|
jscott991 -> RE: The AI CSA in the West (6/14/2009 5:24:03 AM)
|
1st Turn Power + 0 CSA resource income = 191, 93, 30, 100
1st Turn Power +2 CSA resource income = 264, 118, 43, 139
That's a big jump. But it means nothing in terms of behavior. Three games played now through late 1863. Three times the AI pulled his forces east (gradually and not in a rush).
In the last several games, I did NOT build Meade's second eastern army up higher than 50,000-60,000 men. This meant that the CSA AI actually outnumbered me at the end of the game in the east (about 150,000 Union field forces to 161,000 CSA troops in Richmond). In the game I just quit, he actually beat me in Fredericksburg and took out the fort there, plus Fort Monroe.
Of course, he still lost the game. There wasn't a single container in Miss., Louisiana, Tennessee, or Georgia, and the game ended when I took Savannah, New Orleans, and Little Rock.
So, it's lose-lose. If I pack the West, he gives up and marches east.
If I keep men in the east, he panics and marches east.
But no one else sees any of this. It's remarkable. This happens in every single game I play, whether its power +0, power +1, or power +2 (in fairness, I stopped the +1 game pretty quickly, deciding to just go straight to +2). One thing that bugs me about this thread, is that people seem to think this all happened in one game. There are five different games documented here. 4 on First Sergeant +0 and 1 on First Sergeant CSA +2. I could have documented three more +2 games (and still can document one of those), but it doesn't seem to matter. The consistent theme is that as the Union army gets bigger, the AI simply surrenders the game by marching his entire army to Richmond and either sitting or bouncing back and forth attacking Fredericksburg or Rappahannock (and even after a victory, he withdraws to Richmond). Gil R. argued that the problem is that the AI has too few resources to compete on First Sergeant. I submit that's flatly wrong. The AI can't compete regardless of its resource base. I submit what to me is obvious: the AI can't compete in this game competently at all; the higher difficulty levels MASK this by penalizing the human and stacking more bonuses on top of the CSA's already formidable combat advantage. When you distort economic and combat reality to that point, are you really playing the Civil War?
So for the AI to play semi-competently, I need to have fewer brigades than the Union had historically, and I need to be suffering from a combat penalty.
Terje - I do win lots of battles agains the AI where he suffers terrific losses (he doesn't understand the basics of quick and instant combat; instead of leaving an army in reserve to reinforce a 27 brigade army, he attacks with tons of brigades at once, which is easy fodder for a human with one 35 brigade army and another waiting to reinforce; I refer to this as the ping-pong effect; he bounces back and forth between Rappahannock and Fredericksburg, getting savaged by McClellan calling on Meade or vice versa; this would NEVER happen if he sat in Fredericksburg early in the game instead of Richmond, but he won't do it, another major AI flaw). But I never allow the AI to lose more than 2 brigades by surrender in a single combat. I reload the game, regardless of the outcome, if he does. Despite this the AI does suffer heavily from the brigade surrender chance after a quick or instant combat. He loses at least 10 brigades over the course of the war just from his insistence on attacking big containers with little ones and then instantly retreating. That's why I would love to turn this chance off entirely. At this point, I don't even think a reduction is the answer. It needs to go.
Edit: Oh, one other question, raised by Terje. The western AI doesn't seem to consolidate its units into army containers nearly as well as the eastern. Little containers are dangerous and lead to more brigade surrenders.
|
|
|
|
