SVF 2.0

[berto]

Oh, and I now see Burnside's Command, on Union ships, lying in wait on the James River. Sneaky, sneaky...

[Chilperic]

My preference goes for an AI behaving historically with possibilities of divergences to surprise the opponent player. Now, the alternatives must remain as close as possible to real outcomes, like a 1864 campaign in 1861 or to plans that were never enforced ( let's say the Grant proposal in 1864 to invade by sea North Carolina with the Potomac Army...).

If we look the other side of the hill, do we expect a player to repeat the same strategy than the historical one? [;)]

[Chilperic]

I should add my method allows generally to get both a plausible AI in simulation sense,and a worthy opponent in gameplay terms. FY AI is yet a bit better on this, because of some differences in game mechanism and maybe in the hard AI itsef, even if I'm not sure of that. Or it's just because the map is smaller especially for the Eastern theater, so the AI is more prone to follow an unexpected path...

[berto]

SVF 2.0 public beta test 20121019. Late October 1861. Union AI, Confederate Human.

Is McClellan Doomed?

Leaving Jackson in control of Alexandria -- after the twin attacks on that place, his forces were too low in cohesion to move -- I have rushed Beauregard and Johnston south to Louisa VA, just northwest of Richmond. Lee recovers to the south of Richmond. McClellan is now surrounded:



The opposing forces in the general vicinity of Richmond:

[*]McClellan's AoP, now reinforced by Burnsides' Command: ~1500 Pwr

Versus:

[*]Beauregard's AoP: ~750 Pwr
[*]Johnston's Command (in the same region with Beauregard): ~350 Pwr
[*]Lee's Virginia District Command: ~450 Pwr

Southern forces are supplied, Northern forces are cut off from LOC. (Autumn rain and mud about to hit.) The Union forces at Richmond should decline in effectiveness over time. The south can draw on additional reinforcements, but can McClellan?

Is McClellan doomed?

If I can destroy McClellan's AoP at Richmond, retake the Southern capital, and menace Washington DC in force next spring, can I yet turn the tide?

Union Morale 138, VPs 751, Combat Losses 14376

Confederate Morale 60, VPs 643, Combat Losses 12957

Saves, Logs, Scripts available here.

[PanzerHue]

hi,
its great to follow the development of the SVF2 and being a long time fan of this mod I cant wait to try it out.
CHLIPERIC: wich files should I send you? and where should I send them?
best regards,

[Chilperic]

Hi Panzerhue,

sorry for the delay but i've been fixing a nasty bug in Fatal years today.

If you can't upload here the file, you may alternatively mail them at:

fatalyearssup@yahoo.fr

or upload them on the AGEOD forum ( thre's a SVF 2.0 thread buried in the mod subforum). For many reasons, I don't log anymore there, but As I'm forced to lurk the AGEOD website for info, I will download the files.

Feel free to post your comments on the Matrix forum. I read them and I reply them. That's not to say I'm always agreeing, but I will always explain why I've chosen to do this way in my mod. From time to time, I may agree too.

Thanks in any case for the input. A good game isn't only design, but playtesting in the real sense of the term, ie at the basis the reports of several players who have each one their way to play. The AI must be able to cope with different human strategies and the historical events,data must be cross checked by a group to be as close as possible of the reality. [:)]

[berto]

On the Politics screen, I have the option to Move Capitol. It only mentions costs, not benefits. Why would I now (under the current circumstances, the loss of Richmond) or ever want to move my capitol?

[berto]

SVF 2.0 public beta test 20121019. Early November 1861. Union AI, Confederate Human.

Retaking Richmond.

Referring to the previous screenshot above, I can see that the Union AoP at Richmond is too strong to attack. No sense trading equal casualties, 50/50, or getting the worst of it in assaults. Two of my three ranking generals in the vicinity of Richmond -- Beauregard, J. Johnston -- are stronger on the defense than on the attack. Generally speaking, I want to entice the Union to attack me, not vice versa.

And I want to ensure that the Union AoP's lines of communications remain cut. But I imagine that it seized huge stores of supply on capturing Richmond.

My strategy will be to avoid battle, stay on the defense, and hope that the AoP's supplies run out. Then I will attack a weakened AoP at the opportune moment.

I need to block the Union force at Bedford VA from reinforcing McClellan's force, if at all possible. So I send Beauregard to Buckingham VA. A. Johnston and Lee will remain in place at Louisa VA and Henrico VA, respectively. This will create a three-force Confederate arc to the west of Richmond. To the northeast, I tighten the noose. I leave Jackson at Alexandria. The Union has sizeable forces at Washington DC. I could send Jackson southward to help save the capitol, at the risk of losing all of northern Virginia north of the Rappahannock or worse. Another risky gamble!

I hit the Next Turn button ... and here is the result:



The block failed. The Bedford VA force has reinforced McClellan. [:(]

McDowell's Command has moved on Lee south of Richmond at Henrico VA. Lee declined battle and retreated to Petersburg (Dinwiddie VA).

