CSA 4202 to USA 1959 when the game ended in late August 1864, but things were often closer than that final ratio makes it seem.

The Eastern Theater:

The war in the East followed the historical pattern: after the CSA won a sound victory at first Manassas, the Union protected Washington, DC in strength and even attempted a peninsular landing. When I forced Butler out of Fort Monroe in late 1862 (after a couple of failed sieges) such landings were no longer possible for the Union, so they shifted efforts to forays south from Maryland.

Under my command General Lee made three invasions of the north. The first (Fall 1862) smashed up against strong Union armies just across the Pennsylvania border and came reeling back after heavy losses on both sides. (This first invasion coincided with the first heavy Union push towards Richmond--the campaigns began simultaneously and on parallel, opposite courses.) Lee's second offensive (Summer 1863) laid siege to Washington, but only briefly. In a great battle Lee defeated Union forces at the capital's fortifications, but his own heavy losses caused Lee to retreat to Virginia.

Lee's final offensive began in May of 1864. Leading his own core force and accompanied by two 65,000+ Corps under Jackson and Johnston, Lee pushed directly for Washington and lay siege to its fortifications. A series of hard-fought battles through the summer months left Lee's force unbowed. Finally, in late August Lee's army smashed into the city and seized the White House. The game ended immediately in CSA victory, as the Union NM had dropped to 45, a full 15 points below its defeat threshold.

The Trans-Mississippi Theater:

After a series of small actions in Missouri in 1861, the war beyond the Mississippi settled into a pattern of infrequent skirmishes and small raids. I never had more than Price's forces, Watie's indians, and few others in the theater (including a small cavalry division under Van Dorn), but the Union did not seem interested in pushing hard in this region. It became more quiet as the war went on.

The War in the West:

The pattern here was the most unexpected. Kentucky seceded as usual, but a series of sharp engagements in the newest Confederate state left it mostly in the hands of the Army of Tennessee under A.S. Johnston. I strengthened this Army from time to time, but it never came to match the weight of the Army of the Potomac under Lee.

The Union seemed to allocate some of its best commanders to the Kentucky theater, but they seldom made more than a token effort towards taking the state, let alone moving down into Tennessee or advancing along the Mississippi river. Every season a small force would cross the Ohio and be beaten back. It seemed to me that the Union AI was holding back substantial forces, perhaps even forces that could have wiped out A.S. Johnston's had they made a concentrated effort. Almost every turn I saw faces familiar from Union history pacing the Ohio river, but they never made good on their apparent threat.

Such was still the case when the war ended, although Spring and Summer 1864 did see some interesting river forays (mostly exploratory, it seemed, or maybe just fishing), as well as a small landing at Pensacola, Florida.

Conclusions and Questions:

I'll list these as numbered questions:

1) I greatly enjoyed this campaign against the AI, although the Union's behavior in the West perplexes me somewhat. It was definitely more aggressive in Virginia than in Kentucky/Tennessee. Are there different "theater-level" AI's at work on the computer side? Does the aggression setting under Game Options tend to unbalance the game (e.g. causing the AI to make kamikaze attacks)?

2) I noticed that "epidemic" events rarely ever cost me strength, only cohesion. Is this WAD?

3) This is a very stable game. I had not one crash!

4) The final tally of losses was 271,983 Confederate casualties and 404,704 Union. This seems plausibly close to history (which saw CSA losses of 93,000 killed + 123,000 wounded; Union losses of 110,000 killed + 275,000 wounded). I was playing without the hard attrition rule enabled. I don't recall any single battle result striking me as ridiculous.

5) This is the only full campaign that I have played to completion against the AI. I'd love to hear what other players' experiences have been.

Let me know!

[Banks6060]

In reading your story I find several similarities to my own campaign. I too am playing against the AI as the Confederates and find it remarkably odd that there have been no significant attempts so far by the Yanks to cross the Ohio.

I am admittedly only in Spring 1862 and not very far into the whole experience, but I will also say that overall, against the AI, the game hasn't proved much of a challenge at all. I wonder if it is designed that things are much more difficult when playing as the Union, or if, as with many other strategy games, the AI just can't compare to the strategic brain of another human being in the level of its decision making.

