Improving Warfare

[Taefin]

I’d like to hear folks’ thoughts on the dynamics of warfare in the Victor OpenDev. After a few games playing as aggressively as I could, at various levels of figuring out the game, some of my observations were:

I enjoy how much units get worn down as you press toward an opponent’s city. Having a few hostile territories to cross, where the AI will engage you in as many battles as they can, I found my armies dropped to mostly half health by the time I initiated the siege. This creates a neat dynamic where you want to begin the siege with a smaller force of healthier units while you spend the handful of turns of siege healing up the others. This should give the AI a good chance at sortie/reinforcement. However, I have not yet seen them do either.

I thought the discrete battle mechanic and 3 rounds of combat per turn worked well for allowing ground to be gained/lost and making more memorable moments than the feeling of trench warfare at every era of Civ6.

But it might make it a little too easy for a slight strength advantage to snowball. The victor can usually move weak units off the battle line and avoid most casualties while annihilating the opposing force.

I wonder if a mid-battle retreat option would help. Perhaps one side could opt to retreat before making their moves for a combat round, or perhaps before resuming combat on the next game turn. Perhaps the other side’s units could automatically attack any adjacent retreating units. This would help the AI lose fewer units as they wear down the advancing force. It would also create another use for cavalry, to catch up to and engage a retreating army.

The AI really needs military units in their cities (and more population/districts) to ward off a siege/assault. A combination of a mid-battle retreat option, and more direct programming of AI for when to switch from open warfare to city defense, would make siege warfare a lot more enjoyable.

Also, it seems the AI needs to be more aggressive in declaring war, especially against an opponent who is engaged in another war or not defending their territory. For this to work, the war support might need to have a fix where a side cannot be forced into vassalage unless at least one city is taken. Or perhaps not having enemy units in your territory should give a upper cap on war support, so that a player cannot take a vassal simply by repelling an attack, but would need to advance, or else negotiate a more mutual peace.

 

[Boris Gudenuf]

I'm primarily a military historian (16 books and writing another now) so I was really interested in the Humankind 'battle system' from first hearing about it - and because I think the 1UPT version of battles in Civ VI is wretchedly out of scale both in time and space in a Grand Strategy game like a 4X 'semi-historical'.

I’d like to hear folks’ thoughts on the dynamics of warfare in the Victor OpenDev. After a few games playing as aggressively as I could, at various levels of figuring out the game, some of my observations were:

I enjoy how much units get worn down as you press toward an opponent’s city. Having a few hostile territories to cross, where the AI will engage you in as many battles as they can, I found my armies dropped to mostly half health by the time I initiated the siege. This creates a neat dynamic where you want to begin the siege with a smaller force of healthier units while you spend the handful of turns of siege healing up the others. This should give the AI a good chance at sortie/reinforcement. However, I have not yet seen them do either.

The game simulates "attrition" very well, in that outside of your own territory it is very hard to keep up Unit strength. I would not be surprised to see something in the late Industrial/Modern Eras (which they've kept under wraps in all the Open Devs so far) that simulates the massive Infrastructure of Supply Lines, Supply Depots, Replacements, Hospitals, and other military 'overhead' that keeps front line strength up in modern armies. Famously, the US Army in WWII had almost 5 men in administrative, supply, and other non-combat positions outside the divisions for every man in a front-line division. Would be nice to see that 'price' shown in a game for once.

I thought the discrete battle mechanic and 3 rounds of combat per turn worked well for allowing ground to be gained/lost and making more memorable moments than the feeling of trench warfare at every era of Civ6.

But it might make it a little too easy for a slight strength advantage to snowball. The victor can usually move weak units off the battle line and avoid most casualties while annihilating the opposing force.

The tactical system has a huge amount of interaction among the units modeled in Special attributes and bonuses, far more than are usually identified in Civ games. For instance, trying to keep your archers safely back away from enemy melee units and discovering that an enemy Horseman not only ignores all Zones of Control, but can also zip right around your flank and hit your archers from behind in one Bound. Definitely an Oh S**t! moment, as we say. Or discovering the joys of Superior Terrain, with your pkemen charing downhill onto a hapless enemy swordsman. Memorable.
By scribbling furiously throughout the Victor Open Dev, Ive managed to compile a list f the Special Attributes assigned to units and Emblematic units in the first 4 Eras, and it's two pages long! That's an incredible variety of tactical variations and bonuses.

I wonder if a mid-battle retreat option would help. Perhaps one side could opt to retreat before making their moves for a combat round, or perhaps before resuming combat on the next game turn. Perhaps the other side’s units could automatically attack any adjacent retreating units. This would help the AI lose fewer units as they wear down the advancing force. It would also create another use for cavalry, to catch up to and engage a retreating army.

