Design Tips For VR Games

Editor’s note: Tadhg Kelly is a games industry analyst, design consultant and the creator of leading blog What Games Are. He is currently writing a book called Core Game Design. You can follow him on Twitter here.

In the past I’ve written about my strong skepticism regarding virtual reality. The first time I had a demo (from Palmer Luckey, in London) I felt sick as a dog for hours and this made me dubious. Second and subsequent demos softened my skepticism from “doesn’t work at all” to “might work for some“ and there it has remained. The headsets are inching toward becoming products and the media is broadly supportive but I’m still 30/70 on whether VR will ever make sense. There’s a lot, perhaps too much, about it that’s not quite there.

However I don’t want to bang that drum today. Last week, just prior to E3, I was fortunate to finally get to experience the now-famous Valve demo and yes it’s pretty cool. The bit with the whale is great, as is the bit where you can paint in 3D space. Some of it is less so – there’s a crude archery game and a fantasy adventure room that don’t quite hang together – but on the whole it’s sweet. You get a sense of how VR as a platform might work. Putting on my game design hat it’s made me think about the challenges of building for VR, the good and bad practices that loom large over it.

What I Mean By “VR”

As a product VR’s shaping up to be a few things:

  1. The main headset, wired to a computer
  2. Headphones, whether integrated or separate
  3. A positioning sensor
  4. Dedicated positional controllers, one in each hand
  5. Compatibility with standard joypads

These are essentially what Oculus announced and they broadly mirror what Valve’s produced as well. They imply a number of constraints from the get-go, such as:

  1. The primary location where players will actually use VR is in the home office or the bedroom. Most of the time players will be sitting down and will likely have a range of head movement of 270-ish degrees.
  2. Early adopters of VR are going to overwhelmingly be PC people rather than console people (because PCs are where the necessary power is at) and VR is likely to be more interesting to those who want to play in private rather than in the TV setting. Luckey can waffle all he likes about virtual worlds and social networks, but the primary use case will be absorbing single player games.
  3. The likely tastes for VR are going to divide between two distinct game types: 1) Indie experimental faire that’s comparatively cheap to make and all about experiences, and 2) Adjuncts to existing AAA games that have a mouse and keyboard version for everyone else. There isn’t going to be a significant dedicated AAA market for VR games for a long time, if ever.

Now that that groundwork is laid, here’s my thoughts on design principles for VR.

1. Which Way Is North?

One of the most important design hacks in the first person shooter is the targeting reticle. Why? Because it’s just hard to know where the center line really is – and therefore what your bullet will hit – when looking at a 2D projection of a 3D world. For VR that will be a problem (likely with the same solution) but a larger problem is overall sense of direction.

In demos that I’ve played, for example, I’ve often missed key animations/moments (in the Valve demo I almost missed the whale entirely because I was looking at the deck). Even with the ability to look around I can still very easily get lost and audio doesn’t really help this problem. Most headphones are capable of stereo and a crude amount of forward/backward positioning but nothing more precise, and so can’t be relied upon. Meanwhile more fancy room solutions (like 5-1 speaker setups etc) are fixed in position while the player’s head moves. Thus they too are unreliable.

So a compass is going to be a necessary component of most VR UI, by which I mean an arrow or directional indicator that the player can always see. It doesn’t need to be a big element but players are going to need it otherwise they will get lost and frustrated. Moreover that element will probably need to be dynamic so that when something important is happening (hello, big f-ing whale) it can flash to tell the player to turn around. Expect to need to use a lot of those kinds of tricks.

2. Don’t Make Me Move

My biggest realization in trying the Valve demo was that physically walking around was mostly a bad idea. The demo has a good solution for showing you boundaries of safe space (a white grid fades in and out as you get close to the edge) but that wasn’t the issue. The issue was that sometimes I could feel myself stepping on the cable connecting the headset to the computer and I became very fearful of tripping. That duality of immersion and physical danger is not good.

