Symmetric vs. Asymmetric Design Techniques


I want to go over some of my thoughts on symmetric and asymmetric design. I have come to conceptualize these two terms very differently than their traditional use would otherwise prescribe. Symmetric design seeks to create a system of consistency and relativity within a game. It is sometimes viewed a bit more narrowly, understood as simply applying mimicry within a game’s design. For example, it might be limited in its application to providing equal starting conditions. I believe the concept can be applied in a far more expansive way with the goal of providing continuity within design. Whereas symmetric design provides structure, asymmetric design seeks to provide chaos. One of the consequence of chaos is unpredictability, which can have an extremely positive or negative affect on a player’s experience. It may force them to adapt and therefore create a dynamic challenge or it may result in endless and insurmountable degrees of frustration. This article seeks to explore these two concepts and provide some ideas on how to extend or manage the application of each.

Applying Symmetry to Design

Symmetric design in its application is a tool that allows us to create continuity within our design. Lets assume for the sake of discussion that we are working on a MOBA. We have 10 heroes and we want to make sure that no one hero is stronger than another. We would likely want a way of measuring the value of each hero relative to each other. To do this, we would create a formula which factors in each hero’s statistics and spits out a value. I have highlighted how to go about this approach in a previous article, located here. In addition to being able to compare hero value, this step provides a framework to better understand the impact of the changes we make to each hero.

Note: Our single consideration is to have each hero equal in value. We are not necessarily applying mimicking to our design, as each hero could potentially have their stats applied differently. Instead, what we are seeking to do is make each decision available to the player hold a similar weight.

In terms of process, this should be seen as the first step but not the last. The goal of the symmetric process is to create a baseline within the game. A baseline seeks to establish relatively equal values within each hero with the idea being that equal values equates to meaningful decisions. We also want to ensure that the baseline matches our intended gameplay.  If we are aiming for quick matches but instead get long drawn out matches due to each hero having too much health, than the baseline needs to be adjusted. By establishing a baseline, we should have a stronger understanding of how each variable impacts a heroes value in addition to knowing the game is fairly stable. This provides us with a strong foundation to begin iterations, this is typically where I would begin to add elements of asymmetry. Periodically, it is always a good idea to go back and recalculate the baseline to get an idea of how your changes have affected the game as a whole. It is also important to note that the symmetric process is not perfect, as a designer we must be able to make decisions outside of what the numbers would otherwise dictate. The introduction of asymmetry will often conflict with your baseline, trust your gut to make the right decisions.

Asymmetric Design and Balance

From the perspective of application, where symmetry seeks to create relativity within a game, asymmetry seeks to do the exact opposite. Which begs the question, why? I think the why really depends on how and to what the asymmetry is being applied. I view asymmetry as the application of chaos. Chaos has many consequences, it results in unpredictability, it breaks apart patterns, and it thereby forces changes on the player’s behaviour. The consequences of chaos effectively answer the why, it may not be the only way to achieve these consequences but it sure is an effective means. In other words, we might seek to apply asymmetry to force adaptation on the player. I believe the application of asymmetry must be very intentional and carefully groomed. It is an extremely difficult concept to execute as it not only require good intuition but our most expensive resource of all, time. The remainder of this article will focus on the grooming of asymmetry as it relates to balance.

If we have two distinct heroes, one with a high power single target nuke and the other with a stun, how can we determine whether these two abilities are balanced within the game? A symmetric analysis would seek to evaluate the two abilities by breaking them down into some relative value. Since gameplay is often very complex, this is not always achievable. In comparison, an asymmetric analysis would instead seek to understand the value of the gameplay produced. Once this value is defined, we can then determine whether it enhances or takes away from our design intent.

Note: An asymmetric analysis is inherently subjective, for this reason I strongly recommend ensuring that all assumptions as they relate to the game are questioned and answered. Ideally, we engage in the analysis with as many objective facts as possible.

