How to balance asymmetric civilizations

We know by now that AoE4 will be asymmetric. I’m not sure if several civilizations can have similar designs or each one of them will be different.
This makes balancing a huge challenge. I have played AoE2 and SC2 a lot, and I think the solutions are as follows:

1-More armor types or unit tags.
In AoE2, units have armor and pierce armor, but they also secretly have something like infantry/archer/cavalry/cav archer/siege/camel/elephant/ship/building/stone building armors, which allow units/buildings to take or negate bonus damages. This means you can generally find a good counter to a certain unit because every civ has access to a relatively similar unit pool. But it also means that for a civ that lacks a certain unit (or the said unit is not practically viable by lacking a specific upgrade(s) ), some matchup will be unfavorable.

SC2 does essentially the same by giving units tags: bio/psyonic/mechanical/light/armored/massive etc. But it lacks the feature where you can give a unit resilience against some bonus, which you could do in AoE2 like with cataphracts and condottieros. In this sense, AoE2 is better.

SC2 has 3 asymmetric factions, and balancing is a nightmare (not saying they did a bad job, just that it’s difficult to pull off). For instance, if you hope to see more mutalisks in ZvT by decreasing Thor’s anti-light bonus, suddenly, phoenix becomes OP in PvT.

If AoE4 is to have more such factions, balancing should be a serious issue. One way of doing it is to invent more armor types, so that increasing the damage from unit A to unit B will not affect the damage from unit A to unit C-Z, or at least affect as few others as possible. Kinda like how eagle warriors have their own armor type. Except now, we can perhaps give mangudais their own armor type as well.

Of course, computation wise, we don’t want to assign a new armor type to each unit type, but having more than the 11 in AoE2 (as far as I know) is desirable.

2-More maps
Different civs prefer different maps. Making more maps so that all civs can find their maps to excel on is helpful, because now you don’t have to make sure all civs are on equal footing on arabia. This also ensures that players won’t just play one civ. In SC2, a lot of players (especially competitive ones) only play one race, because each race is supposed to be accessible on all maps, and learning a new race is exhaustive. This creates a toxic player base that constantly whines about balancing issues, e.g. you can always find some terran players bickering about balance issues no matter the win rate.

These are the two ideas that I can come up with. I think more specific armor types is definitely necessary if we ever hope to balance so many factions.

I wouldn’t consider using the sc2 method because they have pigeon holed themselves into a mess.

TvZ is terran favored at the highest level.
PvT is protoss favored at the highest level.
PvZ is zerg favored at the highest level.
And they can’t fix any of it without destroying a civ in another matchup.

Nerf T because tvz? Great. Worse in tvp.
Nerf z because pvz? Great. Worse in tvz.
Nerf p because of pvt? Great. Worse in pvz

That’s why I didn’t suggest the SC2 model. But SC2 does have some features that enables easier balancing.
Zerg has no mechanical units, meaning if we buff viking anti-mechanical bonus, it only affects TvP.
Only Protoss has shields, meaning if we buff widow mine anti-shield bonus, it again only affects TvP.

If AoE4 can have units/buildings with similar attributes, e.g. stone building bonus might not affect nomadic civs, anti-shepherd/farmer bonus when raiding nomadic/agricultural civs’ economies. The idea is, as long as the CPU can handle it, the more the better, the more independent they are from each other, the better.

There is a limit how much factions can be asymmetric and still playable.

If we simply take SC1 vs SC2, one thing is clear, in SC2 the factions are more similar than in SC1.

example SC1 Zerg: Guardian flying siege unit. its a unique unit only to zerg
SC2 protoss Tempest is very odly bigger version of that unit
SC2 Zerg Brood Lord just has different attack animation “units instead of ball”
SC2 Terran Liberator in Defender Mode is same vs ground units only

That’s why I think the team we have should rather focus on playability of the game,
not some wacky weird mechanics.

That sound reasonable first, but on second thought, balance is a thing of whole picture.
A unit should always be in the food chain or otherwise it does break the balance.
So can you maybe make a chart how your idea is supposed to work?

Or we might end with mangudais, not have a unit to counter it.

Mangudai is the unique unit of the Mongols, cavalry archer with an attack bonus against siege weapons. Problem is, except Skirmishers there is no real good unit to counter him.

