Retreat

[MrT]

Ah Mike okay, thats a worthy fix in its own right. Nothing worse than the loser chasing the winner round the field of battle lol.

[Pocus]

Thanks Generalisimo.

MC still seems to be the driving factor for retreat paths. I can have a depot with 68% mc and a region with 90% control and I bet they will retreat to the region with 90% control regardless of where this region is. Maybe just me but this has been the pattern in my tests so far.

Then boost ctlRetreatAdjDepot and you should have your troops attracted by an adjacent depot...

[Gray_Lensman]

deleted

[bigus]

Then boost ctlRetreatAdjDepot and you should have your troops attracted by an adjacent depot...

Yes I have. But what if your not adjacent to a depot, city or fort?

I always thought the ctlRetreatLandLink meant landlinks to offmap areas so I never really paid attention to it. I will try this when I have some time.

[enf91]

Hang on... the "landlink" trait counts the number of regions adjacent to each potential retreat region and assigns a favorability index. Could a parameter be added that causes the engine to check the regions adjacent to each potential retreat region for depots or other structures?

[Gray_Lensman]

deleted

[kwhitehead]

Rebel forces retreating from Memphis will always go to Corinth if Held by Confederate forces. t

No for sure this one isn't true. Recently got into trying to retake Memphis with the AoM. Memphis was the only region held between the Mississippi and Tennessee Rivers. I moved into Memphis from two of the eastern regions. All regions around Memphis including Corinth and all the rail lines were under CSA control. I lost the battle and two Corps ended up on the west side of the river. To make things worse it treated it like a retreat across a river (no rail effect I guess) so an extra 10,000 men were lost.

This is more like a bug than bad parameters. It shouldn't happen and so far it has never happened except in Memphis. My fights in Nashville always retreat southward. But it is hard to be sure since I haven't watched for a duplicate set of circumstances.

[bigus]

Here is a test scenario and Control&Retreat.opt file for the scenario.

If you change any parameters in the Control&Retreat.opt file then you should restart the game.

Note: I've always tested using extreme values. I've left it this way. The original values are noted at the end of the comment.

Edit: updated.

[ATTACH]12558[/ATTACH]

[Ebbingford]

I don't know if it helps at all but have you seen this thread and Pocus' reply?

http://www.ageod-forum.com/showthread.php?t=18354

[Gray_Lensman]

deleted

[Ebbingford]

Pocus answered on 10th August, 1.16beta6 was released on 7th August, so I would guess not.
..............That's if it was present in AACW in the first place.

[Pocus]

Here is a test scenario and Control&Retreat.opt file for the scenario.

If you change any parameters in the Control&Retreat.opt file then you should restart the game.

Note: I've always tested using extreme values. I've left it this way. The original values are noted at the end of the comment.

ctlRetreatLandLink = 35000
ctlRetreatPrevSubSpaceCoeff= 25000

[ATTACH]12497[/ATTACH]

You are in effect removing the influence of all others parameters here...

These changes are dynamic, no need to restart a game.

If there was a bug in the code, it will be fixed for AACW too as the retreat code is common to all AGE games.

[bigus]

You are in effect removing the influence of all others parameters here....

Yes, I was testing. I used extreme numbers that might have offset the other ones. It is not meant as a replacement to the existing OPT file but rather numbers that I found worked. I will try and tweak them down and try and maintain the same result which is not having Johnson retreat across the Mississippi.

These changes are dynamic, no need to restart a game..

Yes. Restart the scenario is what I meant. Edited in the post above.

If there was a bug in the code, it will be fixed for AACW too as the retreat code is common to all AGE games.

Excellent!

[tagwyn]

Pocus: Its your game mes amis. You fix it. t

[bigus]

I've created two new test scenarios (three total) to test the Control&Retreat.opt file.

Test scenario 1 is a one on one with Johnson and Grant.
Test scenario 2 has a depot in Oxford.
Test scenario 3 has the depot and a level 3 city in Marshall (Holy Springs).

Starting with test 1. The problem as I see with this is if you move Johnson into Shelby to attack Grant in Memphis,
Johnson usually retreats to Edmund (2). Ideally you want him to retreat back to where he moved from (1).

Solution: ctlRetreatPrevSubSpaceCoeff= 25000 (or 250)

[ATTACH]12546[/ATTACH]


Test scenario 2. Move Johnson into Austin and wait for Grant to attack. With the vanilla I still had Johnson retreating to Edmund (1)
which is still across the Mississippi.
After a tweak I had Johnson retreating to Helena (2) which is still no good. But the Ideal retreat path here would be toward Desoto (3).

Solution: ctlRetreatLandLink = -10

[ATTACH]12547[/ATTACH]

Test scenario 3. I figure the ideal retreat path here would be toward the depot first (1) instead of the city in Holly Springs (2). Some might disagree.

Solution: ctlRetreatAdjCity = 5

[ATTACH]12548[/ATTACH]

Unless I'm missing something, you do have to restart the Game to see any changes to the OPT file (ALT Tabbing to change and save the OPT file does not work).


Here are the test scenarios and Control&Retreat.opt file.

[ATTACH]12549[/ATTACH]



I still have to test these out a bit more. Overall it seems good.

