[R&F] The State of Combat Units (Post Spring 2018 Patch)

[Sostratus]

If a combat system features several unit types and a rock - paper - scissors approach, it needs to implement those correctly.

Civ6 combat seems to have a very "age of empires" RTS inspired theme. It's quite a good system improvement over civ5 (ranged beats all) and they put some detail into it.

There's the classic "Ranged beats Infantry beats Cavalry beats Ranged" cycle in the conceptual design. These are some of the other specific modifiers:

One feature of the ranged line I tried to highlight in prior posts is that they have a ranged strength 10 points higher than their melee strength. This means they take 50% more damage from non ranged attacks.
Ranged deal half damage to cities, to give siege weapons a place; bombard units deal half damage to units, to keep them in check.
If melee and anticav were better balanced with each other, they would still both be countered by ranged; this is why I suggested earlier that melee should be heavy infantry that can shrug off ranged blows rather than general-purpose units that slaughter anti cav. I just think that melee should have a more fundamental interaction with unit classes that aren't its fellow foot soldiers.
Mounted have the innate advantage of movement, which helps them negate the perk of ranged units-range. Unfortunately, they have no downsides- they are just more mobile melee units. IMO, cavalry should suffer a nice -10 to -17 penalty against cities to really orient them as dominant on the field but encourage players to have other units to crack defenses. (I also think they shouldn't get siege tower/ram bonuses, or defense from terrain. But, at least a penalty vs walls.)
This was a feature they added into Civ5 when horseman spam was too dominant. I'm surprised it's not carried over- after all, mounted has quite literally no drawbacks at the moment (other than needing resources which is a flaw that cannot be fixed by Civ6's resource system; it's either irrelevant if resources are common or makes the game heavily luck based if the resources are scarce)

When it comes to what is the correct level of production costs for the units, that answer has to be based on a lot of quality multiplayer data and fine tuned adjustment. Basically what competitive games like starcraft do- the real world is the best way to find out how people value intangibles like mounted movement, range vs melee, etc.

 

[Boris Gudenuf]

I've always viewed the Archer unit in Civ as the exception to the historical rule: the trained, disciplined troop of soldiers who can stand their own on the battlefield, such as the English or Indian bowmen, or Italian mercenary crossbowmen. Because that's how they're depicted in the game. The conditions that gave rise to those types of troops in real life were, of course, very rare.

Even here, the surrounding conditions and tactical combinations had as much or more to do with the 'success' of archery as did the archers themselves. The ancient Indian archers were part of armies that also had heavy chariots, elephants, and infantry armed with long two-handed swords - lots of people to distract anyone from charging the archers! Aside from the knights and stakes already mentioned in the earlier post, you put your finger on one major advantage of the English 'yeoman archer': the men trained together in their villages, and on the battlefield were commanded by familiar sheriffs from back home, so it was possible to direct and coordinate their fire on targets, instead of merely filling the sky with arrows flying hither and yon. Still wouldn't protect them on an open field when faced by horsemen, but if anything slowed down the horses, the longbows could smother selected groups (like the king and his bodyguards, the supporting crossbowmen, and other Lucrative Tactical Targets), which simply could not be done by a motley group of archers, no matter how good at shooting, assembled for the occasion with no prior training and organization.

The major problem for me (in this otherwise excellent thread) is the assumption
by some people that units, e.g. archers, are idealised versions of a real
collection of men armed with bows and arrows.

The thought of archers occupying a region and shooting arrows at an opponent
for 20+ years just doesn't work for me.

I prefer to think of the pressure of the technology of archery applied to
a particular region of the map. That also includes enhancements gained via
other means, civ-specific characteristics for (in this example) archery, and
pressure via many other means and methods applied to that region and surrounds.

Thank you! for reminding all of us that in the game's Time and Distance Scales, all tactical considerations are Fantasy.
However, having embraced 1upt, Tactics, however realistically inane in the context of the rest of the game, becomes part of the game: there is simply no other reason for having 1upt at all.
That means we have to look at the tactical considerations of the units, even though we know that any single unit has to represent an army-sized (whatever size that is in the given Era) group ranging over X 100s of square kilometers and 1 - 40 years each turn.

