12 min read

SteamCrew VR, a postmortem for a Mobile VRJam project

This article is a postmortem of the Oculus Mobile VRJam 2015 Platinum Winner game SteamCrew VR for GearVR. It describes the process, design choices and problems we we went through during development.

SteamCrew VR is a virtual-reality cooperative arcade/simulation putting 2 players in a Steampunk submarine and was selected as the contest's platinum winning project. It runs on GearVR using Samsung Note4 and Samsung S6.


Tools used: Unity, Blender, Photoshop, Inkscape


First I think a little background would help contextualize things;

I. Some Background

We've been working on VR games for a few years now and this is our second game jam, in 2013's Oculus VRJam we scored in the Top10 with EpicDragon VR without winning any prizes.


In march 2015, we were working on a non-VR game prototype, our VR were projects on hold, not knowing when headsets would be ready for consumers.


It's been tough working on VR in France, we were earlybirds and had a very small social network even after 2013's satisfying results. We decided that it would be nice seeing Oculus and other VR developers in person so we registered for GDC aside of our goal, pitching our non VR game.


Turns out GDC was all about VR, it made us rethink our prioritiesWe got in meetings with Oculus, Valve and heard about the Mobile VRJam being set up in the coming weeks. At first we were a little skeptical about GearVR, it seemed a weird little gadget that could not rival bigger and more powerful headsets running on PCs or consoles. We got to try one and we were surprised by how polished the whole experience was.


It made us eager to find concrete applications and think about practical design cases that could define it's use. We did a lot of Oculus DK1 and DK2 demos around Lyon. It was always cumbersome to carry our computers around, we saw the Gear as a far more accessible VR device to new and casual users. It is an ideal mobile device for these local multiplayer games we love playing with friends.


As we were tossing ideas around during the evenings at GDC, I had a concept in mind that I wanted to try. A weird mix of the interaction frenzy you experience when playing Spaceteam. The immersion of playing Guns of Icarus Online and the synergy required between pilot and gunner to be effective at flying choppers in the Battlefield series.


After a few team talks we agreed to work on this idea when the jam would start.


II. Jam's Start


Our team was composed of two full time developers from Agharta Studio, Aurélien, our programmer; and myself, in charge of design, graphics and sounds. Our two friends Sandra and Mathieu from Blackmuffin were doing part-time concept art and helping with promotional visuals and video trailers.


We are used to work remotely but for the jam we decided to rent a 2 desk office in order not to waste time in sending tons of mail.

At the time I stumbled upon Clockwork Game Design by Keith Burgun. Using his ideas, I tried to keep the smallest number of mecanics, finding if each one was really justified and on focus. Some ideas were tweaked but a lot were scrapped.



Our goal was to define practical uses for GearVR players, encouraging them to get together and play instead of feeling isolated in VR. We wanted to have social interactions at the very core and build mechanics around it. Building multiplayer features was required, along with a very simple lobby.


The game had to be easy to engage in for new players, be confortable and maximize accessibility through the simplest controls possible.


I have trouble with swivel chair or standing experiences, I get easily sick of all the rotations required by interacting with gameplay elements behind your back. It's also very restrictive in terms of player postureWe decided to go for a control system that was similar to the one we designed for EpicDragon VR, based on gaze driven rotations.



We chose to constrain playtime to around 12 minutes max though increasing difficulty to ensure players would take breaks out of the game and debrief their game, giving them regular opportunities to continue or stop playing.



At first we were hesitant to go for a space sim but we figured ocean depths were a great setting for contemplative and the slow pace needed for a confortable VR experience.



Immersion was another important factor, players had to see each other move in the game universe. As the GearVR features no movement tracker, there was no need for complex IK rigging and we could easily do full body anims on actions. Early on I wrote a little piece of story for the game to help us characterize the universe and give names to objects.


We did a hard choice at the end of GDC, we are used to build game engines for every project we work on. We wanted to use the jam as an opportunity to learn Unity, also because it seemed multiplayer was going to be easier to implement. We did struggle to get it to do what we wanted at first but we finally managed to understand the engine's subtlties halfway through.


It was important to offer a full gameplay loop and design 100% exclusive assets not to have any biases on how to use these and allow us full creative freedom for the gameplay.


III. What went wrong ?


On Graphics :


I love color saturated hand painted textures, it reminds me of that awesome SEGA arcade game look. We did use that style for EpicDragon VR and I think it worked well.


However, during the course of developing SteamCrew VR, we faced a few problems.


I started texturing very early on and I felt reluctant in doing structural modifications for the gameplay once things were textured. Everytime I had to do big adjustements it was a struggle to keep clean UVs. It prevented important modification from being easily made and overall was not a very good choice.




Add that to the fact that our designs were closely bound to gameplay constraints, like the sub's layout, how equipement was installed. It had to be fuctional and look cool.


It was really hard to communicate with our concept artists the exact thing we wanted without them on-site. I had to make extensive artistic blockouts versions of the cockpit and other things to pass on to Mathieu and Sandra for detailing, making sure gameplay concerns were adressed.



In the end I'm still not satisfied with our interior layout, players cannot see each other enough and the sense of presence could be enhanced immensely if they were sitting side to side.


Games like DRIFT and Neuralgia from the same jam did approach this problem through better and more pragmatic artistic choices using simple textured polygonal style. Aknowledging they were jam games with time constraints, I think it helped them keeping reasonable production times.