The AI already knows (or at least has demonstrated them to me!) some tricks with Retreats. Because an enemy that retreats before combat stops all movement by the opposing force, I have had an AI attack me with a single unit against my army, then run away, do the same thing with a different unit the following turn, and basically fix my entire army in place with two lousy scouts - until I split my army into two groups and used one to pursue the retreating scout and wipe him out. I've also had an inferior enemy force make excellent use of the battlefield terrain - setting up as far away from me as they could, so the 1st or 3 rounds is used up just getting to them, or using constricting terrain to keep only a portion of their force in harm's way. A mid-battle retreat option is not necessary, I think. Although I could change my mind after I see what off-battlefield artillery, naval or air support from many tiles away does to an inferior force later in the game.

The AI really needs military units in their cities (and more population/districts) to ward off a siege/assault. A combination of a mid-battle retreat option, and more direct programming of AI for when to switch from open warfare to city defense, would make siege warfare a lot more enjoyable.

Siege warfare during its prime time in the 17th - 18th centuries was a very different form of warfare. It was understood that you needed one force to prosecute a siege, and another 'covering' force to keep enemy field armies from interfering in the siege. On a couple of occasions I had AI forces go around my besieging force and do some Ransacking in my nearest region, forcing me to either detach part of the besieging force or call up other forces to chase them down. That's in keeping with historical practice, but I agree the AI at the moment does not seem to pay enough attention to garrisoning its cities.

Also, it seems the AI needs to be more aggressive in declaring war, especially against an opponent who is engaged in another war or not defending their territory. For this to work, the war support might need to have a fix where a side cannot be forced into vassalage unless at least one city is taken. Or perhaps not having enemy units in your territory should give a upper cap on war support, so that a player cannot take a vassal simply by repelling an attack, but would need to advance, or else negotiate a more mutual peace.

I remember the AI being much more aggressive in the Lucy Open Dev. In Victor, they seem remarkably passive, at least with me. On the other hand, every game (I managed to complete 6) included several "Foreign Wars" between AI factions out of my sight, so the passivity was by no means Universal. In all 6 games 1 to 3 Factions ended up as Vassals of other Factions, so somebody was fighting and prosecuting wars to the point that the other side was completely beaten.
On the other hand, I agree that the 'War Support' calculation perhaps needs to include some more factors. Battles without friendly casualties, or avoiding any Ransacking or loss of your own Outposts - signs of a One Sided War in game terms, should have some kind of effect on your own people's perception of the war as Painless and the other side's view of the war as Hopeless, and shift the War Support in your favor. Right now, regardless of how successful you are, War Support trends almost always steadily down the longer the war lasts. This may be true in Modern times, when mass communications keep the dark side of war firmly in the collective consciousness, but was not true earlier, when the nastiness was out of sight as long as you weren't on a battlefield or in the path of the rampaging army (from either side!). Even in the 20th century (Modern Era) Victories could be announced without (much) contradiction and earlier loot and captives coild be paraded to the public to convince everyone that the war was Going Well.

 

[Taefin]

Awesome points, thanks for sharing!

I too have noticed AI fighting and putting each other under siege, a welcome relief from Civ6, and I notice with a few vassals under their belt, an AI can really take off in the tech tree and fame. Also I’m guessing the war support system and unevenness of battle resolution helps one AI actually win wars against each other.

 

[8housesofelixir]

AIs in Victor opendev really have a skill in concentrating their main force in the fog of war, and surprised you with throwing them out via huge reinforcements.

I had two battles with fairly developed AIs, in which I opened the battle with attacking a relatively isolated army, with only another army in my sight. The Tactical Map unfolded with my main army facing about 8 AI units.

On the very next turn, in both cases, AI immediately send 3-5 armies under the fog of war, the Tactical Map expands, and I found myself facing 20-30 units instead of the 8 just encountered. The tensions of the battle suddenly raises and I was forced to concentrate all my resources into this battle as well. Even though I came out with a victory in the end, the battle still feel like an uneasy win.

In short I do respect AI's operational skill in case of a main force engagement.

 

[mitsho]

Those stories are great.

I had two situations. One where I didn‘t pay attention and let auto-resolve - bad idea, but it was also a badly chosen battleground (too tight for my cavalry). And one where I was put under siege, tried a sortie which misfired (I was testing a bit) and then would have retreated back mid-fight. But I could only put auto-resolve and see the rest of my forces slaughtered. Maybe instead of a retreat a „white flag“ button? You keep the units alive, pay some ransom or I don‘t know. Maybe that button already exists and I didn‘t see it?