Island No10 confederate trap?

[DDLAfan]

Confederate army retreats from Columbus, Ky, but instead of going south to Humboldt they retreat into the swamps of Island No10. Thing is they can't reach it in 15 days, so they are stuck in Columbus indefinitely as a punching dummy with diminishing cohesion getting attacked each turn until they are destroyed. I lost a 50,000 man army in ACW1 and the same thing is going to happen here.

Is there any chance we can get the retreat parameters modified so that either a)retreating armies leave the combat space regardless of distance, or b)if they retreating army can't reach the adjacent space in 15 days that in then looks for an alternative retreat route.

Otherwise my next game I'll just blow up the No10 redoubt so it doesn't become a death magnet for retreating troops

[Merlin]

One of the defining points of a PBEM player is that they use other units to shape their advances and retreats. (SP players can use this too, provided you catch your opponent seriously off guard.

I see your partisans; that's all the absolute bare minimum is all I ask.

[Captain_Orso]

Confederate army retreats from Columbus, Ky, but instead of going south to Humboldt they retreat into the swamps of Island No10. Thing is they can't reach it in 15 days, so they are stuck in Columbus indefinitely as a punching dummy with diminishing cohesion getting attacked each turn until they are destroyed. I lost a 50,000 man army in ACW1 and the same thing is going to happen here.

Is there any chance we can get the retreat parameters modified so that either a)retreating armies leave the combat space regardless of distance, or b)if they retreating army can't reach the adjacent space in 15 days that in then looks for an alternative retreat route.

Otherwise my next game I'll just blow up the No10 redoubt so it doesn't become a death magnet for retreating troops

The retreat code looks at where the closest forts, depots and cities and their sizes are in this order of importance. Chance then decides in which direction your stack will actually retreat.

I believe, but I'm not sure, that the number of days to the above list of locations is weighted into the equation.

If on the turn after the battle your stack has not yet entered the target region, technically it is still in the original region and you can simply cancel that stacks move and either stay put and defend or order a move to another region.

One of the defining points of a PBEM player is that they use other units to shape their advances and retreats. (SP players can use this too, provided you catch your opponent seriously off guard.

I see your partisans; that's all the absolute bare minimum is all I ask.

My hovercraft if full of eels

[Merlin]

My hovercraft if full of eels

That's what I get for posting while tired and multi-tasking. I can't remember what I meant to say with that last line

[loki100]

The retreat code looks at where the closest forts, depots and cities and their sizes are in this order of importance. Chance then decides in which direction your stack will actually retreat.

I believe, but I'm not sure, that the number of days to the above list of locations is weighted into the equation.

the retreat code is in 'gamelogic in the 'settings' directory. In full they are:

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 = 30 // 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


// Added September 22, 2012
cltRetPenaltyPerNmySU = 4 // retreat penalty (in interest pts) for each nmy SU in retreating region
cbtRetreatMax = 100 // Max retreat chance

Key bits are there is no time factor (I made this mistake and had to check in a PoN game - I think it has changed from the earlier AGE games as a lot of people believe it does), but that depots are very attractive as a target province.

Won't address the Island 10 issue, but if you want to stop retreats into enemy held territory adjust ctlAllowRetreat= 0 to a higher value (ie you can only retreat to a province you have a minimum MC in. But ... be careful, this will also lead to outright surrenders if a stack can't retreat.

If you get trapped in this sort of situation, I'd split my army up, put the fast stuff on passive, find something half decent (ie has some cohesion) and give it defend at all costs, and hope you can slip out while your rearguard saves the day.

[Captain_Orso]

We had huge discussions about retreating in beta. There are a number of issues that cannot be coded to meet all situations such as the game has no concept of which direction stacks are facing when in battle, which should dictate in which direction they would retreat.

The closest thing which would work in most situations is the "ctlRetreatPrevSubSpaceCoeffH = 250 // Coefficient applied to the interest if the region is the one where we are coming from" parameter.

Of course if you are returning from enemy territory and could at least not distance your stack from friendly territory by retreating to the side it would be better to do that.

If you passed a depot while moving into the battle region and then had to retreat, it is unlikely that you would retreat toward the depot, because "ctlRetreatPrevSubSpaceCoeffH" will nearly always retreat your stack from where it just came.

I don't think a stack will simply surrender outright if it cannot retreat. Elements which are in bad shape (hits/cohesion) can surrender through combat, but otherwise surrender en mass occurs only if your stack in inside a location and besieged.

[Captain_Orso]

that's what i get for posting while tired and multi-tasking. I can't remember what i meant to say with that last line :d

lol

[ArmChairGeneral]

Merlin, if I understand your post, you are suggesting positioning some troops in a nearby city/fort other than Isl. 10 so that your force will retreat to that region instead of Isl. 10? I have not seen in the code or in-game that friendly units draw retreaters, only that enemy units repel them....

That being said, I sometimes manage to use the repelling effect to herd enemy retreaters to regions I want them in by occupuying the region I do not want them to go to with a second stack, reducing their interest in that region. (Never works when they are on the march, they are almost always going to go back the way they came no matter what you do, but sometimes you can lay a trap for them....)

Isl. 10 is very tricky to manage overall. You often can't retreat from it in one turn, and credible invasion forces will have enough pwr to take the redoubt in an assault so you don't want to have a lot of stuff inside it either. Relief from Memphis or elsewhere takes a long time to arrive (swamp), so holding out for friendlies to lift a siege is not a good idea either. As the OP noted, Isl. 10 is also dangerous to troops in surrounding regions because the redoubt will draw retreaters from some regions.

