Matrix Games Forums

SVF 2.0 public beta test 20121004. Late November 1861. Union AI, Confederate Human.

Battle of Manassas Again!



Nice victory.

But I'm a bit concerned about the pace of warring this early in the war. In the Real War, in 1861, there was only a single major engagement in the east -- First Manassas. In my test game, there have been multiple Battles of Manassas, and more than a half dozen major engagements, many of them initiated by the Union AI. In the Real War, after its loss at Manassas, a chastened Union retired to Washington DC to lick its wounds and regroup. It didn't venture forward again until spring of the following year. In this game (and generally in the SVF mod?), is the Union AI perhaps too motivated to push, push, push forward? Do you have mechanisms in place to recreate the Real War, post First Manassas defeat, Union meekness and hesitancy?

The Confederacy has finally chosen the Paper Money Printing option, BTW.

Union Morale 96, VPs 672, Combat Losses 19735

Confederate Morale 122, VPs 780, Combat Losses 16499

Unlike my first test game, where I questioned the low loss levels, in this second test game the losses seem maybe a bit too high.

Thoughts?

Saves, Logs, Scripts available here.

SVF 2.0 public beta test 20121004. Early December 1861. Union AI, Confederate Human.

First Union AI Coastal Ops, Saint John's FL:



Good to see some coastal ops at last. (Although I think you said they are in the stock AI, not new scripts you've prepared for SVF.)

Fortunately, in anticipation of this sort of thing, I had thought to garrison most of the Southern coastal forts with home-state militia. So I'm not caught completely flat footed and unprepared.

Union Morale 96, VPs 746, Combat Losses 19955

Confederate Morale 122, VPs 831, Combat Losses 16849

Saves, Logs, Scripts available here.

Manassas value must be ressetted after Summer : next version.

Florida: the good news is the operation is compliant to my own amphibious events: Florida was picked because the event fired.

Thanks for the work. [:)]

SVF 2.0 public beta test 20121004. Late January 1862. Union AI, Confederate Human.

Mid Winter Battles.

In Early January, Bull Nelson attacked a commanderless garrison at Columbus (Hickman KY). Columbus fell to the Union.

To the west, Franz Sigel conducted a mid winter assault, in the snow, against Earl Van Dorn's command at Rolla MO. Sigel was repulsed, and retreated northward.

To the east, no action for a couple months or more.

In Florida, Stiles' Brigade abandoned the siege of St. Augstine and turned to invest Ft. Marion. I am assembling a counterforce in the vicinity.

The situation in the center:



Union Morale 93, VPs 884, Combat Losses 21275

Confederate Morale 121, VPs 975, Combat Losses 18460

Saves, Logs, Scripts available here.

The Ft marion preference is intended: Union AI will search to capture fort before the other regions for supply purpose...

This is really good stuff [&o]. I will start SVF 2.0 when my real life starts to ease up.

SVF 2.0 public beta test 20121004. Early February 1862. Union AI, Confederate Human.

What in heck is Sigel Doing? Marching westward through the winter mud away from his sources of supply? (The previous turn, he was east of Jefferson City, at Union MO. IIRC.)



At the siege of Ft. Marion FL, is Stiles' Brigade withering away from lack of supply?

Late last month, the Union Army of the Potomac sallied forth from Washington, Mud Marching south towards Fredericksburg. Seeing as "Yup, it's pretty muddy," and observing the well-entrenched Jackson Command at Fredericksburg, the Union AOP returned whence it came -- Alexandria VA. They needed the exercise maybe? [;)]

Union Morale 93, VPs 930, Combat Losses 21275

Confederate Morale 121, VPs 1023, Combat Losses 18460

Saves, Logs, Scripts available here.

Sigel: Lexington is a supply source and the move is limited to Versailles region. Now, why Versailles has been picked? I don't know. As I don't know what next Sigel move will be. Not a clever AI action, but nothing disasterous.

McDowell: here the log

4:29:39 PM (Reporting) Agenda 1 : Agenda: Attack, Value: 9975, Region: 133 Nelson, VA
4:29:39 PM (Reporting) Agenda 2 : Agenda: Keep Position, Value: 74550, Region: 119 Stafford, VA
4:29:39 PM (Reporting) Agenda 3 : Agenda: Defend, Value: 26376, Region: 116 Fairfax, VA
4:29:39 PM (Reporting) Agenda 4 : Agenda: Attack, Value: 3450, Region: 147 Richmond, VA

The hightest value, and by a large margin, is Agenda 2, keep position at Stafford. The picked agenda was the third, defend Fairfax. That's the result of the hard randomization routine of the AGE AI: it happens from time to time. Let's imagine it's the game repritition of the 1863 Meade Mud march...The key is to compute if this AI behavior comes from time to time, that is acceptable, or often, and so has too be reduced.