For me units don't need to be very historically correct. I am fine with generic units being an amalgam of several historic inspirations. Nonetheless I would appreciate balance changes but those should evolve around the tactical / game side of things. If a combat system features several unit types and a rock - paper - scissors approach, it needs to implement those correctly. If melee crushes anti-cav but anti-cav struggles against cavallery there is something wrong.

No question. Whatever mechanism/factors/numbers/characteristics are implemented in the game, they have to work in terms of the game. I'm far more of a (military) historian than I am a gamer, but I will be the first to argue that the most historical mechanism or unit that does not play, will not be played and is a waste of game designer's time and game player's money AND time.

And when we argue 'tactical' considerations for 'tweaking' units to make them playable - and enjoyably playable - 'rock-paper-scissors' is precisely the right approach. All tactical combinations are attempts to resolve the deficiencies of some units by combining them with terrain, position, weather, or other units to minimize their failings and maximize their strengths. Right now, as stated, Anti-Cav seem to have NO strengths in the game, and that is the crux of our problem.

 

[Boris Gudenuf]

Again, an outstanding analysis o the Game Problem with Units!

Civ6 combat seems to have a very "age of empires" RTS inspired theme. It's quite a good system improvement over civ5 (ranged beats all) and they put some detail into it.

There's the classic "Ranged beats Infantry beats Cavalry beats Ranged" cycle in the conceptual design. These are some of the other specific modifiers:
View attachment 495872

One feature of the ranged line I tried to highlight in prior posts is that they have a ranged strength 10 points higher than their melee strength. This means they take 50% more damage from non ranged attacks.
Ranged deal half damage to cities, to give siege weapons a place; bombard units deal half damage to units, to keep them in check.
If melee and anticav were better balanced with each other, they would still both be countered by ranged; this is why I suggested earlier that melee should be heavy infantry that can shrug off ranged blows rather than general-purpose units that slaughter anti cav. I just think that melee should have a more fundamental interaction with unit classes that aren't its fellow foot soldiers.

First, Suggested Game Definition: the Civ VI 'Melee' units appear to be representing groups with individual weapons (swords, 'dagger-axes', maces, clubs, etc) who do not require fixed formations and group orientation, like spears or pikes, to be effective. Historically, the big advantage of swordsmen of all kinds was their flexibility: not requiring fairly rigid formations, they could move and fight equally well in hilly or wooded, bushy terrain where spears and pikes were handicapped trying to maintain their formations. The historical Disadvantage of swordsmen and similar troops was that they required continuous training and practice to maintain their proficiency both as individuals and units, and well-forged and tempered swords are more expensive than simple spear/pike points. This suggests that Melee should have (as we've discussed already) higher Production and Maintenance Costs than Anti-Cav in general, and perhaps an Increased advantage over Anti-Cav on Hill, Forest or Rainforest terrain/tiles, but the exact numerical changes will need Play Testing.

Mounted have the innate advantage of movement, which helps them negate the perk of ranged units-range. Unfortunately, they have no downsides- they are just more mobile melee units. IMO, cavalry should suffer a nice -10 to -17 penalty against cities to really orient them as dominant on the field but encourage players to have other units to crack defenses. (I also think they shouldn't get siege tower/ram bonuses, or defense from terrain. But, at least a penalty vs walls.)
This was a feature they added into Civ5 when horseman spam was too dominant. I'm surprised it's not carried over- after all, mounted has quite literally no drawbacks at the moment (other than needing resources which is a flaw that cannot be fixed by Civ6's resource system; it's either irrelevant if resources are common or makes the game heavily luck based if the resources are scarce)

Historically, it was very, very hard to get horsemen to get off their horses, and practically/physically, it is almost impossible to get a horse to climb a ladder. That makes any kind of Mounted Unit almost worthless when it comes to taking a city. The penalties for attempting to use them in this way should be severe: at least a 50% reduction in Combat Factors. This is true for ALL Mounted Units throughout the game: taking tanks down city streets without copious infantry support is not a way for any tank crewman to reach retirement age in any army in any tank.