The idea of this post is not to invent a system that creates a balanced game in the get-go, but rather invent a system that enables easier balance fixes. Mangudai is OP, because it has a high dps, high mobility, and counters siege weapons. If we give some civs one tech that grants anti-anti-siege armor, or their camels additional anti-mangudai armor, then these civs can hard counter the Mongels.
However, if we don’t have unit tags such as anti-anti-siege armor or anti-mangudai armor, we can’t have these simple fixes. The point is that we create many armor types and their corresponding attack types, while we may not use them upon game release (because we can’t foresee all the possible ways of abusing one unit to make it OP), we can reserve them for balance purposes afterwards.

And how are people supposed to understand those metrics? I understand where you’re heading, but by your suggestion people will lose the overview what the system actually does and is supposed to do.

It’s easier to understand from unit to unit how it supposed to work. Example:
I have no idea why Siege Tank does loose vs Immortal, I just know he does.
Than you can make conclusions what does lead to it.

Good point. Too many tags does tend to confuse the big picture, and making tags too specific might be a bad idea.
I still think we need more than what we had in AoE2 though.

Another idea I have is to allow unit prices that are not multiples of 5. In AoE2, without civ bonuses, unit cost for a certain type of resource is usually 5N where N is an integer. Allowing a unit to cost 42 food, instead of having to choose between 40 and 45 should increase the balance flexibility. I don’t think this will put too much additional burden on the player.

I think the first thing you need to do regarding balancing a game, is to not be so obsessed about game balancing.

Lets be honest, what in real life warefare is balanced? Having military or technological superiority is actually what differentiates powerful nations from less powerful nations. Nations specifically invest money to make sure their units are more capable than the enemy.

When you design a game with unique faction play styles, what balances out the game is that there are a wider range of strategies available to succeed or lose. Its more dependent on the players skill and knowledge of the strengths and weaknesses of a particular civ.

It also results in more varied games, you have to be able to adapt based on the specific game rather than just spamming the same strategy in every game.

For example, one nation might have strong Age 1 and 2 units - that means probably best strategy is a quick military boom and rush.

Then you might have civ with strong Age 4 units, in which case the best strategy is probably going to be a economy and defence boom, with plans for a age 4 overload attack.

1 Like

well,it isn’t same as sc2 civ design
accroding to Adam Isgreen.
in aoe4,there is a civ spectrum:
most familiar to most different civ
most familiar civ is english,classic aoe2 gameplay
most different civ is mongols,a very different gameplay

Age of Empires Online has to date the best example of balanced, fairly asymmetric civs. Each of AoEO’s six civs share an awful lot in common, feel naturally related, and are easy enough for anyone to pick up and play. But they are each fairly complex and offer a large number of unique features. The civ design in AoEO is as perfect as I have ever seen in this franchise.

Here is way more information on this topic than anyone here wants:

Every civilization in AoEO has 18 buildings except the Norse (16). Looking deeper, we can see that every building falls into one of five categories: Common Buildings, Improved Buildings, Advanced Buildings, Custom Buildings, and Unique Buildings.


Common Buildings : The game files tell us that there are 13 buildings that every civilization in Age of Empires Online shall have by default, and that they each shall appear at certain ages. Age 1: Town Center, Storehouse, House, Watch Post, Dock, Wall, and Barracks; Age 2: Market, Farm, Guard Tower, Armory; Age 3: Fortress; Age 4: Wonder.

We call these the Common Buildings, and, in fact, that’s what the game files call them, too. In almost all cases, every Common Building in every civ is identical. However, there are two exceptions, Improved Buildings and Advanced Buildings.


Improved Buildings : Sometimes a Common Building in a civ has a special technology or other attribute that improves it a little. Examples of these Improved Buildings are the Persian Storehouse (it heals), the Celtic house (it allows you to Age up), the Babylonian clay Wall (cheaper in Age 1), and the Norse Fortress (Burning Pitch). Arguably the Egyptian Armory is an Improved Building (it has cheaper, combined armor techs) as is the Greek Guard Tower (the Academy gives it Murder Holes and higher health and damage). By our count, every civ so far has one Improved Building.


Advanced Buildings : Meanwhile, sometimes a Common Building is replaced by a completely advanced version of that building that is still available in the same age. Examples of Advanced Buildings are the Norse Longhouse (House) and Outpost (Watch Post) and the Babylonian Ox Cart (Storehouse), which is actually a unit that replaces the role of a building).

Other than the 13 Common Buildings, each of the Original 6 civs have 5 other Custom Buildings (except the Norse, who inexplicably have only 3). In almost all cases, every Custom Building is some form of the following: Archery Range, Stables, Siege Workshop, Temple, and Academy (a building with only techs).

