Fixing the enemy units behind-the-lines problem

[LDiCesare]

Also, I think it would look better in the long run if only the terrain tile & terrain overlays were colored (not roads, cities, units, etc.)

That's what I did at first, (since it was easier), but you wouldn't see the color in Rome on the picture if I didn't also filter the overlay.
If you like the green part, I can change all the other filters to behave the same.
I only implemented the map filter for now, so I'll promote it and implement its effect on military conquest.

[alms66]

Originally posted by LDiCesare That's what I did at first, (since it was easier), but you wouldn't see the color in Rome on the picture if I didn't also filter the overlay.

You might not, I'm certain it would be perfectly clear to me.

Originally posted by LDiCesare If you like the green part, I can change all the other filters to behave the same.

I'm not sure what you mean...

Originally posted by LDiCesare I only implemented the map filter for now, so I'll promote it and implement its effect on military conquest.

Again... not sure what this means...

[LDiCesare]

You might not, I'm certain it would be perfectly clear to me.

Well, if I can't see it, then I won't be the only one. Units keep their color, and so do walls.

I ment grey not green. The filters are currently of two types:
- Switch the color to redder and leave it as normal if the value to be shown is 0.
- Be grey if the value is 0.
I can make all grey/colored instead of the current normal/colored.

I ment I committed my changes (map filter) to the repository and am working on the square ownership thing.

[alms66]

I see.
If it were up to me, I'd say all gray if the value is 0. The filters were never meant to be on all the time. They were a way to query for information. Once you've got your information, you would switch back to normal view.

[alms66]

Originally posted by LDiCesare
Here's what I propose, and an illustration of it.
Snip...

btw, in case you guys hadn't realized it yet, this is basically what demipomme suggested - or at least has the same problem as what he suggested, except instead of having to leave a single unit on every tile, you have to leave a group of units in an area to maintain control (and the number in the group will most likely be the same as leaving a single unit in each tile - it's just slightly more convienent)...

[LDiCesare]

I don't think demipomme suggested leaving units, but losing personnel which was used instead as "law enforcement", which would be abstract and not require management.
Also, if you don't conquer a square by moving it there, then the ai will have less incentive to do so: If the square doesn't change hands, then it is not a good target, so it would send troops far away only if they had a known goal far beyond the lines. So you would get rid of the problem by removing the incentive for the ai to attack. It would then have to fallback on other plans, like ergeing with other units and defending.

[LDiCesare]

I modified the code that takes ownership of a square:
1) It's computed only at the end of a turn, not at each tick, thus fast units won't conquer everything in their path by crossing 3 squares in one turn.
2) You can take control of a square only if your military influence is the strongest in that square (should be refined when alliances exist).
(I still have to adjust the ai to take point 2 into account properly. The current effect is units in enemy territory stop moving and stay idle.)
A nice side effect of (1) is to speed the game a bit.

[alms66]

Originally posted by LDiCesare
I don't think demipomme suggested leaving units, but losing personnel which was used instead as "law enforcement", which would be abstract and not require management.

Originally posted by demipomme
I think that the 'land' problem is more problematic than the 'unit' problem. The only reason why I want to track down and kill the stray enemy units is because if I don't, they conquer my entire province except by well-defended city. If it were the case, and I have already suggested it, that control of a square reverted to its original owner unless held for several turns, then a single enemy unit wouldn't be able to take large swathes of my territory. The second advantage is that the enemy unit would have to stay still to gain anything and thus I can track it down while its stays still.

I think this solution is realistic as it models for the change of local government & police etc. of the square. The same can be applied for province capitals i.e. you get to keep the entire province if you hold the capital for several turns.

Now I don't know where his original post on the idea is, but this was my reponse, and seeing as I was never corrected, I assume I understood his post correctly:

Originally posted by alms66
The idea in and of itself is sound. A unit should have to move into a tile and stay there for that turn, at least, to change ownership. Having more than 2 turns would probably not be worthwhile though. It would bog wars down as players have to leave individual units behind the lines to "occupy" the tiles in order to gain control, or worse if we require the occupying force be larger depending on the population and resistance of the tile. Then, they'd have to move them once ownership changed, it would be a hassle.

Now, all you've done is grouped those single units that would have had to stay behind to take control of individual squares and allowed them to take larger swaths of territory - but they still have to stay behind to control that territory or it reverts to the owner. It's still going to be a hassle to leave the units behind, just less of one than what demipomme suggested. Obviously, I haven't given the thing a try, and I could be wrong, but from what I've read here, the above seems true enough to me.

Originally posted by LDiCesare
Also, if you don't conquer a square by moving it there, then the ai will have less incentive to do so: If the square doesn't change hands, then it is not a good target, so it would send troops far away only if they had a known goal far beyond the lines. So you would get rid of the problem by removing the incentive for the ai to attack. It would then have to fallback on other plans, like ergeing with other units and defending.

You really do a remarkable job of giving me responses that simply confound and confuse the hell out of me...
What's this about?

[demipomme]

Just to clarify what I did and did not say on the subject:

I did suggest leaving behind one unit per square to conquer land. However it was pointed out that this would require too much micromanagement, so instead I suggested the 'law enforcement' approach.

I still think the one unit per square stay behind to keep is good. It creates micromanagement which forces attacks on the province capital. I think this is more realistic. How many nations were conquered by taking all of their territory and then the capital last? Not many that I know of - armies have historically targetted the capitals to take out command and control.