This, by the way, would also allow the game to show some 'special' Mounted Units: Alexander regularly got his Hetairoi or Companion Cavalry to dismount, pick up shields and go up ladders with sword, armor and shield to clear the battlements and take cities. Peter I of Russia in his initial 'regular Russian Army' had all dragoons as mounted units, and they could all fight mounted or dismounted. The first United States mounted units were 2 regiments each of Cavalry, Dragoons, and Mounted Rifles, but tactically, they were all ready to mount and charge or dismount and shoot, skirmish, defend or charge - a tradition that continued in the US Army's mounted arm right down to 1942 when the last of them traded in their horses for armored vehicles and jeeps.

Disregarding the 'Specials', I suggest that we should look at:
Severe Penalties to Mounted versus cities, forts or walls
Penalties to Anti-Cav facing Melee in 'rough terrain'
Increased Maintenance/Production Costs for BOTH Melee and Mounted units compared to Anti-Cav.
Factors of Anti-Cav and Melee much more equal, without Promotions, in open terrain
Factors of Anti-Cav being significantly higher than Mounted either with or without Promotions, or what's the point of Anti-Cav?

"Rock-paper-scissors' again, or perhaps "Rock-Paper-Scissors-Lizard-Spock-Massacre"...

 

[criZp]

which means the force of the arrow is simply that of gravity - all the 'pull' of your bow disappears when the arrow hits the top of its arc.

Just wanted to comment on this. Gravity does not affect the horizontal velocity of arrow, only the air drag does.

And some people have commented that civ gets "anticav units" wrong. What one has to remember is that anticav units are not dependent on any resources in the game, so they has to have a drawback, to make the strategic resources and their units worth going for. The reason anticav and cavalry/melee cost the same is because the latter require resources.

But the combat system would perhaps be more interesting if no units required particular resources, instead the resources gave other combat benefits of some kind.

 

[Boris Gudenuf]

Just wanted to comment on this. Gravity does not affect the horizontal velocity of arrow, only the air drag does.

I was not speaking of horizontal velocity, but the velocity that remains after an arrow fired above the horizontal to get more range, reaches the top of its arc of flight. Beyond that point, the initial 'pull' or force of the bow is not acting at all, it has been superceded/overcome by gravity, and the arrow falls back to the target with only the force of gravity minus any drag. The 'direct fire' or horizontal velocity of the arrow is, of course, primarily affected by the force imparted from the bow and drag. Note, however, that by the middle ages the effect of drag was pretty well understood even if the mathematics of the physics wasn't, and the shafts of the 'clothyard' arrows of the English longbows were tapered - streamlined - to reduce drag.
Gravity does effect the horizontal velocity, though, in that the horizontal velocity is actually a vector of the initial force from the bow directed either horizontally or, usually, slightly elevated, and the vertical downward pull of gravity, which begins to operate as soon as the arrow leaves the bow. Fire an arrow precisely horizontally, and it will hit the ground simultaneously with an arrow dropped from the same height, drag being identical. Gravity is one of those constants that effects most movement in almost any direction relative to the planetary surface - God Knows I spent enough time incorporating its effects into calculations as a Fire Direction Chief in the army!

And some people have commented that civ gets "anticav units" wrong. What one has to remember is that anticav units are not dependent on any resources in the game, so they has to have a drawback, to make the strategic resources and their units worth going for. The reason anticav and cavalry/melee cost the same is because the latter require resources.

But the combat system would perhaps be more interesting if no units required particular resources, instead the resources gave other combat benefits of some kind.

Resources is a whole other bag of saber-toothed worms that should really be addressed separately and at length. I will agree that 'Resource Requirements' are almost always an abstraction, and in the game frequently just flat wrong.