Frequently these buildings use civ-specific names (the Sacred Grove and Great Hall are essentially Temples, and the Bard Hall serves as an Academy). Custom Buildings often can become available in either Age 2 or 3 depending on the civ.


Unique Buildings : A few Custom Buildings stand out so much that we call them Unique Buildings. Sometimes, Unique Buildings serve as some advanced form of a Common Building that does not actually replace it, such as the Persian Immortal Camp (Barracks), Celtic Gold Mines (Storehouse), and Babylonian Gardens (Farm). Other times, Unique Buildings combine the features of two or three buildings into one (the War Academy trains priests, has tech upgrades, and houses unique toggle techs and the Ziggurat trains priests, has tech upgrades, and ages up). There are five Unique Buildings in the game, and they are available anywhere from Age 1 through Age 3.


Note that no two civs occupy the same field. Instead, each new civ focuses on different buildings. The civs dance around the chart filling holes and never quite overlapping.

So each civ has about three buildings that do new things (plus a couple that just have unique names).

It also shows that civs got a bit more unique over time. The Greeks (our 1st civ) act like a baseline that every other civ jump off from in different directions. This makes some sense, too. Nobody wants a new civ to feel boring.

Just like with buildings, we started our unit design by first analyzing what everyone could agree on: the designs of the Original 6 Civs. In the game files, every unit is categorized as either a Common Unit, a Military Unit, a Unique Unit, or a Scout. This can be pretty dense reading, so please buckle up.


Common Units : Just like Common Buildings, every civilization in AoEO shares the same Common Units. There are four Common Units: Villagers, Caravans, Fishing Boats, and Merchant Transport Ships. Though these units can have small statistical differences in different Civs, to date, no Civ has anything particularly unique with any of them.


Military Units : Almost every other unit in the game is a Military Unit, and so most of our analysis focuses on these units. They include all Land Military , Siege , Priests , and Ships .


Unique Units : In rare situations, a Civ can have an extra unit that is neither a Military Unit nor one of the Common Units. There are currently three in the game: the Celtic Sacred Deer , the Babylonian Ox Cart , and the Norse Raven .

The total number of units varies among the Civs and ranges from 19 (Celts) to 24 (Norse). However, the four Common Units and Scouts are essentially identical in every Civ.

The game files describe Military Units in three ways: their Class , their Role , and their Attack Type .


Classes : There are six different Classes of Military Units in Age of Empires Online: Infantry, Archer, Cavalry, Siege, Priest, and Ship. (You will see below that our analysis adds a 7th Class for those three Unique Units, as well.)


Roles : In addition to being a member of a single Class, each unit also has one of 13 different Roles:

  • Basic (e.g. Sparabara, Bowman, Asabara, Trireme)
  • Anti-Infantry (e.g. Hypaspist, Axeman)
  • Anti-Archer (e.g. Slingers, Carpentom)
  • Anti-Cavalry (e.g. Spearman, Camel Riders, Podromos)
  • Ranged (e.g. Mounted Archer)
  • Heavy (e.g. Cataphract, Champion, Sapper, Berserker, Chief)
  • Ram (e.g. Ram and Siege Tower)
  • RangedSiege (e.g. Catapults, Ballista, War Wagon, Catapult Trireme)
  • Trebuchet (e.g. Palintonon, Stone Thrower)
  • HealOnly (e.g. Priestess of Ra, Druid)
  • Unique (e.g. Priest of Set, Unique Units)
  • Convert (e.g. Priest of Ptah, Magus)
  • Anti-Ship (e.g. Ram Ship and Fire Ship)

Note: Certainly any unit can be balanced to have more than one role (an Anti-Cavalry Camel Rider may also be effective against Archers, for example), but for Unit design purposes, it is helpful to see that the original Devs creates the Civs with a specific Role in mind for each Unit.

Attack Type : Every unit is assigned one of four different Attack Types: Melee, Ranged, Both (such as Immortals), or None (such as most Priests).

Note: The “None” attack type refers to being able to attack enemy military units and does not include Priests’ ability to attack and kill huntable animals.

Overall, the 7 Classes, 13 Roles, and 4 Attack Types combine for many dozens of possible permutations of Units. So in order to get a handle on how to choose among them, we summarized this information for each Civ and spotted patterns. We don’t have space to go through everything, but let’s take the Greeks as an example since they are considered by most the “baseline” civilization.



So for example, on average, each civ has 7 ranged units, but one civ has just five (Celts) and another civ has as many as nine (Greeks).