I personally don't like the military influence approach for the following reasons:

-given the size of tiles the distance of military influence is unrealistic
-why would the governor of a square ever turn over power just because the neighbouring square had an army in it? no matter what the size of the square people only give up power when they really have to.

[LDiCesare]

You really do a remarkable job of giving me responses that simply confound and confuse the hell out of me...

Didn't I say I was poor at communication?
Consider this: The ai sees an enemy square. It decides to attack it because there is something to be gained there (namely the square, though the goal could be to destroy enemy units). If the square can't be conquered by sending some armies there, then it's useless to attack that square with this army, so the ai will choose another, better plan.

why would the governor of a square ever turn over power just because the neighbouring square had an army in it?

The rationale is the other (stronger) army is harassing the occupying army or can make occasional raids from its position and still deal enough blows to the occupying army that it can't control and benefit from the square, though it should probably be able to pillage it.

[alms66]

Originally posted by LDiCesare
Consider this: The ai sees an enemy square. It decides to attack it because there is something to be gained there (namely the square, though the goal could be to destroy enemy units). If the square can't be conquered by sending some armies there, then it's useless to attack that square with this army, so the ai will choose another, better plan.

Ok, sorry to be a pain but, now I understood what you said and I'm still confused, particularly by this line:

Originally posted by LDiCesare
Also, if you don't conquer a square by moving it there, then the ai will have less incentive to do so:

I don't see how it relates to my suggesting that the new 'control' method of conquering, is just demi's idea, minus a bit of micromanagement. You no longer have to leave a single unit in each square you conquer to hold the tile as per demi's idea, you simply have to leave a small army (probably about the size of 1 unit per tile of area you want to maintain control in) in the area to maintain control of the tiles you've conquered (your new way).

[LDiCesare]

Yes you have to leave some units but you can keep them near the borders because the ai won't try to cross the "zone of control" unless they know a big juicy target lies somewhere beyond. In that sense it is similar to demipomme's idea.

[demipomme]

Originally posted by LDiCesare
The rationale is the other (stronger) army is harassing the occupying army or can make occasional raids from its position and still deal enough blows to the occupying army that it can't control and benefit from the square, though it should probably be able to pillage it.

I'm glad there is a rationale, but I would point out that control is usually maintained through the rule of law and law enforcement, rather than by the army.

Occasional raids would decrease economic effectiveness of a tile, but not give control. A useful way of modelling this would be for a tile of civilisation X near to a big army of civilisation Y would stay in control of civilisation Y but with some production going to X (to represent raids) and some production lost entirely (to represent disruption of an enemy army raiding).

I still think the most realistic and simple way for terrain to change hands is through the capture and retained control of the capital.

[LDiCesare]

but I would point out that control is usually maintained through the rule of law and law enforcement, rather than by the army.

I tend to disagree. That's true when there are no riots, and historically it's not very obvious, particularly in occupied countries. Furthermore, some law enforcement is actually military (that's the case in France where there is a military force -gendarmerie- whose main mission is one of police).

Limiting conquest to control of the capital is indeed simple, but prevents one from conquering part of a province. I thiink there should be a way to take control of a few squares anyway. If you consider the fall of Constantinople, when the Byzantine empire finally fell, it didn't control much more than its own capital, so conquering squares and eating part of another province should be possible..

[Mark_Everson]

Hi All:

I have now played the code update with the military influence effects in it. That is commit #177. Thanks for implementing it Laurent! I haven't played it a huge amount yet, but based on limited testing it does appear that the changes help a fair amount. The influence map is also fun to watch. (BTW There is a bug where the influence map doesn't update properly. The bright red from Hannibal's passage through Gaul [while playing Romans] remains even when he is long gone and in the Italian peninsula!)

I think influence will be especially valuable when we get to square reversion as discussed above. Basically, the military influence that a square is bathed in should have a large effect upon its reversion chances. People aren't stupid, if they know that their revolt can easily be crushed, they will tend not to revolt, and wait for a better time.

I think that between the current changes in the square conquest model, and a reversion implementation, we should be about 70% of the way to solving the problem. As has been mentioned before, some way to slow down or reduce the size of small invaders that would allow them to be finished off more easily would also be beneficial. I think the idea of having to leave behind mini-garrisons to establish control of a square, as suggested by demi pomme, could be interesting. I'd be in favor of trying it if we can figure out a way to do it with relatively little effort.

I actually don't much like the current implementation where taking the province capital results in flipping the whole province ownership. (Aside from where enemy military units are present) We simply did that as a means of avoiding having to take every province square-by-square. Taking the capital is important because it is usually a center of administration, and transportation among other things. However, if a unit should sneak into the capital while greatly outnumbered by opponents in the rest of the province, I think the effect should be much less severe than we have currently. Now that we have the influence mapping, my proposal would be that when you take the province capital, you also get along with it any squares in the province where your influence is positive. That way, a small unit sneaking in around larger opposing forces and grabbing the provincial capital will only result in the loss of the capital and perhaps a few squares around it. Conquering the capital with a large TF as part of an even larger invasion force would generally result in flipping the entire province as we have now.

What do people think about trying this change?

In regard to the specific influence formula (total attack of army)/(time in ticks to reach square + 10), I think I would favor something that died off a bit faster at longer distances. Something like raising the denominator to the 1.5 power. Laurent also noted that he cuts off the calculation at two turns worth of move. I think I would prefer it to go out to 3 or 4 turns of movement, unless that becomes too expensive to calculate.