CEP: Armies

[ExpiredReign]

... can the AI retreat mechanic be linked to how much HP is left in their army? For example, let's say I'm attacking with 4 units. The total HP of my army is at 400. All 4 units take some damage, and the total HP of my army now goes down to 190. From a player's perspective, it makes sense to pull your troops back and heal them. Can this type of logic be coded into the AI? Of course, it will be more complex than this simple example (i.e. have to think about what the total HP is, and how it changes as units die).

I think this idea has great merit.

If the AI calculates the attack/defeat ratio based on 'army status' rather than, or as well as, individual unit health or numbers then it surely would make a difference.

Humans will break off combat if, despite having more archers say, they have no melee units for the coup de grâce.

Not sure how much coding this will take though.

 

[stackpointer]

I suspect this would take a ton of coding if it is possible and would increase AI turns. I suspect that programming tactical movement for the AI is incredibly difficult to do. Which is why the AI is so bad at it and why it's taken so long for the AI to be improved to the point it is today (still bad).

Although, if you are just modifying the conditions that the AI will retreat instead of the actual tactical movement then that should be much easier to modify. I think that those conditions might be located in the DLL files though.

 

[mitsho]

I see two sides to this

Army Composition: A issue of the economics side. Compared to the optimal mix, AI armies imho should have more vanguard/soldier/mounted/et al. units since they lose them more often than the ranged ones, possibly something prioritizing rebuilding them at home would be helpful as well.

Combat issue: Extra sight for AI units certainly helped in GEM and should be implemented, though that only goes so far. Can go on the Admiral for Ships if we don't want every ship to have them... BroOftheSun's proposal seems like worthy to try as well.

The congestion issues at last come from having many units in the field. If there are few, movement should be easier for the AI to calculate, no? So they should concentrate on better (i.e. upgraded = cheaper costs for them) and more experienced units. But what if they'd attack not one city, but two? Eh, probably not...

I'd say one of the biggest problem for the AI is that it choses to attack the easiest target, and not the nearest one. In all my conquest game - unless they built the Great Wall - I always attack the civ next to me, and not the one behind that, even if the latter is a small Babylon that may potentially grow into a behemoth...

 

[ExpiredReign]

...
I'd say one of the biggest problem for the AI is that it choses to attack the easiest target, and not the nearest one...

That is very observant, and relevant. Now that you mentioned it I too must concede that the AI does this in my games as well.

I wonder if a 'DistanceIntoEnemyTerritory' or 'DistanceFromFrontier' type factor could be introduced to handle this behaviour?

Something to make the AI re-evaluate the wisdom of targeting that weak interior city.

 

[Dunkah]

Combat issue: Extra sight for AI units certainly helped in GEM and should be implemented, though that only goes so far. Can go on the Admiral for Ships if we don't want every ship to have them... BroOftheSun's proposal seems like worthy to try as well.

This did make the Ai better in GEM.

However I hated the fact that Artillery or English Longbowman, or any unit with a range increase, (which was pretty common in GEM after Thal added the automatic promotions), could see me and attack my from the fog of war without me even knowing they were there.

There was one particular game where the AI had like 4 units out of sight. I would all of a sudden lose a unit. boom. gone. since i had fighting going on in other locations I didn;t think much of it. Then another one went. I went to investigate and before I could see the units boom that unit was gone.

I literrally had to bring my navy in from another direction to actually see the units before I could even react to what they were doing.

It would be cool if we could somehow give them the extra sight without allowing them to use that in combat.

 

[mystikx21]

We could potentially limit what AI units get sight bonuses. That would still mean artillery/longbows could kill you, but at least you'd see there's an army over there harassing your units because of the scouts.

 

[Ahriman]

Yeah, I wouldn't add sight bonuses to tanks, archers, or siege units.

The longbow problem was mostly because their ranged strength was too high, and because they had indirect fire.

 

[Stalker0]

The longbow problem was mostly because their ranged strength was too high, and because they had indirect fire.

I think that range 3 may simply be too abusable.

Just as we removed move after attack for hunter ships because of abuse, the same may need to be considered for longbowmens, it is a crazy powerful ability.

 

[mystikx21]

You can get it on any archer unit. I don't see the problem with a 3 range unit. The problem with longbows was that they had it to start AND had indirect fire AND were still strong ranged units (and could get 4 range because of the GEM change to the way longbow range worked).

2 of those three is probably fine. A minor strength reduction would work too.

 

[Stalker0]

You can get it on any archer unit.

I actually feel that the range promotion is broken myself

 

[mystikx21]

Right, I seem to recall that debate over battleships at range 3 back in GEM. I think that promotion is fine since it requires a good deal of investment to get it, but this is also because there are later game units that already fire that far by default.

The AI doesn't seem fundamentally worse at handling range. Move after attack with high movement rates, blitz, and the actual unit AI of ranged units in general, sure. I think that's more where I'm coming from on that front. The human is better at using 3 range units because the human is better at using ranged units basically overall.

 

[Ahriman]

The midgame unit design just doesn't work at all. Industrial era musketmen (strength 25) are barely stronger than medieval longswords). That can't be the intended design....

