Development Process Behind Upcoming Online Co-Op Roll Together
Nearly Done Games spoke with us about their upcoming co-op platformer Roll Together, discussing the gameplay mechanics, the technical development of the physics and active ragdoll systems, and their marketing approach.
Please introduce your team. How did the development of Roll Together begin?
We are Nearly Done Games, a two-person indie studio based in Hamburg, Germany. We have been close friends for years, having met through the video game industry where we both built our careers.
Hasan Olgun – Technical Director: With more than 7 years of experience in the video game industry and a software engineering background, Hasan handles all engineering, systems design, and technical architecture.
Gökhan Altay – Game Director: Bringing more than 10 years of industry experience with an artist background, Gökhan is responsible for the game design and creative vision, art direction, and level design.
The name Nearly Done Games is actually an inside joke that every game developer can relate to. One of the biggest challenges in our industry is that so many projects never reach the finish line, whether it is because the project loses its charm, the scope was not properly planned, or countless other reasons. Completing a game from zero to one hundred is one of the greatest achievements in this field. We wanted to acknowledge that reality with a name that is equal parts honest, sincere, and a little bit funny.
As for how Roll Together began: we knew we wanted to create a co-op game with a fun, catchy mechanic that had never been done before. Our brainstorming centered around the idea of two players controlling the same object simultaneously. Initially, we almost went with a game where two players control the same bicycle, but we felt that would not give players the tangible, visceral feedback we were looking for.
We wanted something where you can clearly see the effect of every input, something with more chaos, more physicality, and more laughs. Around that time, we also noticed other developers starting their own similar mechanic projects, which pushed us further to find something truly unique. That is when the first seeds of Roll Together's core mechanic were planted: two players, one massive roller, total chaos.
First roller controller, applying forces on the capsule collider
What sort of gameplay experience can players expect? How do you keep it engaging with such a simple concept?
Roll Together is a chaotic, physics-based co-op parkour game where two players must steer a massive roller through treacherous obstacle courses. The twist is that each player controls only one side of the log; one handles the left, the other handles the right. When both players roll in sync, it feels so satisfying. When they're not, well, it's chaotic, funny, and a genuine test of patience.
Players rolling in perfect sync
The concept is easy to learn, but surviving the obstacles is a test of true patience and friendship. What keeps it engaging is the constant push-and-pull dynamic between the players. Every moment requires communication and coordination. The physics engine makes sure that no two attempts ever feel the same; a tiny difference in timing can lead to completely different outcomes, which keeps players coming back to try "just one more time."
Ragdoll chaos in action upon failure
We also ensured there are multiple ways to play. You can cram onto the same keyboard for some classic couch co-op chaos, use controllers for a more comfortable shared-screen experience, or play online co-op with a friend across the world. Rather than adjusting difficulty through checkpoint frequency, we carefully place checkpoints so that we can focus on crafting the level's own challenge.
Split-screen toggling, implying all the possible ways of playing
What references or influences defined how Roll Together plays and looks?
Our biggest gameplay influences came from the co-op space, specifically games where players share control of an object. Chained Together and Carry the Glass were our primary mechanical inspirations, the idea that cooperation is not optional but physically enforced by the game's design. When it came to pure entertainment value, the constant fun and chaos factor, Fall Guys was our north star. We love how Fall Guys creates that feeling of joyful mayhem where every round is unpredictable.
On the visual side, Fall Guys was always at the back of our minds as well. We wanted Roll Together to have that same kind of appeal: vibrant, cute, and irresistibly charming. This was important to us because let's be honest, nothing tests a friendship quite like one player sending the other flying off a cliff. But alongside that frustration, we wanted the experience to also feel heartwarming, entertaining, and visually delightful. It needed to look sweet enough that you would keep coming back even after your partner sent you flying off the edge for the tenth time.
Could you walk us through the technical development process?
Physics System: Building the Core Mechanic
The physics system went through an extensive number of iterations before we landed on something that felt right. We experimented with a single raycast wheel system, a custom collider wheel system, and a flat rigidbody collider physics system; each had its own problems with feel and controllability.