 

[acluewithout]

Giving light cav a -17 to cities or walls seems like a no-brainer.

I wouldn’t give a negative to heavy cav - they are meant to be fast and hard hitting units. They should be able to take cities. For Knights, just assume they’re getting off their horses. I’d rather see Knights balanced by increasing their costs, so you have to pay for that fire power.

Anti-Cav also need a buff against ranged like melee get, ie through their promotions. Their vulnerability to ranged compared to melee and heavy cav really undermines any offensive role- even just a limited support role to harder hitting units. I’d replace to +vs cav promotion with a barding promotion (like melee get) and then have both tier 1 promotions give an additional +5 vs cav in additon to their other bonuses.

What might also help balance units would be having some soft limits on unit numbers. One way to do this is to have each copy of a resource allow you to have x number of a particular unit: say 1 iron lets you maintain 4 swordsmen or 2 Knights (or combinations of that). Encampments could then boost those numbers, but you’d still need resources for healing, upgrading, and just building (up to your cap) in the normal way.

That would buff units that don’t require resources, because you’d need those units to bulk out your army. But. That’s a pretty big change and introduces a lot more micro. Not sure I’d really want that.

 

[Pietato]

Pike and Shot are not vulnerable to melee.

 

[Trav'ling Canuck]

And some people have commented that civ gets "anticav units" wrong. What one has to remember is that anticav units are not dependent on any resources in the game, so they has to have a drawback, to make the strategic resources and their units worth going for. The reason anticav and cavalry/melee cost the same is because the latter require resources.

In Civ 5, where the number of units you could field were based on the number of strategic resources you control, I would agree.

In the Civ 6 system, you either have the resource or you don't. If you do have it, it doesn't cost more to build the unit. Developing the resource also doesn't cost more, as you'd happily use the Builder charge to build the mine anyway (especially for flatland resources where the resource makes the tile much more valuable, except when it pops up right where you planned to put a district ). The only cost is the opportunity cost of not trading it to someone else, which is an economic cost if there's a willing buyer (who you're willing to trade a strategic resource to - many players wouldn't trade one to a neighbour).

So that just leaves the situation where you don't have the resource and you have to trade to get it. Then, indeed, you're paying a cost to get it which should be built into the cost of the unit. I'd guess that the frequency of this is low, based on personal experience, which is that if you don't have it yourself, the odds that an AI has 1 they will trade to you is slim.

 

[Flaxton]

Is there a mod for these suggestions, in summary I see:

• Anti cav should be cheaper to build and less maintenance
• Melee can stay where it is
• All cav, more to build and more maintenace, penalties against cities
• Ranged - as is.

Plus, the English special unit should be the Longbowman: 2 shots a turn
Thanks!

 

[Pietato]

Siege also need to join the game.

 

[Boris Gudenuf]

In Civ 5, where the number of units you could field were based on the number of strategic resources you control, I would agree.

Civ V had a mechanism where there was a Unit Cap based on your total population, and if you exceeded it your Production and, I think, your Population Increase dropped by X for every unit you exceeded the Cap (It's been almost two years since I fired up Civ V, so don't quote me on the exact penalties). This seems to me like a better proposition for limiting 'Unit Spam': if too much of your population is slogging around carrying weapons instead of producing goods or raising a family, your Productivity and Population should suffer. This would also allow us to Modify the Penalties by, say hiring mercenaries from City States, Barbarians, or other Civs (something I think the game needs in any case) or having Industrial/Modern Era 'Conscription' Policies/Civics that allow you to raise the Cap by drafting a larger percentage of the population: a real 'Total War' mechanism.

Siege also need to join the game.

Ah, yes, the Siege Elephant in the Room.
Siege Unit are tricky, because up to a point they are almost completely divorced from the regular battlefield, and after that point they are almost indistinguishable from regular units.