For instance, in Age 1, we see there is a rebuttable presumption that a new civ have a single Age 1 Spearman. Though we can certainly go a different direction (like the Persian Sparabara), but if we do so, then that civ would almost certainly need to have at least one Anti-Cavalry unit no later than Age 2. Since most civs have an Age 2 Cav unit, they’d need a counter.

Similarly, if we give a unit an Attack Type of Both (melee and ranged), in order to keep some balance, we likely want to limit the civ to no more than one Age 3 unit (which is exactly what we are doing with the Centurion).

Finally, we also observed a unit design rule whereby if a building trains new units (or technologies) in more than one age, then it trains them in successive ages without ever skipping an age. This is a subtle but super important point. In other words, if a building trains a new unit in Age 2, then it would never skip Age 3 and have the next new unit available only in Age 4. So if the Barracks has an Age 3 Unit, you know it also has an Age 2 Unit. And if a Dock has an Age 4 Ship, it always has an Age 3 Ship, etc. Now this rule is less about overall unit choice and balancing and more about creating a tight, intuitive civ design. And it’s much harder than it sounds to pull off. It’s really tempting to ignore this rule because it creates a hurdle that doesn’t at first seem all that important. I mean, what’s the harm really if a Stable has an Age 2 and Age 4 Unit but skips Age 3? But if you skip an age, it really would feel like a sloppy tech tree that was just thrown together. Unless at some age the building completely stops training new units or techs (such as the Celt Archery Range, which has no Age 4 unit), players intuitively expect a new unit in each successive age. And if this wasn’t really a design rule, then surely some building in one of the Original 6 Civs would have ignored it. Yet every single building follows this rule. And the civs are stronger and more intuitive because of it. Again, breaking this rule won’t necessarily make an unbalanced civ, but it sure would make a disorganized, unpredictable one that lacks the same quality and artistic flourish of the original civs. So we had to follow this rule whether it was convenient or not. And it gave us absolute fits.

This is the kind of hidden detail that tends to go unnoticed until you try to create your own civ and realize something feels terribly wrong.

Building a civ is so much more than coming up with fun ideas and snapping your fingers. Everything has to fit just so. If we were to shift one unit from one Age to another, it would throw off the balance in a different area, requiring us to shift something else. In this way, one little change echoes across the entire tech tree causing all sorts of unexpected headaches. At times it felt like they were using tweezers to move sand from one side of a scale to the other.


Well I think what is there more important, is a clear visual difference and visual feedback.

Age of Empire 1 and Online did not do a good job there. As units are dressed similar. And let’s say if it gets RTS hectic for many, as you lose the overview who is who and who is where.
Here example as all units are simply lazy and outdated dressed in blue

A professional designed even in Middle age, does there do something much more clearer.
here you can instant tell on closes alone what weapon do units have

here you can see a very good example how to tell units apart,

archer is dressed very simple red shirt, blue jeans, spearman has 2 colours shirt and grey jeans
you can instant see who is archer and who is spearman. In AoE1/online they look like same unit.

Another things are hats, lol with a simple hat you can already see how different units are, armour can look different as one of the units has grey leather armour and its visible different from Iron Armour and closes. Also a simple trick is to put symbols on the shields, units can for example even carry one on the back, it does not have to be obvious a sword or bow/arrow, but some kind of clear symbol.
There are also things like colours, look how similar are horse units by Online, in AOE1 at least archer horse is black, sword cavalry horse is white and improved units has a clear golden armour. The shields can have different shapes, to better tell units apart. There are many tools to give player clear visual feedback.

The big problem is to give by different factions and different game teams, clear visuals who units are and to who they belong.

here is an example how same unit types could be told apart by different factions,
like give them different gear look, shield shape, helmets , colours pattern .

1 Like

I agree that visual cues are essential and enjoyed your post very much. However, I disagree that AoE1 and AoEO have problems in this area. AoEO communicates through its units quite well.

One way of creating unit diversity without creating a risk of confusion for new players is establishing a universal general look for different unit types, then creating units that adhere to that general style while still looking different from each other. here’s an example:


Even though all of these Units are completely different Units from different Civs with different names, icons, sounds and everything, one look and you can tell that they are anti-infantry infantry units. How? they all wield 1-handed weapons and a shield.

Here’s a second example:


These are two completely different heavy infantry with area damage - the Norse Berserker (on the right) and the Celtic Champion (middle).

