Sometimes to do something good, you need to go through a bunch of fails. Which in our case came in the form of the game about two warships battling each other. At the time, we were asked if we could make a VR game that would not be a shooter but still would be somehow related to the games Wargaming is famous for. My colleagues still are trying to forget that nightmare, and I'm sure if there were a way of wiping out their memories of this particular episode, they'd gladly agreed to it.

Our warships even sank silently. We failed to spark interest in players, but the shock and disappointment from this project served as good preparation to have another go at a game about tank battles in VR.

And that is how the work began.

The results of the first prototyping were scathing.

Since you can't even think about making a game about tanks in VR without knowing for sure that you do know how to solve motion sickness problem - that was our first task. So we created a moving platform, put it on uneven terrain, put our VR headsets on and there it was.

"See! I told you it's impossible," I said.

One of our first ideas was to increase the number of details. I assumed that dizziness might be diminished if there were more visual details. We added some rocks on the terrain, planted a few trees, and built a railing around the platform. The effect was encouraging, but we had to do better than that. So we experimented with the density of details and different types of environments. For example, in a stylized world the brain was less affected by motion sickness than in a world pretending to be realistic. We then spent quite a lot of time coming up with the next clue. The fantastic thing about our brain is that it prioritizes every task. So if you give it a simple and relatively easy objective, it will switch to it — given the environment is immersive enough — and for a while forget about the dizziness.

"And when were you going to think of testing the actual game?" my inner voice gloated.


Jokes aside, this was a groundbreaking discovery for us. From that moment on, we started to tackle the problem of motion sickness in all directions at once, combining different approaches both in visual style and gameplay mechanics.

Our first idea about how the vehicle should be controlled was using levers.

In the process, we found that there is merely a universe of opinions on how those levers should exactly work. At first, right and left controllers were respectively bound to right and left tracks. The game became a struggle with the tank rather than a fun experience.

Watching those poor souls trying to tame our tanks, we showed compassion and came up with the idea of a joystick in VR. Easier said than done: We never got tilt-controlled direction combined with angle-controlled acceleration to work in virtual reality. It was too hard to distinguish unnecessary movements from those meant to control the tank, so we went back to the idea of levers.

But this time each lever was responsible for the rotation of tracks: forward or backward, whereas the direction of movement was bound to the player's line of sight. Mastering the movement of the tank became much easier. It took us a few more iterations to figure out how we shoot, how we aim, and how different speeds of the tower rotation work. And the easier it was to control the tank, the less motion sickness we had. But right when we thought that it was over, we changed the whole system again.

And here is where I should thank the father of PC World of Tanks - Slava Makarov. Although the initial idea was born somewhere in our "Slytherin" R&D department, it was Slava who came up with the right formula: Why rely on a tilt, when we can mount both the acceleration and direction on the triggers. The one on the right controller would drive you forward, the other - backward. Since that final touch, the controlling of a movement became almost telepathic. Our players stopped thinking about how to control the damn thing and focused on where to go and who to kill.

Every now and then we had entirely unexpected setbacks.

One of those was finding the right tank model. Now, if you ask people what the model of a tank should be in a Russian PvP VR tank battles set somewhere around WW2, ten out of ten will say the T34. Well, I've got bad news for you.

You see, in our prototypes, we used a T76 model. And a T76 looks like pretty much everyone's image of a modern tank. It has very, well, tank-like proportions. And when we changed it to a T34, it was a disaster. Its proportions make impossible for a player to see the base of the tank. So when the tower turns while the tank is moving, and there is nothing to catch your eye, motion sickness strikes back.

Once we got over an overwhelming panic attack, we started researching every Soviet tank from those years. Light tanks were too high and too fast. Heavy tanks were to slow. So we ended up with the medium tanks, and finally found T44-100.

Our initial idea was that we stick to the maps that PC players are used to in World oF Tanks, which meant a rugged terrain, warm colors, a bunch of trees for players to cut through, and that's it. We had neither time nor the resources to go with more than one map, so we had to make sure that the map we provided at the start would have everything we needed: enough landmarks and a landscape that was not going to make rounds too messy or too dull.

"And the range?"
"What about it?"
"Well, how big a map are we talking about?"
"For Christ's sake, man, three ranges of an arrow? How would I know?"

Well, it may not have been arrows, but the diameter did depend on the projectile's speed, the tank's pace, and the average time we intended players to spend heading into active fight zones.

And if that were not enough, we got into a fight on what time of the year it should be. While our Slytherin aesthetes wanted it to be winter and snow, I was adamant on bringing players into bright autumn or summer which, as I thought, would be more fun.

When it comes to making VR game, one of the biggest headaches is the problem of how a players see themselves, and how other players see them. In our case, other players should have seen each other as tanks. Although at earliest stages we argued about that too:

"And how do you think it 'd work out? Behold, here goes a tank with a couple of hands and a helmet floating above its tower"
"What if we show a person's torso coming out of fade?"
"Well, good luck having tankish centaurs!"
"Guys, if those are centaurs then what does it make a gun sticking from... Oh, wait!"

Freudian mention of a gun sticking out from right where player's legs should have been put an and to the dispute. Our final decision was quite elegant: no torso, no helmet, just a pair of hands holding the tops of the levers, with all the controls player might need. The button on the left controls the locking view and zooming in, the one on the right controls shooting.

All too often, (Russian) game developers get stuck at the prototype stage and then wonder why it didn't work out in the end. That must come from our school education mindset over here. Our parents teach us that if we solve a math problem with a minor mistake in calculations or formatting, it's still okay as long as we get the answer right. Well, moms may love us for what we are, but the rest of the world doesn't really care about our brilliant ideas or insights. What it needs is a tangible result.