As in your game, there has been NO attention payed to the Missouri Theater in mine. Something which I plan to fully exploit. The Union continues to plunge itself into Northeastern Virginia only to be beaten back time and time again with staggering losses. I have on SEVERAL occassions decided not to attack Washington simply because I wanted to keep playing because the AI would do things like move their entire Eastern army back to Baltimore or Annapolis and leave several small detachments in Washington.

It's been laughable at times. I suppose these are all things you have to take in stride when playing against the computer though. Maybe you and I can match up on a PBEM game or something. I'd really like a strategic challenge after what I've experienced so far.

[veji1]

Indeed, it seems to me that playing the CSA against a UNION AI should be at least a bit challenging. The AI, if not very clever should at least be able to use the Union's sheer numbers to build up strong offensives in the West and the TransMississipi. The way I see it the AI should focus on concentrating its forces on its main bases and move forward once they have enough forces to hold the territory.

Take the April 1861 campaign : In the TransMississipi the AI must : Concentrate at Saint-Louis, Move forward to Rolla with a strong concentrated force leaving a strong garrisson (300 or 400 pwrs worth) in Saint Louis. Build depots and Garrison them heavily, rinse and repeat.

In the West it should concentrate at Cairo and Louisville, have a strong garrison in all depot cities (builiding up to 200 or 300 pwrs), and from there proceed forward with an army flanked by 3 corps (one in front, one on each side, with each corps touching another) so that it has a mutually supporting force that is wide enough to envelopp CSA forces if they concentrate in a single place..

Even without fancy amphibious operations, and better still in conjonction with them it would at least force the CSA player to be cautious and prepared, I mean I never actually have to protect central Kentucky because the Union never tries to go for it methodically...

The 1864 Grant method isn't to complicated for an AI : Concentrate and bring forward... It should be able to keep the player a bit of balanced..

It seems to me that in this game the AI is a bit too "intelligent" for its own good : It watches the player too much instead of focusing on its own objectives..

In the East I believe their is an "objective" issue : The Union AI should consider Harpers Ferry and Winchester secondary objectives ! The priority is too keep a strong corps at Alexandria, move to Manassas, if it's a defeat fall back on Alexandria, if it's a victory consolidate...

One of the main issues in the east for me is the inability of the AI to keep entrenchments : It seems to move all its troops around all the time : A Corp should stay entrenched at Alexandria all the time, same for the province north just accross the Potomac and West to Washington... The Harpers Ferry area should be a defensive theater of operations for the AI... It should attack towards Manassas, Stafford, Fredericksburg, Charlotteville... etc.

[Banks6060]

Just sat down for another session last night. Incredible. The Union AI does it again. retreating after yet another battle in Mansassas, this time with a relatively lighter casualty count, all the way back to Annapolis. Leaving an open road to Alexandria and Washington for two full Confederate Corps easily numbering over 50,000 men and each with pwr ratings over 1100.

On top of this, Union advances appear to be terribly segmented and their attacks even more uncoordinated than they historically were. Now I'm not one to bitch about things being EXACTLY the way they were, but the AI was stupid enough to send one Corps against an entrenched position occupied by two Corps of my own. There were no "March to guns" orders executed by the AI as no Yank reinforcements showed up, proving that the AI never intended to move more than the solitary Corps into my position. I like winning, but I also like a challenge and the AI offers none.

They are still just milling around the Ohio river at Evansville and Cairo...gathering forces for sure, but showing no signs of operating further south...(cavalry raids and recon, river blockade, etc....)

This really has been the "dream game" for me, but the AI certainly hasn't lived up to the task. I am officially posting a new thread to ask for a human opponent.



Let me also add that the AI's organizational skills really seem to be lacking. I keep facing Armies filled to the gills with eroneous units like Naval Engineers and Marines. maybe those were present in historical field armies operating in Northeastern Virginia? I dunno.

[Bertibums]

@ paul roberts: maybe if you tried to increase difficulty and/or aggressiveness, the AI would put up more of a fight? i mean you can't really blame a game for not being hard enough if you haven't tried the max. difficulty settings, can you?

[Paul Roberts]

@ paul roberts: maybe if you tried to increase difficulty and/or aggressiveness, the AI would put up more of a fight? i mean you can't really blame a game for not being hard enough if you haven't tried the max. difficulty settings, can you?

Oh, believe me, I was not complaining. I hope I made clear that I had an very enjoyable campaign against the AI (something I can't always expect with other games!). My notes on the Union's behavior in Kentucky were aimed at constructive tweaking.