In the before times, I completed a PhD in Computer Science at the University of Waterloo's Games Institute where I studied the design of asymmetric co-operative games. At the heart of my work was the iterative design and development of my own medium-fidelity prototype game Beam Me 'Round, Scotty! (BMRS).
By developing my own prototype game, I could exert fine-grained control over it's mechanics, sound design, visual aesthetics, and interface. I used BMRS to conduct numerous player experience studies with real players and uncovered many interesting concepts and quirks of asymmetric co-operative play.
Some of those design ideas are mentioned briefly below.
It comes as no surprise that changing a player's perspective can drastically alter their experience. When designing asymmetric games however, different players can be presented with different perspectives relative to each other; affording unique gameplay possibilities. Having one player know or see something their teammates cannot is a powerful way to promote social connectedness.
While the goal of asymmetric play is typically to generate distinct experiences between players, the pace of play is one design element that inevitably bleeds or osmoses from one player's dynamics into the other.
Balancing the relative strength/effectiveness of different game mechanics is a critical aspect of game design in general and this is still very much true in the design of asymmetric co-operative games. If a player is significantly more powerful, effective, necessary, or involved than their teammates, it can lead to boredom not just for the underpowered player but for the entire group. Meaningful social interaction (rather than mastery) is often the true motive for co-operative play and having one player dominate the play experience due to imbalance can easily torpedo a game's fun.
Even when a game is designed to emphasize co-operation, some players will seek out ways to compete between each other. Depending on your design intent, this need not be a problem and there are numerous, subtle ways to alternately accommodate or discourage micro-competitions within a larger co-operative arc.
In Beam Me 'Round, Scotty!' for example, different variants of the healing ability alternately required cooperation, forbade cooperation, or left the decision to cooperate up to the players depending on whether who could trigger the healing ability (Kirk, Scotty, both), how the ability was targetted (beam, area of effect, injection), and which player the ability could effect (self, other, either). Team Fortress 2's "Medi Gun" is a well-know example of requiring co-operation.
A major challenge of asymmetric game design is creating distinct aesthetic experiences that appeal to different player tastes in the same game. In my PhD research, I adapted Hunicke et. al.'s "Mechanics, Dynamics, and Aesthetics" framework when building out and then bridging Beam Me 'Round, Scotty!'s multifaceted play experiences.