Matrix Games Forums

 Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

To many early withdraws

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [Current Games From Matrix.] >> [American Civil War] >> Forge of Freedom: The American Civil War 1861-1865 >> Public Beta Feedback >> To many early withdraws Page: [1]
Login
Message << Older Topic   Newer Topic >>
To many early withdraws - 11/13/2007 3:05:00 AM   
dolphinsfan9910

 

Posts: 81
Joined: 6/25/2006
Status: offline 		I definately agree that in the face of definate defeat an army should retreat. I'm glad the days of 3 brigades taking on an army are gone........But where are my get down drag out battles in which I win by a miracle?

One of my strategies for the South is to plunder and get the North's will down. I usually shoot for Saint Louis also. Toward the middle and end though, every battle I had the Union (Union -10 will strategically) withdrew at around 9 (willpoints) on the tactical map. Sometimes with 2 to 1 odds in there favor.

It puts the AI at a massive disadvantage to have them withdraw after getting down to 9 will to fiight during every battle. For example, I was looking forward to a battle in which the union had 170,000 troops and me 100,000. After much probing, the union lined up man for man with my line and made a direct assault. I beat down the first wave, but my boys were done morale wise. The second wave of men would have broke my line, but the union's will point hit 9 and they retreated.

After the AI gets -5 or so will strategically, all tactical battles are more like engagements, never balls to the wall battles.



< Message edited by dolphinsfan9910 -- 11/13/2007 3:06:07 AM >
Post #: 1
RE: To many early withdraws - 11/13/2007 4:29:16 PM   
ericbabe


Posts: 11927
Joined: 3/23/2005
Status: offline 		We can try calibrating the AI's will to fight retreat ratio based on some other factors, such as total battlefield strength.  Is anybody else experiencing this?

_____________________________



(in reply to dolphinsfan9910)
Post #: 2
RE: To many early withdraws - 11/13/2007 8:35:26 PM   
Ironclad

 

Posts: 1924
Joined: 11/22/2006
Status: offline 		My experience is the attacking Union AI triggering retreat consistently at one third WTF. That may occur when the Union still has superior numbers in the field so I suppose that is an issue although the experiences of the Army of the Potomac show this happening in the actual conflict.

I suppose it would be good to have an element of uncertainty as to when the early retreat is triggered rather than knowing it is always at the one third point. Not that I would want it moving up to the half way point which would be too premature.

< Message edited by Ironclad -- 11/13/2007 9:05:17 PM >

(in reply to ericbabe)
Post #: 3
RE: To many early withdraws - 11/13/2007 8:44:53 PM   
ericbabe


Posts: 11927
Joined: 3/23/2005
Status: offline 		One-third is what it's triggered at.  Is this too soon to retreat based on what you've seen?

_____________________________



(in reply to Ironclad)
Post #: 4
RE: To many early withdraws - 11/13/2007 9:13:43 PM   
Ironclad

 

Posts: 1924
Joined: 11/22/2006
Status: offline 		See my edited post. How about randomly triggering it between one third and one quarter for a larger force (eg 25,000 plus)?

I have just recalled our previous discussion about this. The problem would be solved if you can build in an override so that the standard trigger point is ignored if the attacker still has more unbroken troops on the field than the defender. The voluntary retreat would then only kick in when numbers fell below the defenders.

If its not possible to count unbroken troops could you use the number of unbroken brigades (much less accurate of course).

< Message edited by Ironclad -- 11/13/2007 9:34:17 PM >

(in reply to Ironclad)
Post #: 5
RE: To many early withdraws - 11/13/2007 10:17:17 PM   
ericbabe


Posts: 11927
Joined: 3/23/2005
Status: offline 		That's a good idea.

_____________________________



(in reply to Ironclad)
Post #: 6
RE: To many early withdraws - 11/14/2007 2:23:56 AM   
dolphinsfan9910

 

Posts: 81
Joined: 6/25/2006
Status: offline 		I would like to see some randomness to it.

(in reply to ericbabe)
Post #: 7
RE: To many early withdraws - 11/14/2007 11:34:50 AM   
jkBluesman


Posts: 797
Joined: 2/12/2007
Status: offline 		One random factor could be the event that the army commander is wounded. Beauregard decided to call off the battle at Shiloh when he was suddenly in command and Hooker at Chancellorsville is another case.

_____________________________

"War is the field of chance."
Carl von Clausewitz

(in reply to dolphinsfan9910)
Post #: 8
RE: To many early withdraws - 11/14/2007 4:17:56 PM   
Ironclad

 

Posts: 1924
Joined: 11/22/2006
Status: offline 		Good point. In practice I don't think the override would impact on the CSA since I would expect their forces to already be below the 100% level when the WTF trigger is operated. Certainly I wouldn't want any possibility of the CSA AI being forced to retreat just because it was at less than 100% numerical odds. In fact it might be better to exclude the override from the CSA AI, given the suggested amendments listed below.

Apart from Chickamauga, Wilson's Creek and Pea Ridge I can't recall another battle where the CSA forces possessed numerical superiority. When they did launch major invasions of the north or northern occupied territory they were outnumbered, sometimes heavily so.

Some other thoughts:

An override at the 100% defender level may be close or even worse than the numerical position suffered by the Union army when an automatic retreat happens. I just don't know, never having analysed the unbroken totals at 0 WTF.