I would start answering this question by first identifying what we know to be true. This is where we might use symmetric applications to better understand how these two abilities compare to one another within the game via a formula and its respective evaluation. The point of this first step is to gather as much information about the two abilities as possible so that we can better inform our analysis and future decisions. Finding similar abilities is always a good idea, as it provides some comparable reference points. If the game has been released and we have access to telemetry, we should also seek to gather as much information about player behaviour as it relates to these two abilities and any similar ones we have identified. Does each ability see equal usage relative to hero selection? Can we some how break down the efficiency of an ability? i.e. damage done or impact on a team fight. The type of telemetry we look for is really dependant on the game, what we should seek to establish is a stronger understanding of how the abilities affect player behaviour. This in turn should be compared against our own understanding of the game to get the clearest picture possible. For example, the high damage ability might see greater usage because its easier to execute than the stun. Perhaps the other ability is an AOE stun with a high cooldown which is reserved for large team fights. Therefore, players might use the stun ability less frequently but the impact of the ability in a clutch team fight is far greater than any other. Telemetry will not always answer the questions we pose and we should be prepared to accept this rather than force a false truth.

Pure data will never equate to our own understanding of the game. There is a great deal of discussion on data-driven design which I fundamentally disagree with largely because games are too complex to breakdown and understand with mere numbers. Instead, I advocate something more akin to data-informed design. Data can helps us see things outside of our narrow perspective, but it can rarely provide answers to our complex questions especially when they relate to player behaviour. What I am trying to highlight is a process to better understand player behaviour. Ask questions, question the answers, and be thorough. This step is very scientific in nature, we must assume nothing to be true and thereby question everything. Keeping an open mind helps us remove assumptions from our analysis and therefore broadens our perspective.

These first steps help us identify player behaviour and how that translates to gameplay. We now need to evaluate gameplay specific factors to understand whether the current behaviour we have observed is desirable. This requires us to identify whether our design goals are being maintained and whether that said design is the right direction to move forward in. The first question requires a very clear design vision. If we do not have a defined goal or direction, than the chances of us making the wrong decision is high. Let us assume we are making the next MOBA, if teamplay is a major design pillar then we need to make sure our decisions reflect on that stipulated goal. Its very difficult to evaluate the next question if decisions within the game do not match the goals, as it would not be clear whether the goal was wrong or the execution failed. More simply, lets assume I am seeking to build a treehouse which can hold the weight of four people at one time. If I build that treehouse with the wrong materials, resulting in the materials to fail and the treehouse to collapse when four people are within it, was the design of the treehouse at fault or the materials I chose to use?

Which leads us to the next question, is the outcome of the execution resulting in more interesting gameplay? We know that we want to hold four people in our treehouse, is this fun? Assuming we executed properly, i.e. used the correct materials, then there are many variables that need to be discussed in order to fully understand this question. I will go over a few points, but this is likely worthy of its own article. This is where our intuition as a designer and understanding of design principles comes into play. I typically start by evaluating counterplay, interplay, and teamplay but these variables are largely drawn from my experience as a RTS designer. Other genres might find other variables more relevant. Moving on, we may ask what the opposing players can do to react to the nuke or the stun? How do these abilities encourage teamplay or counterplay? How does the outcome of these abilities affect the players emotionally? Does it result in frustration, satisfaction, perhaps a feeling of accomplishment? Again, the type of questions we ask really depend on what type of game we are trying to make and are relative to our design pillars. Counterplay, interplay, and teamplay are three important questions which I believe apply in almost all multiplayer game types. If any of these questions result in negative consequences, we need to then weigh the benefits vs. negatives.