And that’s in Valve’s wide open demo space. We’ve seen this problem before with Kinect, where Microsoft’s notion of the standard living room proved wildly optimistic (well-paid tech people living in Redmond tend to have big houses) for the real world. Similarly testing labs for VR may have lots of open space but the likely locations for VR to be used will tend to be cluttered. So you’ll need to figure out how to get your players moving through your game world without physically moving.

With joypads this should be relatively doable. Much as first person shooters have WASD and mouse to manage both axes of movement, in VR it can probably work much the same – although the combination of user head turning and body turning in VR FPS is generally not comfortable, and not solved yet. Another option is to take movement away entirely. Endless shooters and runners could work very well for example because they remove the conflict between the player’s need to see and the body’s need to turn. And yet it’ll be pretty sad if it turns out that the best kinds of game in VR are those that are on rails.

Gestural controllers present a different problem. One possible solution would be to assign movement to a left-stick (much as a joypad) but that might feel limiting, as though the player only has one limb because the other is effectively a gear stick. Another option would instead be to take advantage of the positional reading to design tank controls. If I push both controllers way from me that might mean “move forward”. If I push the left forward and pull the right back that might mean “pivot right” and vice versa.

3. Never Seize The Camera

You may already know how annoying cut-scenes and quicktime events can be in games. When badly implemented what they tend to do is simply take control of the camera (and other controller) functions temporarily and then place you into autopilot so that you see the Big Important Cut Scene That The Studio Spent All Its Money On. Please promise me that you’ll never do that in VR.

Why? You’ll disorient players, make them spin-dizzy and vomit. And then two years from now I’ll be on this blog writing about your most hated game saying “I told you so”. Listen: free camera movement is the whole selling point of VR. It’s supposed to be immersive because the world is all around the player, inviting discovery and self-driven experience. Forcing that out of their control just for the sake of a story point is not going to help you win friends and influence people, trust me.

4. Pincer Controls

Much as Kinect was sold on the idea that it was reading your full body, VR is likely going to sell its gestural controllers on the idea that it can track you in full real space and such. This means you’re likely thinking “lightsaber fighting at last” but not so fast. For alas, much like Kinect, the reality is much cruder.

Gestural controls in VR seem to be good at reading orientation (is my hand up, down, tilting) but bad at reading exact points of activation and release. For example the controls work really well in Valve’s 3D painting demo because it’s relatively free flowing (and easily the most fun of all then demos). But when trying to reach and grab objects in a scene to pull, push, throw and release? Then the controls are about as precise as robot arms on an undersea sub.

Of all constraints this one may be the hardest for many games to overcome. Limited or dumb movement means that players will grow to hate any game that uses them. They’ll just be too fiddly. An LA Noire style detecting game of examining clues, for example, could very easily be the dullest experience if half the time the player struggles to pick up clue objects in the first place.

VR will probably fare better with gameplay types that are ranged (shooters) or more pattern-oriented about actions (like Street Fighter, where the input is essentially a very fast Simon Says game). Anything that’s trying to mock actual movement, however, is a long way from being credible. Put the lightsabers away for now.

5. Depth Perception

Finally a word about optimal distances. A thing you need to realize is that – much as with 3D cinema – in VR there is such a thing as too near. At a certain distance (for me it’s about arm’s length) there’s a point where my eyes struggle to make sense of an object or scene, and it becomes less believable the closer I look. So as a general rule I’d try and avoid designs based on close work. A good example would be something like a bomb disposal game. It probably sounds cool in principle that the player would staring at tiny components, looking at wires and trying to figure out which ones to cut. In practice? Probably not so much. Combined with the imprecision of controllers particularly I think this kind of game would be very hard to get right (as would safe cracking and a variety of similar activities).

So What Can I Make?

Even with these constraints there’s still an incredible canvas at your disposal. I’m not saying that there’ll be a market for your wares – not yet at least – but there’s certainly plenty of room for experimenting with VR to see what it can really do. Good luck!