The bulk of the Union forces are now south of Richmond. A reduced McClellan AoP remains at the Southern capitol. Is now the time for Beauregard and J. Johnston to strike? Darn, but I wish my best attacking generals were northwest of the capitol and not farther south at Petersburg (Lee) and far north at Alexandria (Jackson).

In KY/TN and out west, nothing much is happening.

All eyes are on Richmond. Tense, tense, tense.

Union Morale 138, VPs 844, Combat Losses 14376

Confederate Morale 62, VPs 688, Combat Losses 12957

Saves, Logs, Scripts available here.

[berto]

It occurs to me that, from a playtest perspective, it is a good thing I lost Richmond so early in the war. Why? Because maybe sooner rather than later your Union AI faces situations -- taking Richmond, possibly also taking all of northern Virginia -- that it might more usually face in late 1862, 1863, or beyond. So, even though it is still 1861, we get to test your Union "mid-war AI", sort of. [8D]

[Chilperic]

The Union AI has almost the same plans for the whole war. The first months of 1861 have special details to adapt these plans to the starting situation, the strategic directions are the same.

What produce variety in AI play is the choice between several plans by theater and the AI adaptation to the current display of forces. So any game should show both same AI concerns for primary objectives and surprises when and how it will try to capture them. That's about the same system than in FY, the major difference being a much smaller map in SVF, limiting the variations.

You could have another game where Union AI could stay on the defensive in Virginia during the whole year before starting offensive operations in 1862. That's a possible outcome. Mre generally, offensives ops will be the norm in Virginia in 1861 with variations for the axis.

Now of course, in 1862, Union Army will be larger and so Union AI yet more aggressive.

I think Union AI has chosen to live from supplies captured at Richmond during winter, as movement is very slow during this period. Now, of course, in RL, the uNion navy would have pushed to free the James River to supply Richmond. That's something I could try to add to the AI. Next version.

[berto]

SVF 2.0 public beta test 20121019. Late November 1861. Union AI, Confederate Human.

Union Takes Petersburg.

I attempted a risky commander switch, with Beauregard moving south to take command of the Petersburg defenses, while Lee moved north to assume command of the Confederate AoP. Very risky, because it temporarily left two forces under the command of much less capable subordinate commanders, Huger and E.K. Smith (at Petersburg).

Banks took advantage, another gamble lost:



The Union has taken Petersburg!



Grimmer by the month. [:(]

I must say that, setting aside the wisdom/plausibility of the initial Union move against Richmond, after it took Richmond, the Union AI is acting very capably.

If it has not made a fatal mistake, and faces eventual starvation due to lack of supplies. A very big if.

In the center, Grant has assembled a force at Camp Robinson (McCracken KY). Out west, Fremonts faces off against Price at Springfield MO.

The Union has built up a huge, probably insurmountable lead in NM. This could be a very short-lived War of Southern Rebellion. [8|]

Anyway, it's entertaining. [8D]

Union Morale 145, VPs 929, Combat Losses 16581

Confederate Morale 62, VPs 734, Combat Losses 16027

Saves, Logs, Scripts available here.

[berto]

ORIGINAL: Chliperic

I think Union AI has chosen to live from supplies captured at Richmond during winter, as movement is very slow during this period. Now, of course, in RL, the uNion navy would have pushed to free the James River to supply Richmond. That's something I could try to add to the AI. Next version.

Anticipating that (assuming a plausible Union behavior, even if the current AI is not programmed for it), I am rushing as many Southern fleets to the James River as I can.

[Chilperic]

the next version of the offcial AACW will maybe add some of the AJE tols I would need to help the AI to be more plausible in its Forward to Richmond plan, by protecting its LOS or open a naval one.

Anyway, it changes from the usual thread: what house rules to play against the AI?[:D]

[berto]

SVF 2.0 public beta test 20121019. Early December 1861. Union AI, Confederate Human.

Attack on Richmond Imminent

Beauregard has successfully joined up with and taken charge of the sizeable Confederate force southwest of Richmond at Camden VA. And Lee now commands the newly rechristened Army of Northern Virginia, sited west of the capital, ~1600 Pwr strong.

Fortunately, the Union army remains divided, with the bulk of their forces, under Banks' command, ~1000 Pwr strong, resting at Petersburg. McClellan's Army of the Potomac HQ command remains at Richmond, ~600 Pwr strong.

Lee's command is out in the field, a bit distant from immediate sources of supply. The Union supply situation is unknown. Vis-a-vis Lee, McClellan's command may be at its weakest. Now may be the best time for Lee to strike. I propose for Lee to Sustained Attack (All Out Attack is too risky) McClellan the immediate next turn. An attack on Richmond is imminent.



To the north, Jackson remains at Alexandria, keeping watch over northern Virginia, and keeping a close eye on the equally strong Union forces defending the Union capital.

Out west, Price gave up Springfield MO to Fremont, retreating to the southwest without a fight. In KY/TN, force buildups continue.

Union Morale 145, VPs 1009, Combat Losses 16581

Confederate Morale 62, VPs 778, Combat Losses 16027

Saves, Logs, Scripts available here.