Quick and Dirty Guide to Siege Units: Battering Rams, Siege Towers, Catapults, and Bombards have almost no effect on regular military units/targets. Try dragging a Siege Tower across a battlefield, and you will inconvenience the enemy only because he will be laughing too hard to charge and massacre you for the first few minutes - same with a Battering Ram. Catapults, to my knowledge, were used on battlefields exactly twice, once by Phillip of Macedon and once by his son Alexander, and both times they were the smaller bolt-throwers which are only marginally siege weapons at all. The Roman Carroballista or cheirobalistra were cart-mounted 'catapults' that could be used on the battlefield, but again they were bolt-throwers designed to break up dense enemy formations or pick off their leaders at twice the range of an ordinary bow. Firing a .6 meter iron bolt at a stone wall might eventually chip it away if you have a decade or two to work at it, but it is not a serious 'siege engine'. The larger stone-throwers that could smash walls, like the Bombards that slammed 100-kilomgram stones into masonry, could break walls, but also took anywhere from 1/2 hour to several hours to load, aim and fire. They are effective only against units manning walls or sound asleep.

Basically, up to and including the Bombard, all the game's Siege units should really be Support Units, unable to operate alone or defend themselves in open combat.

And then, with the 'field cannon' ALL the smoothbore, muzzle-loading cannon are manufactured and operate in much the same way, and the only difference between 'siege guns' and 'field artillery/batteries' is the size: firing up to a 12-pound shot, it was a field cannon; firing 18 to 24-pound shot, it was a Siege Gun (the 24-pounder was, in fact, considered the optimum balance of weight of shot versus speed of firing and ability to move the gun). Siege Guns very rarely appeared on the open battlefield (Marlborough and his artillery commander, Brigadier Blood, at Ramillies, Frederick II at Leuthen) but it was remarked on as an aberration in every case.

So, after the Renaissance, the game to maintain its artificial division of all Units into rigid Classes, has to 'manufacture' an artificial Class of Siege Units. This is by no means impossible, we just have to 'enlarge' the minor differences between 'regular' Field Artillery and Heavy Artillery.

My suggestions, (from the background of both military history and 20 years as an artilleryman in the Atomic/Information Eras!)
Siege Units:
(Battering Ram)
(Siege Tower)
(Catapult)
(Bombard)
- These are in parenthesis because they also have Support Unit Characteristics: they have no effect on Combat Units and can 'stack' with regular Land Combat Units
Heavy Howitzer - which can use the current 'Artillery' graphic, because that actually shows a heavy howitzer of the 1915 - 1918 period.
Heavy Rocket - representing the large and long ranged but area effect truck/track-mounted individual rockets of the late Atomic Era - too inaccurate to do much to fast-moving tactical units, but able to do lots of damage to city structures.
The last two can also be used against Land (or Sea) Combat Units, but that's not their primary function. The Heavy Howitzer should have to 'set up' before firing, but the Heavy Rocket launch vehicles are designed to be able to 'shoot and scoot' so they do not have to Set Up first.

Ranged Units:
Slinger
Archer
Crossbowman
Field Cannon - which really appeared at the end of the 15th Century, so is a Renaissance Era weapon, not Industrial, and about 170 years after the first Bombards (dates: 1311 CE first Bombard, 1480 CE first 'true cannon', both in France)
Field Artillery - Early Atomic Era, perhaps at Tech: Advanced Ballistics (first 'electrical' computers were used to calculate Firing Tables for WWII-era artillery) - Graphic similar to Civ V's 105mm howitzer/Artillery Unit graphic
Rocket Artillery - the current Civ VI graphic shows a Multiple Launch Rocket System (MLRS) which, like the Soviet 'Guards Mortars' multiple rocket launchers, is designed for maximum fragmentation-anti personnel effect on a battlefield, not the maximum Blast effect needed to take down city structures.

- And in case you wondered, the Machine Gun, AT Crew and Modern AT units are ridiculous manifestations of historical Ignorance that have no place in a game the scale of Civ VI: all of them represent weapons that were added to other units as support weapons, and should be Promotions or, perhaps in a new category of unit modifications, Augmentations of existing Melee or Mounted units. The Modern Era and later units to make any sense have to be representing Division-sized historical units, and no one no where ever formed Machine Gun, Antitank Gun, or Antitank Rocket Divisions: The Soviet Army in WWII formed Antitank Artillery Brigades, but that is as close as anybody ever came, and those brigades were by doctrine to be used to support rifle or armored units, not as separate organizations.

