When: Sep 2018 (2 weeks)
Where: Entertainment Technology Center, CMU
Roles: Game Design, Producer
Team Size: 5
Platform: HTC Vive + Trackers
Engine: Unity
The Process
For this second round of Building Virtual Worlds, we were tasked with creating a game for a naive guest. This was the most difficult round. Additionally, we had to use HTC Vive Trackers to bring in real world objects to the virtual world. Lowering the skill floor as much as possible and properly telegraphing control methods were the hardest parts of this world.

Brainstorming ideas was also inherently difficult. I lead most of the initial discussions as we threw out things and saw what stuck. As ideas become more firm, we started combining them, culminating in robot space garden and Dance Dance Revolution balloon. Seeing other teams go for gardens, we went with the freedom of flight.

Concept art inspired by Ha Long Bay's flowing river and spires - By Team Artist Boyi Liu.
The original idea behind the balloon was simple: flying is fun and screams freedom of choice. There was no specific reason that we chose balloon, as it was just one of the brainstorming words, but it turned out well when we combined it with fur cardinal direction arrows on the floor, similar to DDR. It seemed obvious to all of us that of course you would move a balloon by balancing out your weight with it. This would turn out to not actually be the case in real life, but it worked for us and every playtester, so we rolled with it. We quickly rigged a headset to track its position over an XY plane and played with that.

While standing on (really, moving your head over) different directional arrows move the balloon in that respective direction, we were still stumped on height control. We played with constant height, constant downward height, and constant upward height, but the pressure to use trackers spurred us to use them as our height controllers. I suggested we use them as sandbags placed at their feet. The guest would be primed in the real world by seeing these physical trackers below them, and would definitely look down at them when the headset was on, and then also see the directional arrows and their feet. This chain reaction would hopefully be enough to teach them how to fly when combined with another balloonist in the world who would practice sandbag throwing for them. Ultimately, while fun, the instructors didn't really like us throwing the expensive tracker hardware, so we moved to other ideas.

What we settled on was a lighter using the HTC Vive controller. The back trigger on the controller is very fun to click and is the most easily pressable button. We were suggested the lighter idea and implemented it quickly, giving the guest immediate up control and relied on passive, constant downward control that they could counteract as they pleased.

Now, briefly, level design. We wanted an initial staging ground to test the guest before setting them free, making sure they knew how to fly before completing any challenges. I suggested Ha Long Bay's rock spires and river. The guest would naturally follow the river, but not feel forced to, and have plenty of rocky obstacles to move around as they went, earning their sea (or air) legs in their balloon.

The Result
The final presentation linked below was one of the most relieving moments of Building Virtual Worlds. When the naive guest took off and immediately understood how to control the balloon, a great amount of pressure was let off. But you may have noticed that I haven't brought up the tracker requirement again after ditching the sandbags. We chose to effectively not use the trackers at all—we stuck with the hackjob head-movement tracker from before and tricked the naive guest into thinking their feet were important by putting the trackers on them. Now, they just had two feet models that would make their feet seem important, and help them connect their feet to standing over the arrows to fly. While I'm not sure if this counts as cheating, it worked great as a headfake, and all initial guest interaction fortunately went exactly as planned.

One single comment made me reconsider everything our level design had to offer. To recap, we were using a tight canyon area with rock spires cut through by a river that opened out onto a wider bay. This was to make players firmly understand flight before flying more freely. But that one comment suggested we reverse it: learn to fly in a large open space before doing tight challenges. This was an incredibly better idea; I had been doing the hard canyon part over and over again because I had already learned to fly, naive guests wouldn't have that experience. For us it turned out ok in all playtests, but taught me to freshly reevalute the challenge curve of anything I design and make sure the ability level is right next to the difficulty level.

We did have to cut the ending out after breaking the build three hours before deadline, but that's part of the fun of BVW. Below is the video with the guest who has never seen the experience before.