[Gray_Lensman]

deleted

[bigus]

The "if" leaves a lot of questions and for sure I'm not messing with any of the changes for the Control&Retreat.opt file for the current beta patch without knowing if the aforementioned bug was in the game engine. The v1.16 b6 AACW game engine was released PRIOR to Pocus announcement of finding the bug in the AGE game engine. It could very well be that the next release of the v1.16 game engine might include the fix for his surprising bug find and correct a lot of retreat issues without compensating the variables in the Control&Retreat.opt

No problem. I'm just testing my own set with 1.15 and posting the results.
Since it might be a while before 1.16 is official, compensating is all I have.
I've actually found a problem others might be refering to. This picture shows a union force retreating
on the first couple of days north to its depot (good so far) the next couple of days it is pushed further
north (still o.k) At the end of the turn it is forced to retreat once again but this time it's retreat path
leads it back south!

[ATTACH]12550[/ATTACH]

[Gray_Lensman]

deleted

[bigus]

Check out the thread that Ebbingford links to above in Post #37, especially Pocus reply/response within that thread.

I missed the response by Pocus
This does explain the retreat back to the Confederate depot!

[Pocus]

So yes, this would mean that AACW is fixed on Retreats, as soon as the next beta patch is out

As for 'restart the game', I believed you were saying 'restart the scenario' and not 'restart the game engine'. Sure, the engine don't read dynamically the variables you edit in each of the files (well, it does that, but only when setting up the interface and only for us the devs )

[bigus]

As for 'restart the game', I believed you were saying 'restart the scenario' and not 'restart the game engine'. Sure, the engine don't read dynamically the variables you edit in each of the files (well, it does that, but only when setting up the interface and only for us the devs )

Yes. I have to restart the 'game engine' or 'game' after I change the Control&Retreat.opt file.

[bigus]

The ctlRetreatPrevSubSpaceCoeff value of 25000 works.
IMHO, if you are moving, you should return to the region you moved from.
There should be no doubt about the retreat path.

While static or non moving forces should retreat:
1) depot (controlled)
2) City
3) Land link (lowest MP)

I think this might work with ACW 1.15:

ctlContested = 5 // Minimum control gained upon entering a region (if not passive)
ctlAllowRetreat = 0 // Minimum control to have in a region to allow a retreat into it
ctlRetreatAdjCity = 5 // Interest in retreating toward a region with a city (per level)
ctlRetreatAdjFort = 25 // Interest in retreating toward a region with a fort (per level)
ctlRetreatAdjDepot = 25 // Interest in retreating toward a region with a depot (per level)
ctlRetreatLandLink = -10 // Interest in retreating toward a region, value per land link
ctlRetreatPrevSubSpaceCoeffH = 250 // Coefficient applied to the interest if the region is the one where we are coming from
ctlNoBeachHead = 10 // Minimum control to have so that a region is not a beach/riverhead if you have to cross a body of water before attacking

[Pocus]

No, that's depend of the situation, so the parameters. Say you make a foray into the enemy land, then you find the situation too dangerous, and decide to move back to your soil. But then, you are intercepted by an enemy stack on your way back. With your parameter, you'll bounce back to the enemy soil. With more fine tuned parameters, you'll perhaps make a translation move to a region adjacent which is friendly to you.

[wsatterwhite]

No, that's depend of the situation, so the parameters. Say you make a foray into the enemy land, then you find the situation too dangerous, and decide to move back to your soil. But then, you are intercepted by an enemy stack on your way back. With your parameter, you'll bounce back to the enemy soil. With more fine tuned parameters, you'll perhaps make a translation move to a region adjacent which is friendly to you.

Sounds realistic to me, if a force makes a deep move into enemy territory it runs the risk of the enemy intercepting them and blocking off their retreat path. Historically most Civil War armies a)always took care to guard and protect their lines of retreat as much as possible and b)never really moved too far away from a controlled depot in the face of a strong opposing force. It seems like this should just lead the gameplay to be more realistic, in situations where a player knows strong enemy units might be able to intercept any movement, you simply can't afford to risk any long sweeping movements into enemy territory that might lead to getting trapped there.

If anything, the situation you describe here should be very rare, off the top of my head I can't think of any non-cavalry raid operation where a force moved forward on a path and then decided on its own to move back down that path without being confronted first by enemy resistance. Even considering cavalry raids, most of the successful ones (Grierson's for example) went from one friendly-controlled point to another friendly-controlled point in another area- if at any point Grierson had been intercepted by an enemy force and forced to retreat, he would have just moved back down the path he came because he would have been moving forward all to that point. In short, any movement into enemy territory should be either with the intention of staying there or else just quickly passing through on the way back into friendly-territory somewhere else- there really should be no "voluntarily" moving back unless you are forced back.

[bigus]

I would say "choose your path wisely"

[ATTACH]12644[/ATTACH]

Where would you rather retreat to? location 1 or 2?

I think if you are moving into enemy controlled territory you will plan your move. You want your retreat path to be the region you moved from.
This would apply to cavalry raids as well

After more tests I have found 250 for ctlRetreatPrevSubSpaceCoeffH is not enough. I've gone to 2500.

[bigus]

I thought it was funny no one commented on the -10 variable.

I have this for 1.15....

ctlContested = 5 // Minimum control gained upon entering a region (if not passive)
ctlAllowRetreat = 0 // Minimum control to have in a region to allow a retreat into it
ctlRetreatAdjCity = 5 // Interest in retreating toward a region with a city (per level)
ctlRetreatAdjFort = 25 // Interest in retreating toward a region with a fort (per level)
ctlRetreatAdjDepot = 25 // Interest in retreating toward a region with a depot (per level)
ctlRetreatLandLink = -10 // Interest in retreating toward a region, value per land link
ctlRetreatPrevSubSpaceCoeffH = 500 // Coefficient applied to the interest if the region is the one where we are coming from
ctlNoBeachHead = 10 // Minimum control to have so that a region is not a beach/riverhead if you have to cross a body of water before attacking

[ATTACH]13183[/ATTACH]