Modern era riflemen are only strength 30, Infantry only strength 35. This compared to industrial era gatling guns that are strength 35 with a ranged attack.

 

[Anvari]

Well both units aren't so different in vanilla BNW. LWM have 21 strength for 120 hammer (1 strength=5.71hammer), Musketman have 24 strength for 150 hammer /1 strength=6.25hammer). But you don't need iron so i think they are fairly balanced. The biggest problem are attacking into units and promotions because LWM probably have more, which causes you trouble. I think you just need more Crossbows for backup if you want to wage war in that era.

Usually i try to avoid wars in that era, part is the minor upgrade advantage if you are attacking, compared to Swordman to LWM (14-21 strength or 21-24). The second part is i don't want to get more diplomatic maluses then necessary, so a break in that era is win-win. At least in my eyes.

 

[Ahriman]

Well both units aren't so different in vanilla BNW

But lower level units have been scaled up significantly in CEP. Swordsmen, longswords, knights are tougher in CEP (longswords in CEP are strength 23 as compared to 21 in BNW). So are aircraft and many other units. But the main infantry units are incredibly weak.

Also remember that "muskets" in CEP are riflemen in BNW - they're the industrial era unit.

BNW muskets are strength 24; in CEP they're strength 20.
BNW rifles are strength 34; in CEP they're strength 25.
BNW Great War infantry are strength 50; in CEP they're strength 30.
BNW Infantry are strength 70, in CEP they're strength 35.

Something is wrong, I suspect a strength formula was misapplied. In the current build of CEP the marginal strength increases from tech are very low.

 

[Ahriman]

I think it might help if there was a clear design exposition for military units, including strength values and upgrade lines, and comparisons to base BNW values. That might make seeing the gaps more obvious.

 

[mystikx21]

I think it would be better to have a line that looks like this
Arques: 23
Muskets: 30
Rifles: 40
Infantry: 50
Mech: 60

But then, this is the line we had before in GEM. I'm not clear entirely on why it was beneficial to reduce cost and strength of mainline units. Higher kill ratios are useless if the units are destroyed so easily as to be worthless to build.

Currently the strength is set at .75. It would probably scale better at .9, or something closer to that. Elite infantry no longer is a factor at the gunpowder age, so it doesn't seem necessary to weaken these ground units artificially (or to just use them at 1 values straight off).

Bumping later game ground units overall by about 5% would probably help too (tanks, artillery).

 

[Ahriman]

Well, once again, the simplest thing is just to have a single line of main infantry units from the Renaissance onwards that longswords and pikemen both upgrade into. They don't get the city attack bonus of swordsmen and longswords, but they don't need any particular strength penalty either.

Honestly, I don't see what is wrong with the vanilla values for the infantry units. The vanilla strength values for the main infantry units are the logical starting point for a combat design, then balance the strategic resource units around them.

The current values are broken. We have the GEM values for some units but not others. We have for example bombers that have a ranged attack of 50 and +50% vs land units and ships, vs equivalent era infantry with a strength of 35.

 

[mystikx21]

The issue with the vanilla values is that they put the tanks/horses at a large value disadvantage (along with the early swords units). It's probably easier to balance at a lower level overall for mainline units with tanks/horses having higher value. I think we had a pretty good balance overall in GEM, other than vanguards and a couple of unit role swaps on naval units and that that was probably the proper place to start with unit balance with a few of the new options available from BNW.

What seems to have been the idea was to bring back vanguards as the main units but not as the special unit class. Lowering city strength on the high end helped with this new level of balance, but didn't fully address the change, and left incomplete the unit balance. Why it was necessary in the first place isn't really explained above anywhere either. We could make units cheaper in cost and upkeep without effecting their powers on battlefields as a point of balance, if that was a desirable impact.

 

[mitsho]

I think it might help if there was a clear design exposition for military units, including strength values and upgrade lines, and comparisons to base BNW values. That might make seeing the gaps more obvious.

Like in the excel file in the mod folder of the version you downloaded? Take a look at the CEP_ArmiesCities file, while they have been copied from GEM and some pages are old, the sheet UNITS may contain what you are looking for. (It even has CEP written on top ).

It's hard to have the overview and I don't get to play enough so I won't comment on the strengths and all that, not sure what to argue/think here

 

[Ahriman]

GEM values are probably ok as a start point, but I don't really remember what they were. The ones here:
http://communitas.wikia.com/wiki/Gem_Units

seem to be higher than I remembered. 41 strength knights??
Or maybe these were the hypothetical values once various modifiers were added?

I guess the point is that city strength, defensive building bonuses, and military strengths need to all be balanced together. It's hard to do that without a clear design document with all that in front of us.

But I think a basic design goal should be that the main infantry units should be the core jack-of-all-trades of a post-medieval military. Cavalry and tanks will be faster and have move after attack, and may in some cases have higher stats, but not radically so. Each era should represent something like a 25-40% strength bonus over previous eras.

Like in the excel file in the mod folder of the version you downloaded?

I'll take a look at that, thanks, but I think it is fairly clear that the current stats are broken. [If there was a decision to reduce military strengths, it was applied unevenly.]
So it would be helpful to have a sheet that has a design goal that can be discussed.