In Cosmic Crew, different co-op controllers keep a starship intact

The 2023 Game Developers Conference will once again feature Alt.Ctrl.GDC, an exhibition dedicated to games that use alternative control schemes and interactions in new, exciting, and clever ways. Ahead of GDC 2023, Game Developer will be talking to the developers of each of the games that have been selected for the showcase.

Cosmic Crew sees a space-piloting team work together: a gunner and mechanic run different controller stations, with the gunner spotting foes and shooting them while the mechanic repairs the ship and loads ammo.

The team at Space Cow Games spoke with Game Developer about how iterating on their design for a space combat game would lead to splitting the game into two unique controllers and roles, the thoughts that went into creating two control schemes that make the players communicate in sometimes-hilarious ways, and how it sometimes just feels good to have a controller you're SUPPOSED to beat on.

What’s your name, and what was your role on this project?

Qiu: My name’s Andrew Qiu and I’m the project lead responsible for design, programming, hardware development, and documentation.

Chambers: My name is Sarah Chambers and I’m another designer of Cosmic Crew. I’ve also been responsible for organizing playtesting sessions.

Sun: My name is Suowei Sun and my role in this group is as an artist. I'm mainly responsible for game art design, including 2D art sprites, UI assets, and VFX.

How do you describe your innovative controller to someone who’s completely unfamiliar with it?

Qiu: The goal of the controller is to emulate what operating a space fighter with a friend would feel like. On player one’s side, you have a physical gun turret that can be aimed and fired at enemies. Player two has metal plates that must be hit with a hammer to repair damage. In addition, they also have a slot to insert different types of armor-piercing ammunition.

Because the Gunner is the only one with a working camera, they must identify each enemy’s weakness and direct the Mechanic on which type of ammunition to load.

Chambers: These controllers both emulate the feeling of operating a "real" space fighter, immersing the player in the game in a way that wouldn’t be possible with traditional controls.

Sun: Our game requires interplay and communication between two players. And possibly shouting.

What's your background in making games?

Qiu: I’m currently a third-year student studying game design at Sheridan College. I almost exclusively work with art, programming, and narrative, so it was a super refreshing experience to try developing a physical controller.

Chambers: I am a second year in game design at Sheridan College. My main strengths are design and narrative, as well as managing projects. Making an alternative controller game is a new experience for me, and it has been very exciting!

Sun: I am a third year student of the game design program at Sheridan College. My main areas of work are 2D art, 3D modeling, and level design. Making an alternative controller game gave me the feeling of making a game in a game room. I've never had this experience of making a physical controller before! It was very refreshing and interesting for me.

What development tools did you use to build Cosmic Crew?

Qiu: The game was programmed in C# Unity, with game art being created in Aseprite, Photoshop, and Clip Studio Paint.

What physical materials did you use to make it?

Qiu: An Arduino Uno powers both sets of controllers. The controller itself is mainly made of wood and metal for durability. The electrical contacts on the pieces (for example, the metal plates and ammo magazines) were made from either copper sheets or metal duct. In retrospect, I wish we went with the metal duct for everything—large copper sheets turned out to be horrifically difficult to solder!

What inspired the creation of Cosmic Crew? What appealed to you about having two stations communicating with each other?

Qiu: Cosmic Crew came out of a week-long alternate controller game jam at Sheridan College. At the time, our group’s goal was to create something with a simple scope given the limited development time. So, we decided on a basic top-down shooter that was controlled by a physical aiming device. However, we quickly realized that this would be too basic; we’d essentially just be making one of those on-rail arcade shooter light guns, after all. So, we tried brainstorming more ideas to go along with this mechanic.

What else could we have the player do? Well, they’re on a spaceship, right? How about repairing a ship? They could have a big hammer which they could use to fix damage with. Neat idea! Oh, what if they also had a control panel they could fiddle around with? Maybe they’d need to enter certain keycodes to initiate certain repair protocols. Cool! Wait… that sounds like an awful lot for one player to manage. Hm… how about making it a cooperative game? One player can shoot, and the other can repair!

While we felt this a solid idea, we still felt that the two roles were rather disconnected from each other. How could we get the two players to feel more cohesive; in other words, how do we get the two roles to communicate? To this end, games like Keep Talking and Nobody Explodes and SpaceTeam were pretty big inspirations. These are cooperative multiplayer games that place a heavy emphasis on overcoming communication gaps in order to win. Keep Talking and Nobody Explodes, for example, features one player as an information provider, and the other as an agent that acts on this information. Both parties must work together with their unique abilities in order to defuse the bomb.

This dynamic in communication gap roles is a super interesting style of cooperative games, and one that I really hope was effectively captured in the Gunner and Mechanic roles. In the game, the Gunner acts as the information provider, relaying information about enemy positions and weaknesses to the Mechanic, who must then act accordingly either by switching ammunition type, prioritizing which ship part to repair, or activating special defensive/offensive powerups.

Chambers: Our group was inspired by the traditional space arcade games many of us grew up playing. Being a pilot of a space shooter is a fantasy every kid has had, and we wanted to bring that to life! Although we had seen some modern adaptations of these space arcade games, we realized we hadn’t seen many with physical controllers—and almost none with two players. We were also inspired by the communicative nature of games like Keep Talking and Nobody Explodes, and wanted to combine that with the classic theme of a space shooter to create a unique experience we hadn’t seen before.