It would prevent an early retreat on the scale of Chancellorsville where the remaining unbroken Union forces still heavily outnumbered their opponents. Although that was clearly a leadership issue (Hooker's loss of nerve, affected by his wounding). One could address this by applying different percentage possibilities for different override values. Eg:-

Attacker less than 100% defender - 65% possibility

Attacker less than 110% defender - 20% possibility

Attacker less than 120% defender - 10% possibility

Attacker less than 125% defender - 5% possibility

So in other words the Union AI WTF trigger point for early retreat would most often be overridden at the 100% level; with a union withdrawal then only occurring when unbroken union numbers fell below the total of the CSA unbroken troops.

At the other extreme there would be a 5% chance of the override applying when the Union dropped below 125% unbroken strength of the defending CSA; with the Union AI withdrawing when that numerical position is reached after WTF retreat is triggered (eg remaining Union force totalling 124,999 or less facing 100,000 unbroken Confederates.)

I did wonder about the effect of the override on the improved Union forces of the later war years. However as already noted the Union almost invariably fought with numerical superiority during the war and definitely so when launching invasions into Southern territory so such a restriction on the attacking Union AI doesn't seem unreasonable. Its fair to note however that if the Union does badly in the game, it could restrict their scope to counter-attack with reduced force numbers.

The proposed override still seems a valid option to consider along with the suggested adjustments. In practice this would provide for an element of randomness, although that wouldn't preclude attaching another random effect to the trigger point itself which presumably would apply to both AI's - Union and CSA.

Its a pity we are so close to the gold patch as it would have been useful to have tried these out (if practicable) in beta first.

< Message edited by Ironclad -- 11/14/2007 5:20:17 PM >

(in reply to jkBluesman)
Post #: 9
RE: To many early withdraws - 11/14/2007 5:04:23 PM   
jkBluesman


Posts: 797
Joined: 2/12/2007
Status: offline 		It sounds good in theory but you are right that it is hard to estimate how it would develop in practise.
Concerning the improved Union forces in the later years: they might have improved in the West (espacially the Army of the Cumberland), but the Army of the Potomac only improved in terms of equipment. There were many green recruits in the army in 1864 and the number of deserters was very high.
So I do not think there should be a difference between the early years and the later.

_____________________________

"War is the field of chance."
Carl von Clausewitz

(in reply to Ironclad)
Post #: 10
RE: To many early withdraws - 11/14/2007 5:31:12 PM   
Ironclad

 

Posts: 1924
Joined: 11/22/2006
Status: offline 		Thanks. Much better if it's valid for the whole conflict but it does need testing. By the way the edit, that crossed with your post, simply tidied up my opening paragraph re the CSA.

(in reply to jkBluesman)
Post #: 11
RE: To many early withdraws - 11/14/2007 7:45:00 PM   
ericbabe


Posts: 11927
Joined: 3/23/2005
Status: offline 		The new WTF breaking point for the AI applies only to the attacker (who generally has a much harder time pulling off upset victories once his WTF falls too low), so it often wouldn't apply to the CSA.  I'm adding new criteria that take unit power (strength, firearm, morale, etc.) into account, which should account for the CSA units properly (I hope).

_____________________________



(in reply to Ironclad)
Post #: 12
RE: To many early withdraws - 11/14/2007 8:14:20 PM   
Ironclad

 

Posts: 1924
Joined: 11/22/2006
Status: offline 		Yes, I knew it only applied to the attacking AI. I look forward to seeing the new criteria in action.

(in reply to ericbabe)
Post #: 13
RE: To many early withdraws - 11/14/2007 8:38:07 PM   
Ironclad

 

Posts: 1924
Joined: 11/22/2006
Status: offline
quote:

ORIGINAL: dolphinsfan9910

I definately agree that in the face of definate defeat an army should retreat. I'm glad the days of 3 brigades taking on an army are gone........But where are my get down drag out battles in which I win by a miracle?

One of my strategies for the South is to plunder and get the North's will down. I usually shoot for Saint Louis also. Toward the middle and end though, every battle I had the Union (Union -10 will strategically) withdrew at around 9 (willpoints) on the tactical map. Sometimes with 2 to 1 odds in there favor.

It puts the AI at a massive disadvantage to have them withdraw after getting down to 9 will to fiight during every battle. For example, I was looking forward to a battle in which the union had 170,000 troops and me 100,000. After much probing, the union lined up man for man with my line and made a direct assault. I beat down the first wave, but my boys were done morale wise. The second wave of men would have broke my line, but the union's will point hit 9 and they retreated.

After the AI gets -5 or so will strategically, all tactical battles are more like engagements, never balls to the wall battles.




Dolphinsfan9910:

I forgot to ask whether you are playing the latest public beta - 1.10.9 or the earlier one 1.10.5. Its significant on this matter because the AI early retreat parameters were improved in the later version; being based on one third Will to Fight rather than the earlier level pitched at one half Will to Fight?

(in reply to dolphinsfan9910)
Post #: 14
RE: To many early withdraws - 11/30/2007 6:26:08 AM   
dolphinsfan9910

 

Posts: 81
Joined: 6/25/2006
Status: offline 		When I posted this origionally it was the earlier version.

(in reply to Ironclad)
Post #: 15
Page:   [1]
All Forums >> [Current Games From Matrix.] >> [American Civil War] >> Forge of Freedom: The American Civil War 1861-1865 >> Public Beta Feedback >> To many early withdraws Page: [1]
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.828