 

[Boyan_Sun]

Maybe once I have a better understanding of upgrade costs (I've never looked into it)

Upgrade costs are: 1.5 * cost diff + 10, align to 5, round down.
Professional military policy card should be considered when upgrading, which broke the whole numerical system.
And if you also consider the +50%/+100% prod policies, many are different.

We change gold to production by 2:1(classic rate), the optimal costs can be:
Swordsman (36) = 40/1.5 prod + 85/2 gold = 46.7
Horseman (36) = 80/1.5 prod = 53.3
Knight (48) = 65/1.5 prod + 180/2 gold = 88.3
Pikeman (41) = 65/1.5 prod + 210/2 gold = 95.8
Samurai (45) = 160/1.5 prod = 106.7

 

[AlannaT]

Upgrade costs are: 1.5 * cost diff + 10, align to 5, round down.

It fits most examples I tried (Slinger -> Archer -> X-Bow, Heavy Chariot -> Knight) but until seconds ago I was sure that upgrading a warrior to a swordsman costs 80 gold and not 85. Am I mistaken or is the warrior the exception to the rule or is the formula only close but not exact?

We change gold to production by 2:1(classic rate), the optimal costs can be:
Swordsman (36) = 40/1.5 prod + 85/2 gold = 46.7

If anyone else has trouble following these equations (I certainly had) - here is it step by step:
The most cost efficient way to get a swordman is to build a warrior with Agoge. With 50% bonus you only need 40/1.5 = 26,67 production for the warrior. Then you upgrade that warrior with Professional Military for 0.5 * 85 gold = 42.5 gold. Because 1 gold is worth 0.5 production (wherever this "classic rate" comes from - the game usually works with a 4:1 ratio for purchases) this equal 21.25 production. So in the end you only invested 26.67 + 21.25 = 47.9 production instead of 90 if you had hardbuild it with no policy.
The differing 46.7 suggests that I was right above regarding upgrade costs of 80 instead of 85. Because if you change it to 80, this is the result: 40/1.5 prod + 80/2 gold = 26.67 prod + 40 gold = 26.67 prod + 20 prod = 46.7.

 

[Trav'ling Canuck]

Upgrade costs are: 1.5 * cost diff + 10, align to 5, round down.
Professional military policy card should be considered when upgrading, which broke the whole numerical system.
And if you also consider the +50%/+100% prod policies, many are different.

We change gold to production by 2:1(classic rate), the optimal costs can be:
Swordsman (36) = 40/1.5 prod + 85/2 gold = 46.7
Horseman (36) = 80/1.5 prod = 53.3
Knight (48) = 65/1.5 prod + 180/2 gold = 88.3
Pikeman (41) = 65/1.5 prod + 210/2 gold = 95.8
Samurai (45) = 160/1.5 prod = 106.7

Great structure for analysis, Boyan Sun. Note that I get 47.9 optimal production for Swordsman

This highlights the relative costliness of the unique units that cannot be upgraded into.

For example, the Khevsur (45) like the Samurai is now 160/1.5 = 106.7 optimal production, while the Musketman (55) is 79.2 optimal production (produce Warrior by keeping only 1 copy of Iron and Nitre and then upgrade; if you upgrade to Swordsman first, then optimal production is 87.9).

 

[Boyan_Sun]

Am I mistaken or is the warrior the exception to the rule or is the formula only close but not exact?

No, it is not exact, but error of this formula is no more than 5. Before disassembling, all formulas were conjectural.

 

[Pietato]

So in short:

- Make ranged siege useful
- Nerf knights
- Buff anti-cavalry
- Buff end-game ranged
- Enable upgrading into unique units
- Adjust production and upgrade costs