Communication in video games always ends up being somewhat chaotic, but it is often the player’s favorite part of the game. It can also be hilarious to watch, and we wanted our game to be just as interesting to watch as it is to play for yourself. Communication is a must in Cosmic Crew because players have two different screens, meaning information from one screen has to be relayed to the other player who can’t see it. For example, ships can only be damaged by their corresponding-colored ammo. The player in control of the ammo can’t see what color ships are on the screen, meaning the player controlling the turret has to yell out what they’re seeing. This leads to some frantic gameplay reminiscent of a real space battle!

What ideas went into the two controllers for the game: the turret and the ammo/repair station? How did you create them so they added unique, interesting elements to the game?

Qiu: The original plan was to create a modular control scheme similar to Keep Talking and Nobody Explodes. In this design, there were so many more things to manage: there were stations to quick-change overheated barrels, computer chip stations to change ship configurations, keycode pads, and much more.

The goal there was to create even more hectic gameplay, with players struggling to remember what the controls even did. In the end, though, we realized that all these different mechanics would probably be incredibly overwhelming for the average player. So, in our final design, we tried taking all of the most unique and interesting modules and consolidated them into a more intuitive control scheme. The gunner lever and fire control were combined into a simple aiming turret, and the most interesting elements of the other modules (that is, having to perform instructions based off limited information) were combined into the reloading station.

The hammer was kept simply because it was very fun to hit things with.

What thoughts went into designing a controller that you beat with a hammer? How did you make it durable enough to take a beating? How did you balance it so it didn't exhaust and injure the person hammering?

Qiu: In the version [with] one controller, we found that the most entertaining control was the hammer and plate. Unsurprisingly, all the hammering and smashing meant that it was prone to failure, with players managing to break it multiple times in a single session. In the second iteration, we made the plate from sheet metal, nailed to solid wooden planks. I’d like to see anybody destroy it now!

As for the hammer, I was thinking of using an actual mallet, or some sort of mini sledge to make the repair experience feel all the more authentic. There were concerns of people tiring themselves out from swinging a big hunk of metal, though; or worse, accidentally throwing the hammer at others (think Wiimote mishaps!). In the end, we ended up carving our own cute little wooden hammer. You can smash with it, toss it around all you want—no problems with weight or durability here.

What drew you to make this an experience for two players, rather than an extremely hectic experience for a single player?

Qiu: We thought that forcing a player to learn so many different mechanics by themselves was way too much—better to split the roles up and teach each player the game piecemeal. There’s also just that much more replayability that way; players can play the game from two different perspectives and find the role that they most enjoy.

The game is also made more accessible with two roles rather than one. The Mechanic serves as a more accessible role as those with a slow reaction time can still contribute to their team’s success. Plus, games are always more fun with friends!

Sun: Division of labor is an important condition essential to teamwork. Two players playing separate roles can better recreate the scene when the crew cooperates to control the ship in battle. It will be more intense than a single player to play the game as there's more need for cooperation and communication.

The game features an upgrade path between waves. How did you design these so they enhanced the game without taking away from what the unique controllers brought to it?

Qiu: We designed each upgrade to relate directly back to the way that the controllers operate. For example, gun upgrades enhance the turret’s capabilities in some way—rapid-fire machine guns, laser cannons, or railguns that can destroy any color ship. However, shooting with these upgraded guns results in the gun being damaged more quickly, which meant the Mechanic needed to make many more repairs.

Other upgrades seek to enhance the communication aspect of the game. For example, shield powerups can protect the player ship from damage for a certain amount of time. However, they can only protect against one color (orange shields deflect orange bullets and blue shields deflect blue bullets) and are on a long cooldown after use. Because of this, it’s in the players’ best interests to communicate concisely in order to make full use of these upgrades.

In this manner, we tried to ensure that each upgrade contributes something to enhancing the main features of this game (that is, re-contextualizing how controls work, as well as promoting that ever-important communication).

Banging metal to do repairs in the game adds a loud audio element to the game. How do you feel this enhances the hectic experience of the game (especially one where communication is important)?

Qiu: The clanging from banging on metal with a hammer certainly has potential to add more chaos to the experience, in that players may need to shout even louder to be heard. Still, as much as I wanted to promote chaotic gameplay, I still wanted to make the game as accessible to hearing-impaired individuals as possible. To this end, I’ve added an option to send requests for ammo changes and powerup activation directly to the other player’s screen should talking and shouting not be a viable option. This way, the noise created from hammering will only pose a challenge should the players wish it to.

Has building a game around a unique controller taught you anything unexpected about game design?

Qiu: It certainly has! The biggest thing is just how secondary the game can become when it comes to alternate controller games. If that the controller is interesting enough, the actual game becomes a pretense, or a justification, for just messing around with satisfying inputs. To this end, making sure that the controller is responsive and acts as the user expects may actually be more important than the game itself. Just like most good ideas, the most amazing, unique control scheme in the world can be completely ruined by poor execution.

Chambers: Something I’ve learned is how immersed players can get into a game when presented with controls that mimic the gameplay. It truly shows the power games can have!

Sun: It is important for designers to consider how to ensure that the player does not feel a disconnect between the content on-screen and the controller. We needed to enhance the player's experience when manipulating the controller, just as we enhanced the mechanic's repair experience with the hammer, and also used the turret to enhance the driver's shooting experience.