[Jagger2013]

We had huge discussions about retreating in beta. There are a number of issues that cannot be coded to meet all situations such as the game has no concept of which direction stacks are facing when in battle, which should dictate in which direction they would retreat.

I just wrote up a suggestion last night in one of the developer forums after losing 2 armies in RUS and possibly a third shortly due to weird retreats.

What happens during a retreat? Lose a battle, game engine consults algorithm, game engine plots retreat move.

What if: Lose a battle, game engine checks for a player designated retreat city, game engine plots and executes move to designated retreat city. If there is no player designated retreat city, use normal retreat algorithm to plot and execute move.

So give the player the option to designate a retreat city during the plotting portion of the move. If battle is lost, game engine checks to see if a retreat city is designated. If yes, ignore retreat algorithym, plot move to retreat city. If no, use retreat algorithm.

Note, it doesn't matter where or when the battle takes place. Lose a battle anywhere or anytime, game engine plots quickest move to retreat city and executes. It is the same process as now but instead of using the algorithm to determine where to plot and execute the move, the designated retreat city is used to determine where to plot and execute the move.

The algorithm gives good results most of the time but not all the time. And people have been complaining about the occasional weird and disasterous retreat since the original BOA years ago.

Keeping fingers crossed that it is doable.

[Merlin]

Merlin, if I understand your post, you are suggesting positioning some troops in a nearby city/fort other than Isl. 10 so that your force will retreat to that region instead of Isl. 10? I have not seen in the code or in-game that friendly units draw retreaters, only that enemy units repel them....

No, I kind of half fell asleep when I wrote that.

I was suggesting that players should "shape" the battlefield. For example: Refusing to decisively engage near Is. 10 and instead attempting to force a fight further east and/or south.

We all know the retreat mechanics can be wonky, we know what happens after a ten day march into enemy 100% MC, and we know not to assault extremely well defended positions. I rarely put my head down and try to push a region unless I know I can get out if I lose and the opportunity is relatively consequence free. Instead, I try to stack as many factors as possible in my favor so if I don't prevail I can at least control the aftermath.

That said, if the retreat mechanics could be adjusted to prefer least enemy MC, shortest retreat time, and proximity of friendlies (in that order), I think a lot of the wonkiness would go away.

[loki100]

one change I'd like is for current direction of march to be taken into account. In some cases, I've ordered units to move away from the enemy, they get caught and then retreat as per the rules (taking no account of a direction of march they may well have started some days before). In effect, make a predesignated direction of movement as attractive as 'province moved from' in the algorithm?

Having said that, I'm accepting of wierd outcomes sometimes. There are plenty of examples of simple mistakes in most wars, and for a pre-modern army, managing a retreat from battle was one of the most challenging phases to handle effectively

[Jagger2013]

Having said that, I'm accepting of wierd outcomes sometimes. There are plenty of examples of simple mistakes in most wars, and for a pre-modern army, managing a retreat from battle was one of the most challenging phases to handle effectively

I have done the same thing for a long time. Weird things happen and just accept it. But it is just a rationalization. The reality is an algorithym is choosing your retreat path. Just a mathematical formula. 95% the retreat makes sense. It works the great majority of the time but sometimes it reaches a decision which simply doesn't make sense and can be disasterous.

I have lost 2 armies in RUS in a PBEM game in 18 turns because of retreats that made no sense to me. I almost lost a third and may yet still lose it. So I started looking at the formula closely to try and understand what is going on. Two things stand out to me. First this part is excellent:

ctlRetreatPrevSubSpaceCoeffH = 250 // Coefficient applied to the interest if the region is the one where we are coming from

. However it appears to only apply to a unit which moved during the current turn. Which means a stationary defender or attacker is much more likely to retreat in odd directions than a moving attacker. Two, it appears land links are fairly important. Large and oddly shaped regions adjacent to smaller regions seem a magnet for retreat. Bearing in mind those two factors, seems to explain why I lost 2 and almost 3 armies. My conclusion is certain areas are traps for defenders. In Rus, I simply will not defend certain regions unless I am willing to accept a high probability of losing those armies if I lose the battle. If you have RUS, look at Ekterinberg. Clearing the obvious red position to block white advances on Perm and Kazan and a superb defensive position as well, right? Wrong, it is a trap. Also if I see an opponent defending in those areas, I am attacking first opportunity I get. Killing off armies is not easy in RUS but bad retreats can make it easy.

And again, algorithym works fine 95% of the time. But the reality is an algorithym is chosing a line of retreat based on numbers. And from time to time, the player will lose an army because it was stationary and the algorithym was not set up for an unusual arrangement of regions. A player makes decisions in a completely different manner and would never have made that type of mistake. Ideally, the player should have the option of designating a line of retreat. Then the player can only blame himself if his choice doesn't work out as intended.

[Jagger2013]

In effect, make a predesignated direction of movement as attractive as 'province moved from' in the algorithm?

Why allow an algorithym choose the direction of retreat when a player could make that choice?

As far as I can see, the algoritym choses a direction of retreat and then executes the move. If the player chooses a retreat city during the plot portion of the turn, the algorithym is not even needed, the game simply executes the move to the human designated retreat city-regardless of when and where the battle takes place. Think of the choosing the retreat city as representing your line of communications and the natural line of retreat.

[havi]

why not just make island10 a place where army cannot retreat! in my game against nabulio he just crushed my army comanded by PGT because he was allways trying to retreat that damned island!