In Company of Heroes 2, we recently made a number of major modifications to squad behaviour. These changes improved squad fidelity but also had the consequence of creating more punishing outcomes when the player positioned a squad poorly. This example has a number of consequences on balance and competitive play, but what I want to highlight is the reasoning behind our decision to make this change so to better understand the processes discussed above. A major design pillar within the game is tactical play; hence, our intent. This change noted above increases the emphasis of the players tactical decisions by strengthening the consequences of the outcome. Either the player is strongly rewarded for a proper position, or they leave themselves vulnerable for the opposing player to punish them for poor positioning. Yes they may get frustrated at times when a squad is moved into a bad position, but this frustration is a direct result of their actions which they now have more control over. This decision came down to weighing the pros and cons of the change and comparing them to our design goals. The design goal is to create a more tactical environment. The change increases the players control over their units thereby enhancing their capacity to maneuver units tactically. The consequences create a loop which either reward or punish the player depending on their decision, which again reinforces the tactical play we are seeking to create.

Note: Sometimes we build a treehouse with the right materials but realize its not fun to be in. In the example above, tactical play might not be interesting or fun to the player. Once we determine we have executed on our design intent effectively, we should be prepared to determine whether our intent increased the players enjoyment of the game. If it did not, then we may need to redefine our intent. 


Symmetric design seeks to create a system of consistency and relativity within a game. Its best applied early and often to create a baseline that guides our decisions as designers. In comparison, asymmetric design seeks to create chaos and its associated unpredictability. From an application point of view, it seeks to understand the value of the gameplay produced. Once this value is defined, it allows us to determine whether it enhances or takes away from our intended design.

  1. Anon Reply

    “Yes they may get frustrated at times when a squad is moved into a bad position, but this frustration is a direct result of their actions which they now have more control over.”

    Unfortunately this is a lie. The player has no control over how the squad choses to form itself – that is a fundamental of CoH, and Essense engine games in general. Pretending the player has control is contradictory to the fundamentals of your own production vision.

    The player is at the mercy of a very simple algorithm which arranges entities within a squad – sometimes they chose to blob, sometimes they chose to spread out. Recently the issue has become so bad that squads do not even respect basic player intentions, such as “Remain on a the border of the capping region” and will instead run a few feet to cover, out of the region.

    If you wish to balance the game around the notion that a player has the tools to impart his intention on the squads he commands, you must provide those tools first. Otherwise, balance must be made around the fundamental fact that players do not have controls over their squads. “Move to general World region” and “Maybe do this ability based on 10 hidden parameters” is not sufficient for the former.

    The addition of target prioritization to AT weapons should be the primary study of empowering player with strategic decision, not the removal of input delay and increasing cover-snap range (which ironically is the complete opposite).

    • pqumsieh Reply

      Fair points and great feedback.

  2. Philip Reply

    Quote Anon ((The player is at the mercy of a very simple algorithm which arranges entities within a squad – sometimes they chose to blob, sometimes they chose to spread out. Recently the issue has become so bad that squads do not even respect basic player intentions, such as “Remain on a the border of the capping region” and will instead run a few feet to cover, out of the region.))
    So true.
    If units are at “Hold position”, they don’t evade attacks.
    If units are at “Roam – even limited” a few may follow while others stay behind, thus they get themselves trapped.

    Because of this, some players prefer towers over units.

    Towers might be stronger, compared 1by1 to units for their immobility, but they can’t make use of the square-law.
    Unless there is an all-covering expandable shield wall which inevitable forces the weaker player to play at pressure and higher expenses, this leads toward either (or both):
    * Long-Range Towers rushed in front of enemy towers as both attack force and later for defence.
    * Newbies playing horrible in team-games, because they favour the wrong strategy (which even worse may work against AI-training enemies and other newbies).

    This article is assuming that players are making good choices.

    If players learn to play “wrong” and other players adjust to that by playing “differently wrong”, it might be challenging and still different from what the developers expected.

    In one MMO, I saw that the mage has a Level 60 spell which deals 1400 + 5 damage per intelligence point.
    Since you can only get 5 points per level on that, you can only ever get 25*60= 1500 bonus damage.

    But with 1/2 points on Luck, you can get 200% damage critical hits and you can also use some points for Wisdom (mana regeneration).

    Surprisingly many players laughed at first at my GlüWei Mage at first – they did not do their math.
    But later it owned the battlefield for one update-cycle.

Leave a Reply


captcha *