On Gameplay :


We started serious playtesting very late, found that the game was far from being as clear and as understandable as we initially thought it would be. As the project progressed we slipped focus, hesitating between two opposed directionsShould we aim towards a fast paced game to make sure players had stuff to do? Or towards a Spaceteam-like set of game mechanics with the inherent complexity it brings to an already complex control scheme ?


There were serious issues with how players perceived the submarine's equipement, what could be interacted with and what was just scenery. There was a lack of feedack on what was going on in the Sub. I thought sounds and blinking objects would be enough but we had to work out a whole « gamey » floating HUD layer right at the end.


Overall, our focus of a purely mechanical HUD to reinforce immersion wasn't a great choice retrospectively. We had a complex game that required players to know what they were doing, and we were reluctant to give them visual cueing for the sake of our idea of immersion. Tooltips were added at the last moment to make sure players could at least see what they had to do, which still doesn't replace a properly designed learning sequence at game start.


Controls were extensively tweaked, we radically changed the scheme until the very last days, we did go through at least 5 or 6 radically different iterations and yet we are still not satisfied with the results. I'm still glad we did not stick with the first versions that required very unnatural inputsThe book Game Feel by Steve Swink proved very useful for that matter.


We took a lot of time tweaking damages and enemies health values to encourage players to flank enemies. I'm still convinced we could have done better choices but I guess that worked for the time being.


Last but not least, no time to implement an online mode prevented most of the player to experience the game, which makes us sad but having working local wifi is already a challenge!


IV. What went right ?


We had a strong vision based on a feeling and we wanted to build the mechanics around it. If mechanics did not serve the coop vision, they were out. Our focus of making a strong coop game proved successful, it seems like the type of games Oculus were willing to support, as they hinted in the Jam's rules.


Working in the same office as Aurélien proved to be a huge time-saver for fixing problems. It saved us hours of email debates. We could just chat and decide on quick fixes or scrapping a feature if it was too long to implement.


To digress, it's funny how it went for EpicDragon VR last time. I was on-site with Sandra & Mathieu with the only Oculus DK1 we had. Aurélien did work remotely. He put together a way to emulate Oculus controls through a gamepad in order to test his builds, we're glad we had enough hardware this time ! 


Even if it wasn't the best choice we could do, the game looks turned out good on the graphical part ! It did raise my texturing skills which was a very frightening challenge. Using extensive colored lights baking techniques in blender helped with color mood on textures. I found that we had an agile enough process to change controls, we did not got stuck with choices we did not like for too long, thanks to Aurélien's great skill at coding tools !


In general working with our own assets only code and graphics wise proved a good choice, we did have the flexibility needed to do and show exactly what we wanted. It helped immensely on the optimization process, merging batches and keeping the simplest shaders possible is crucial on GearVR.


Getting working multiplayer is a great thing for us, it's the first time we worked on a live multiplayer game. We distributed calculation between seats, everything related to the enemies is processed on the gunner's side and then synchronizedIt was hard to get right for Aurélien but when it finally worked it saved a lot of cycles for the pilot.


One last thing that did work well was using devblogging as a tool for both marketing and team focus. Knowing that we were a part of an ongoing process that other people could see is was big motivational boost for all of us. Im not much of a PR person and it took some time to maintain but I think it's well worth the effort, we were one of the most followed thread on the Jam's forums !


V. What's next ?


The prize we won encourages us to finish this game. We definitely want to release an expended SteamCrew VR game in the near future. It will go through a big refocus in terms of where we want the game to go but all the valuable lessons we learned from this jam are of immense value for us.


It's probably going to go towards a simplified control scheme and deeper collaboration through a more detection/systems based gameplay.


We are going to add defined objectives to the game in order to engage players in more meaningful ways, solely maintaining a sub isn't that fun in itselfThese goals could also be a part of even more complex interactions between players such as moral choices and risk takingIt also means that we're going to have to define the historical context a little more to engage players in something solid they could build their own stories on.


It's going to feature new biomes, content, new enemies and modular equipement to cater for different play-styles, we want to retain the rogue aspect of short sessions to encourage players to try different approaches.


On the technical part, the game will feature online multiplayer, hopefully with voicechat. We need to design a troll-safe game, it could be a huge problem in a game where interactions are so important. A levelling and friends sorting system seems mandatory for this reason.



Further information on the Game's developement can be found on this thread on the Oculus forums


As a conclusion I would like to congratulate all teams from the Jam, these are many awesome and clever games and I'm amazed by the inventivity there is in this VR community. Last jam we did not won but the experience it gave us was an immense reward and opened us a whole world of possibilities, paving the way for this project.


I would like to thank Oculus and Samsung for the great opportunity this Jam was. Seeing all these projects coming along and this community being built, exchanging with talented and passionate developers is an unvaluable thing.


Obviously it's very sad that most games were not tested based on their descriptions and looks, games have to be played. Testing 300 games is a task of titans, yet people have worked hard for a whole month on these.


Anyway I'd be happy to provide feedback on any of these games !



Thank you for reading !

Latest Jobs


Playa Vista, California
Audio Engineer

Digital Extremes

London, Ontario, Canada
Communications Director

High Moon Studios

Carlsbad, California
Senior Producer

Build a Rocket Boy Games

Edinburgh, Scotland
Lead UI Programmer
More Jobs   


Register for a
Subscribe to
Follow us

Game Developer Account

Game Developer Newsletter


Register for a

Game Developer Account

Gain full access to resources (events, white paper, webinars, reports, etc)
Single sign-on to all Informa products

Subscribe to

Game Developer Newsletter

Get daily Game Developer top stories every morning straight into your inbox

Follow us


Follow us @gamedevdotcom to stay up-to-date with the latest news & insider information about events & more