Why I won't implement full motion controls in my mods

I'm making this post public (as opposed to Patrons-only) so that new users can make an informed decision before they possibly join my Patreon.

My aim with this series of R.E.A.L. VR mods has always been to stay as faithful as possible to the original spirit and gameplay of the games that I port to VR. I want to enable other people (and myself!) to experience those beautiful virtual worlds from the inside, as if you were really there, but at the same time I strive to keep intact the narrative flow and interaction model that the game designers intended.

Take-Two and Rockstar, with their L.A. Noire: The VR Case Files experiment, explored the opposite approach: completely redesign the game so that it would adhere to the paradigm of what VR is "supposed" to be, i.e., with full motion controls and all kinds of comfort options. I played it of course, and really tried to make myself like it... but I couldn't. Besides being a commercial fiasco (which marked an important setback for the VR games industry), the conversion felt clunky and unwieldy, and even with their immense financial resources, they had to cut too many corners for the resulting product to be enjoyable.

The story was butchered (only a small portion of the missions were converted), and what was left was disjointed and hardly engaging. The graphics were dialed down so much to meet frame rate requirements that the game ended up looking like it had come out 10 years earlier; just compare a drive through the city at night in the original game and in the VR release to see what I mean. The motion controls implementation, despite being done by a whole team of people with access to the original animations, assets, source code and everything, was heavily criticized and just didn't feel right. It did offer good caricature material for silly YouTube videos, but choices like having to drive your car with imaginary hands on an imaginary wheel made for a frustrating and ultimately unfun experience.

Games like Half-Life: Alyx are a completely different story of course. They are designed from the ground up around VR interactions. All the physics, animations and character models are built with motion controls in mind (and yet, even with Valve's basically limitless resources, all we get are floating hands without a character body).

As a single developer, with the ambition of converting several huge, open-world AAA games to VR, and with a profound respect for the way those games were originally designed and intended to be played, I don't have the luxury of dedicating several man-years to properly implement motion controls.

A "proper" implementation would be a rabbit hole from which I would never emerge. I will imagine a possible scenario for RDR2, just to give you an idea. The ridiculously easy part would be mapping the motion controls buttons to the game inputs, and having the VR runtime report to me the motion controls position in space. That would only take a few hours, or days at most.

Then the first unsolvable problem would crop up: RDR2 maps a lot of functions to the D-Pad buttons, which are absent on all motion controls that I know of. A clunky workaround was proposed by Oculus, where the D-Pad can be emulated with one of the two thumbsticks, but that wouldn't work for RDR2 because the game already needs and uses both sticks for several different functions. But let's assume that by some awkward combination of buttons and complex gestures we somehow manage to simulate D-Pad input from the motion controls.

At this point the gun position must be tied to the motion control position in space. Flat games usually don't even tie the gun model to the character body; it's just a disembodied hand that gives the illusion of being attached to your body because of forced perspective. By freeing up the gun position in space, the illusion would break, and you would need to disable the character body being drawn (which was the approach taken by the old grandtheftvr mod by Crayfishery). I personally dislike this approach and find that it detracts from immersion, but let's push on anyway.

Now there's the matter of two-handed weapons like rifles and shotguns. Flat games don't implement any special treatment for them, because they are still tied to the camera exactly like normal guns. In VR, people expect two-handed weapons to obey some sort of strange physics where the gun kind of aligns in space with the imaginary line connecting the two controllers. Apart from the fact that I find that ugly and much harder to use than gaze-based aiming, even in native VR titles, let's again imagine that we can successfully mod the necessary physics into the game.

Then people request proper bow aiming. In flat games, the tension of the string cannot be controlled progressively: it's just a fixed animation where it  takes a certain amount of time for the string to go from slack to taut. So somehow we would need to take control of that animation and tie it to the motion control position in space, and modulate the arrow speed and trajectory accordingly.

Then players want every object in the game interactable with your hands. Again, the "grab stuff" animation in games like RDR2 is completely driven by the game itself. Some objects are scripted to trigger the animation, others aren't. When the character successfully grabs an object, the item then follows a scripted path, and at some point is shown to enter a bag or pocket or simply disappears. We would need to somehow rewrite this whole portion of the game physics and interaction model to allow the grab gesture on motion controls to move objects, and implement some holster or backpack mechanics in order to store them once they've been fetched.

And so on, and so forth. Driving vehicles. Climbing ladders. Pushing NPCs. Melee combat. Inventory browsing. Turning pages in books. It's never-ending madness. It's a complete redesign and reworking of the whole game.

I don't want to go down this path. So that's why I decided to draw the line at "no motion controls" :-)