[berto]

SVF 2.0 public beta test 20121019. Late December 1861. Union AI, Confederate Human.

Richmond Retaken!

In order to distract the Union high command and tie down Banks at Petersburg, Beauregard moved on that place, with orders to Retreat If Engaged. The Union had divided its forces still again, sending Banks to Prince George VA to capture the Southern Columbiads there, then onward across the James to Charles City VA. Despite having the numerical advantage, Beauregard lost the Battle of Petersburg:



Of little concern, because Beauregard had served his purpose admirably.

Lee attacked McClellan and won the Battle of Richmond!



A most wondrous and splendid Christmas present!

McClellan was forced to retreat southward across the James River towards Petersburg. The Union AoP is now spread out in various regions south of the Confederate capital:



Lee's command is much weakened (Pwr down to ~1000), what with reduced cohesion after the march to Richmond through the late autumn mud, then the Richmond battle. Lee must now go into defensive mode, all or in part even a Hold at All Cost defense, because we can't afford to lose Richmond yet again.

We are now concerned that the Union AI might, but likely not, retreat towards the sea, move down the James, to try to take Norfolk VA.

The Union AoP might now have good reason to worry. Where will it get its supplies from? From the Union Navy? Note the massive Union naval presence in the James River, as demonstrated by the Battle of James Estuary:



There is still some hope we can contain McClellan's Army, and perhaps even destroy it. General Winter, Lieutenant General Starvation, lead the way!

Out west, Price has joined Hindman to defend Fayetteville AR. In KY/TN, Buell's Army of the Cumberland has started to besiege Ruggles at Columbus (Hickman KY).

Note the dramatic recovery of Southern NM, and the significant drop in Northern NM. We may yet win this struggle!

Union Morale 133, VPs 1046, Combat Losses 18694

Confederate Morale 112, VPs 874, Combat Losses 19887

Saves, Logs, Scripts available here.

[berto]

SVF 2.0 public beta test 20121019. Late December 1861. Union AI, Confederate Human.

Localization bug in message box: "3/4. evt_nam_CSA_Missouri_Partisan"

[Chilperic]

In your saves, there are a few lessons to learn to improve the UnioN AI. For the next version [:D]

[berto]

SVF 2.0 public beta test 20121019. Early January 1862. Union AI, Confederate Human.

Winter Revels.

The Army of Northern Virginia is digging in at Richmond. When it's not celebrating the Joys of the Season. R.E. Lee is the toast of Richmond. Hope abounds!

A supply-short Beauregard has withdrawn to Lynchburg (Florence VA).

Banks has moved down the James Peninsula to take Williamsburg VA. McClellan sulks south of Richmond. Again, how is the Union army being supplied?

Note that I have raised another One-Time Tax, and ordered an out-sized Call for Volunteers. Inflation increases thereby, but we need to bulk up our forces for the hard fighting ahead.



Buell has broken off his Columbus KY siege and withdrawn the Army of the Cumberland northward. Out west, Fremont ponders whether and when to take on Price (at Fayetteville AR).

Union Morale 132, VPs 1092, Combat Losses 18694

Confederate Morale 112, VPs 921, Combat Losses 19887

Saves, Logs, Scripts available here.

I must say that, so far, the Union AI is putting up a good fight, and giving me an enjoyable and fairly plausible game. Considering the cautious Northern commanders (McDowell, then McClellan); given the Union's (and Lincoln's and McClellan's) extraordinary early-War concerns for the safety of Washington DC; and given the breathtaking Sherman-like audacity of cutting cleanly away from sources of supply (something that McClellan would never ever have done) -- again, it is not very plausible what the Union AI did, lunging towards Richmond like that. But, the deed being done, the Union AI is since then acting credibly in its subsequent Richmond Campaign. I don't observe obvious Union AI stupidity (but I'm not looking too hard for it, either).

Not taking AACW quite seriously before, for various and sundry reasons expressed in hundreds of forum posts over the past few years, I am at long last enjoying the game.

The Union AI -- well done so far. (Although I'm still waiting ... and waiting ... for the promised KY/TN and coastal action. [;)])

[Chilperic]

We'll see. The new offical version has been released tonight and could help for coastal ops. Continue your game as I'm not going to release a new one before having studied a bit the new offcial beast.

BTW, it's each day clearer next AGEOD game should be AACW 2.

Oh...I forgot: the Peninsula campaign was done by letting only McDowell corps before Washington. As much I agree with you about the overcautiousness of both Union military and political heads, it remains Union undertook in 1862 a veru boldly operation McClellan led certainly with a total lack of aggressiveness; Lincoln then named Pope who was about the opposite, and if Lee wasn't himself in command, taking the offensive, Pope would vcertainly have tried to advance to Richmond. I don't think another startegy for Union was impossible, even if indeed the chances to choose a less cautious one were slim in 1861.

[Chilperic]

The AI has lived from the supply stock captured at Richmond. Nothing supernatural here!