Ft Marion: yes the uNion unit is starving due to lack of supply. The AI hasn't the tools for evacuation or to attach a supply wagn to an amphibious force. Now, this one operation as tiny, involving only one unit. The Union AI is preparing larger operations. I'm going to see if it's possible to get some supply wagons too, but it will be difficult.

Are you using the latest patch (1.16a) for this Mod?
If not, which patch is this based on?

It works with the latest version [:)]

SVF 2.0 public beta test 20121008. Early April 1862. Union AI, Confederate Human.

McClellan Smacks Stuart.

JEB Stuart's newly formed cavalry division was on a scouting expedition in the Shenandoah. Seeing as how Winchester was ungarrisoned, Stuart recklessly decided to take the place. In a neat countermove, McClellan's (!) Corps moved swiftly into Winchester VA and smacked Stuart hard. Stuart withdrew to the lower Shenandoah to lick his wounds.



In the center, no action since last autumn, force buildup only.



In the west, Fremont is looking to pick a fight.



Union Morale 97, VPs 1108, Combat Losses 21525

Confederate Morale 121, VPs 1215, Combat Losses 20630

Saves, Logs, Scripts available here.

I'm currently working to renforce Union AI interest in KY/TN area. It was a weak spot for it, with too few units devoted to these States. My very preliminary test toninght on the enxt version seems to show I've solved it, as units in WV are coming back to Kentucky. Next release soon.

In my previous report, I forgot to mention:

[*] In early March, Grant made a move against Island 10, but backed off when Bragg showed up in strength.
[*] In Florida, Stiles broke off the siege of Ft. Marion and ... has vanished! Died from lack of supply?

yes.

SVF 2.0 public beta test 20121008. Early May 1862. Union AI, Confederate Human.

The Universe of Battle.

In an early spring (early April) offensive, Banks' Union Army of the Potomac marched southward towards Fredericksburg. Beauregard's Army of the Potomac [I really must rename that!], with Jackson's Corps, left its Manassas entrenchments to move on Bank's right flank at Stafford. Johnston's Corps marched to the sound of the guns up from Fredericksburg. An epic battle, and Glorious Southern victory!, ensued:



To the southwest, a very aggressive Heintzelman handled Stuart rather roughly:



In early May, the situation in the east:



Meanwhile, in Missouri, Van Dorn repulsed Fremont at Rolla:



All quiet on the KY/TN front, and along the coasts.

Union Morale 97, VPs 1215, Combat Losses 33335

Confederate Morale 125, VPs 1286, Combat Losses 31675

Saves, Logs, Scripts available here.

[I wish I could crank out turns faster, and move this playtest along more swiftly; but Real Life intrudes!]

An observation:

All battles thus far have been one-day affairs. Okay for 1861, and smaller engagements. But it would have been better for the epic Battle of Stafford to last for two days (or even more). Why? It's because most Real War large-scale marquee battles were multi-day, typically two. Indeed, consider the Real War Seven Days Battle (much like, in numbers, the in-game Stafford Battle) -- obviously lasting more than a day.

Getting the battle durations right -- this is something I worked very hard at while I was an AACW Beta Team member several years back.

I know that this is probably determined by the game engine, not by the mod. But it would be a shame (to me!) if all battles were to be one-day, hence the game were to fall short of realism on this point.

ORIGINAL: berto


An observation:

All battles thus far have been one-day affairs. Okay for 1861, and smaller engagements. But it would have been better for the epic Battle of Stafford to last for two days (or even more). Why? It's because most Real War large-scale marquee battles were multi-day, typically two. Indeed, consider the Real War Seven Days Battle (much like, in numbers, the in-game Stafford Battle) -- obviously lasting more than a day.

Getting the battle durations right -- this is something I worked very hard at while I was an AACW Beta Team member several years back.

I know that this is probably determined by the game engine, not by the mod. But it would be a shame (to me!) if all battles were to be one-day, hence the game were to fall short of realism on this point.

You're right. That's indeed apoint to be monitored. Once I will have refined AI behaviour in winter ( next version) I will myself go on 1862 turns [:D] to check what's needed.