Side-by-side development iterations
What we ultimately settled on is a dual-wheel raycasting system paired with a Rigidbody core. Both wheels are driven by raycasts, and all forces, dragging, steering, acceleration, are calculated in the background. Each player applies force to their respective wheel in either the forward or backward direction. Based on the combined inputs, the log rolls forward, backward, or steers freely. Every parameter was meticulously tuned to achieve the exact feel we wanted: responsive enough to be satisfying, yet physical enough to create unpredictable and hilarious moments.
Showing the physics debug visualization: raycasts, force vectors, and colliders on the roller
Active Ragdoll System
Our characters use an active ragdoll system, which means animations and ragdoll physics run simultaneously. This is what gives the game its signature feel; even a small impact makes the characters visibly react, and players can see and feel every bump. When players hit obstacles, get caught, collide, or fall off the track, the ragdoll takes over completely, creating those wonderfully chaotic and comedic physics moments that are perfect for sharing on social media.
Networking with Unity Netcode for GameObjects
Implementing networking for a game where two players control the same object presented unique challenges. We quickly realized that the system needed to be host-authoritative rather than client-authoritative, meaning the host's machine is the single source of truth for the log's state, rather than each client owning their own object. This brought up its fair share of bugs and required significant refactoring of our initial implementation.
In the early versions, there was a noticeable latency problem, along with countless synchronization bugs. Sometimes characters on the client side would appear upside down or become invisible entirely. Working through these issues taught us a lot about the nuances of networked physics, and the system we have now is stable and responsive.
Scene Management & UI – Gameplay Transition
We work within a single scene for the entire game. The main menu and the gameplay area exist in the same scene, and we handle the transition between states by toggling objects on and off and rotating the camera. When the camera turns away from the menu environment, we simply disable it. It sounds simple, and it really is, but it is incredibly effective. We are definitely not fans of loading screens, so we approach every project with this philosophy of seamless transitions.
Since we never switch scenes, memory management becomes important to prevent accumulation. We built our own custom object pooling system that works seamlessly with both the physics objects and the networking layer. Beyond that, since the project is not texture-heavy, we handle standard rendering optimizations like batching. We do not squeeze every last drop of performance, but we follow all general best practices and are careful never to introduce anything that would negatively impact the player's experience.
What steps did you take to promote Roll Together and grow a community before launch, and what advice would you give to other developers?
We have been taking a fully organic marketing approach. We are active on TikTok, X, Instagram, Discord, Reddit, itch.io, and have been sending the game to many content creators. One thing we have learned is that there is no single formula for what content performs well. Sometimes, trend-based content gets great engagement; other times, completely organic, authentic posts do better. The key is consistency: keep putting content out there, keep engaging with your audience, and do not get discouraged when a particular post does not perform.
Our biggest piece of advice for other developers: do not treat marketing as something you start after the game is done. If anything, you should start sharing your journey before you even begin development, even brainstorming sessions can be compelling content. You need to grow your visibility just as much as you need to grow your game. Building a connection with the players who will eventually play your game is not an afterthought; it is part of the development process itself.
What is the current progress on Roll Together, and what are your post-launch plans? How has the community responded so far, and what lessons did you learn during development?
The demo is currently live on both Steam and itch.io, and we are actively developing the full version. Development has been ongoing for approximately six months, with work starting in late 2025.
The community response has been overwhelmingly positive. Everyone who has played the demo has loved it, from the art style to the sheer joy of the core mechanic. Players consistently tell us they want to keep playing after finishing our demo, which is exactly the reaction we were hoping for. The most valuable feedback has been about level difficulty, and we are actively using that input to tweak and refine our levels to deliver the most satisfying experience possible for every type of player.
For post-launch, we are definitely planning a new map, but we do not want players to feel like we are just serving them the same game reheated with new levels. We are actively experimenting with new mechanics that will add fresh dimensions to the core experience and give players reasons to come back beyond just new terrain.
Cooking up new levels
The biggest lesson we have learned is something we touched on in the marketing question: visibility is a parallel track to development, not a sequel to it. Start building your community from day one. And from a technical standpoint, lay down a clean, simple foundation and build your features iteratively, rather than trying to design the perfect architecture from the get-go.
Demo final level