  1. They each are much bigger than regular units (compare them to the tiny Celtic Spearman on the left)
  2. They each wield two-handed weapons and swing them in an arc in front of them when they attack.


1 Like

Clearly a lot of work goes into games, the main question is as always, “does the customer see the final product as something to spend money and most importantly time on it”
As customers do have already several products, they have expectations.

Here as example a 10 year older game, again 10 year older game than AoE Online.

If we talk about combat animation, just look how even the simple Spearman does bleed in fight and has at least 4 different attack animations, stab with spear, attack with spear over the head, kick with leg, kick with 2 legs while using spear to hold. Battle Realms despite be 1 decade older, just looks light years ahead.

Just look how different each unit looks, between even same faction. While by AoE Online units identity goes down in masses, by Battle Realms based on closes and animation you can clearly identify each unit from another.

AoE online on the other hand, had compared to competition, just no visible redeeming qualities.
Maybe because they had to add skins by AoE Online, that’s why unit look so lame, anyway the final product just looked inferior on the market. We had games out there with clearly superior looks, better closes variety and more animations.

By Battle Realms, good base build RTS by the way, you can see after some minutes how much effort went into it. Units have this alive feeling, not animated toys. That’s why you do feel much more engaged for the game and have way better emergence into the situation. Just look how heavily damaged units do move. You can see they were hurt and have pain.

By AoE online I lost after some hours my interest, Battle Realms is a game I do play even today.

1 Like

this is a game though, it isn’t fun if you’re stuck at a huge disadvantage.

By AoE you get military or technological superiority by better playing.
That’s why we need a deep and complex base build.

I am offering AoEO as an example of both excellent asymmetric civ design and of clear visual language to help players identify units of otherwise asymmetric civs. Whether or not a different game from 10 years earlier had more visible redeeming qualities than AoEO does not change that AoEO has excellent asymmetric civ design and a clear visual language.

I am not here to argue the merits of Battle Realms or to compare it with AoEO or to otherwise argue about which games any of us enjoy more than others. That is wonderful that Battle Realms had animations that show its units to be in pain. But that has absolutely nothing to do with asymmetric civilizations.


I might be getting a bit off topic here but it’s been quite some time since I last opened the celeste project and seeing your posted images now guys, I just had to comment on how sleek AoEO’s visuals have actually always looked like and how well have they stood 9 years later.

You can say that they are cartoony, that they had paid skins and everything, that’s all ok, outright criticism, but you also have to credit them the vividness and longevity that their art style embodies. If you hate their type of art, then you’ll never like them but they really are not lame at all. It’s a very detailed AAA content.

Considering how AoEO was Microsoft’s last major entry in the franchise almost a decade ago, and how it was viewed as a failure, this kind of images just build up the hype for me on what we can actually expect from AoEIV today, when all lessons are learnt, when today’s computational power is far ahead than 10 years ago and in all the seriousness that the project seems to be treated with by the team.
It has to be something great, at the very least.

Yeah Age of Empires: Online, or Project Celeste has really aged well. More on topic, it’s also a good example of how to balance asymmetric civilizations. The current balance team has done an amazing job if you ask me.

1 Like

“animations” can contribute apparently quite a lot to the game.
Especially as they can do make asymmetric visible.

Just compare it to level of detail from same year game Dawn of War 2.
you get instant the feel how much different factions and roles are.

Considering how AoEO was Microsoft’s entry in the franchise,
if they are not careful, AoE4 might suffer same fate.

1 Like

I like the idea of emphasizing on different animations as an alternate or supplemental way to communicate the asymmetry between the civs. Combat would also be more appealing that way as well. After all, Germanic irregulars would swing their swords in a different way than, say, a tactical Roman unit. Or an English knight should charge differently than a Mongol horseman.

I think in the gameplay trailer, English axemen and Mongol swordsmen had visible differences in their fighting animations, it was a very short glimpse of course but I didn’t just see two guys banging their weapons on the head of each other or swinging them indifferently in the general direction of their opponent.

Also, I believe that MS is as much aware and sensitive of their past failures with AoEO as Relic is with their own ones with DoW3. AoE4 seems to be getting a fresh and ambitious approach after all this experience and knowledge of all sorts of past RTS titles has already been accumulated. Let’s wait a bit more to see what else is in store and have a better and more thorough look of the game, there is no much point in worrying too much and repeating ourselves.
Especially balancing, in the end it can always be sorted out with patches if needed, and the team as mentioned above is already doing a great job with AoEII:DE