In the course of my SVF playtest, I frequently post images, but also .zip files with Saves, Logs & Scripts.

The first time or two I did that, it was tedious, boring, and -- I could see -- eventually prone to error. There has to be a better way...

There is! Here is a Bash script, which I run in Cygwin for Windows, to prepare the .zip files, in one easy command line:

---

Robert@roberto ~
$ cat /home/Robert/bin/mkmodrpt

#!/usr/bin/bash

ORGDIR=`pwd`

if [ $1 = "-z" ]; then
ZIPIT=1
shift
fi

SAVNAM=$1
RPTNAM=$2

SRCDIR="/cygdrive/c/Games/AGEOD/AGEod's American Civil War SVF/ACW"
TMPDIR=/cygdrive/c/Temp

SAVGAM="SVF20beta20120918S"

TGTDIR=$TMPDIR/$RPTNAM

mkdir $TGTDIR 2>/dev/null

rsync -vap --delete "$SRCDIR/Logs" $TGTDIR
rsync -vap --delete "$SRCDIR/Scripts" $TGTDIR
mkdir $TGTDIR/Saves
rsync -vap --delete "$SRCDIR/Saves/$SAVNAM" $TGTDIR/Saves

cd $TGTDIR/Scripts
for f in `ls -tr | head -n -3`; do rm -f $f; done
ls -l

cd "$TGTDIR/Saves/$SAVNAM"
rm -rf Backup[2-9] Backup??
ls -l

if [ $ZIPIT ]; then
cd $TMPDIR
zip -r $RPTNAM.zip $RPTNAM
fi

cd $ORGDIR

exit 0

---

mkmodrpt

[*]mkdir's a temporary staging directory
[*]copies (by means of rsync) the game Logs, Scripts & Saves folders to that staging directory
[*]rm's (removes) all but the most recent 3 script files
[*]rm's all but the last two backup sets
[*]optionally creates a .zip file of the bundle (optional, because I can manually use WinZip etc. instead if for some reason I prefer)

I could fancy this up, by adding command-line options, and suitable code, for different AGEOD games (and game installations), different game playtests, etc. But I'm KISSing this for now.

Here is a sample mkmodrpt invocation:

---

Robert@roberto ~
$ /home/Robert/bin/mkmodrpt -z SVF20betaS2 SVF20Beta20121008TheUniverseOfBattle

---

where 'SVF20betaS2' is the name of my test game, and 'SVF20Beta20121008TheUniverseOfBattle' is the name of my .zip file.

(I have a suite of Cygwin Bash scripts for automating weather mod creation, maintenance & QA, for example:

---

Robert@roberto ~
$ find bin/weathers -type f | egrep -v "~"

bin/weathers/cksum_mkweathers
bin/weathers/cp_mkweathers
bin/weathers/diff_dist
bin/weathers/diff_Weathers
bin/weathers/make_backup
bin/weathers/make_dist
bin/weathers/rsync_Saves

---

For AGElint development, maintenance & operation, there are many, many more Linux scripts besides.)

Without Linux & Cygwin, without programs and scripts, without automation -- without these things, I don't know that I would playtest, mod (weathers and ...), develop AGElint. Too much trouble. In some respects, next to impossible.

Linux, Cygwin, etc. -- Try them, you'll like them!

SVF 2.0 public beta test 20121009. Late May 1862. Union AI, Confederate Human.

A.S. Johnston Retakes Bowling Green.

With a beefed up Army of Tennessee, A.S. Johnston (incl. Bragg's Corps) advanced on Bowling Green (Warren KY). Crittenden's Lawrence Force withdrew without a fight. I like the looks of this.



In the east, Heintzelman's Corps runs rampant through central Virginia. I don't like the looks of this.



In the west, Fremont beat up a vastly outnumbered Price and seized Springfield MO. A major Union AI offensive all across Missouri/Arkansas. I don't like the looks of this.



I have to reconcile myself to: Vastly outnumbered in men and resources, the South cannot hope to hold everywhere. Something's got to give. It's just questions of where, when, and how quickly.

Union Morale 99, VPs 1221, Combat Losses 34135

Confederate Morale 127, VPs 1334, Combat Losses 34815

Saves, Logs, Scripts available here.

[I continue to post so frequently, not only because the circumstances are interesting, but also to provide plenty of data points for developing the 1862 Union AI.]