Breaking the results in ABRISS - build to destroy

This interview is part of our Road to the IGF series.

ABRISS - build to destroy is a game about destruction, having you build up structures only to knock them down in some spectacular fashion. And usually in such a way that you have to hit certain targets. How do you blow up a building in such a way that you hit that little target over there?

For their IGF Best Student Game-nominated title, Randwerk Games eG had a talk with Game Developer about the design details that go into making something satisfying to destroy, the challenges that come up when you're dealing with these kids of complex explosions and collapses, and the thoughts that go into making a building-creation system when you're just going to break the results in the end.

Who are you, and what was your role in developing ABRISS - build to destroy? 

Johannes Knop, co-developer ABRISS - build to destroy: Hi, I’m Johannes Knop. I did a lot of the coding, game design, some sound and music, and some UI work. A lot of business stuff, too. We’re just three people, so everyone does many things. 

What's your background in making games? 

Knop: I have a Bachelor’s Degree in Game Design (almost finished my Masters). This is the first commercial game I've worked on. It’s the first project of that kind for my teammates, too. We’ve worked on many projects at the university before, though. 

How did you come up with the concept for ABRISS - build to destroy? 

Knop: ABRISS is kind of a successor to a game Till Freitag and Friedrich Beyer had worked on for a university course a couple of years back which was called ANTITECT. The premise of the game was destroying buildings and then getting resources depending on how much was destroyed. They approached me in the Summer of 2020 to try to release the project (I think at one point we even planned for a 2020 release?). Then, we got together and made the first ABRISS prototype, which took a lot of ANTITECT’s premise but packaged in a level-to-level puzzle progression with an all new art direction. We formed the Randwerk Cooperative, applied for funding, and have been working on the game since. 

What development tools were used to build your game? 

Knop: It was made in a couple of different versions of Unity - we’ve been upgrading from time to time—and visually, we’ve been making a lot of use of the High Definition Render Pipeline (HDRP) features. The volumetric fog, for example. The physics system is just the stock Unity physics with our own optimizations. We’re usually using Blender for 3D modeling. There’s really nothing exotic in our pipeline. I personally really like JetBrains Rider for programming. 

What interested you in having building creation and destruction in the same game? Why marry the two concepts? 

Knop: It has a neat circularity to it. Building, then destroying, then building again, then destroying again—there’s just no end, I’m sure there’s a lot of philosophical concepts to be found in that. There’s also a long tradition of city-building games in Germany, like the Settlers or Anno series, and we’ve kind of asked ourselves what the opposite of that would look like -a game mainly about destruction. But it still has building in it, so it’s more an anti-building game than a destruction game, if that makes sense. 

In ABRISS, you're specifically building things to break stuff. Did that affect the building creation mechanics in any way? How is it different to create building tools in the game when it's focused on destruction instead of creation? 

Knop: I guess it makes the mechanics focused more on the pure physicality of what you’re building. So, it’s a lot about the dimensions, the weight distribution, and stability. In a typical building game, you would want your creations to stay, maybe to look nice, produce or transform resources in some form, and their shape or centre of mass would probably not matter as much. Even in a game like Besiege, you want your vehicle to remain and be intact while it destroys things. Our constructs are purely functional. 

The building mechanics have to be fast and snappy; no millisecond can be wasted. It’s not supposed to last; it will break anyway. The thing that matters most is if it will fall the right way, fly far enough to reach the target, or reach it with enough speed to destroy as much as possible. 

How did you design your destruction system to make the devastation truly satisfying? What makes for an appealing bit of destruction? 

Knop: A lot of what makes the destruction impactful has very little to do with the kinetic physics of the situation. That’s a big part of it, of course, but a lot of the reason why it can feel impactful is effective use of some built-in features. For example, the field of view of the camera. We have a very narrow FOV, which makes everything look and feel large. The same models wouldn’t feel like larger-than-building-sized megastructures with a default 60° view angle. 

There’s also a lot of sound design. We have several big room effects layered in succession, like delay and reverb, and apply that to most sound effects. That makes it sound far away and like you’re in a gigantic space. A correctly-timed camera shake also does a lot to convey weight and impact. We also do time dilation, so there’s slight slow motion in certain situations. 

The destruction itself, I think, gets a lot of appeal by the sheer amount of parts involved. We optimized every single part so we can just add more. They also subdivide in a smart way, so there’s not many active at the same time. There’s a lot of illusion involved, of course. After a certain size of block breaks, the debris is just GPU particles. But if there’s not that many simulated objects in the scene, but we can make you think there are, do you really care? 

What thoughts went into the various pieces and tools players can use to create havoc? Into the building blocks of the structures themselves? 

Knop: We thought about our building parts in a very functional manner. It was always important to us that all parts have a single function - that there’s not three parts who do kind of the same thing. The gun shoots, the ultra heavy cube is ultra heavy, the rotator rotates. So, you have to combine them for the construct to do something more complex than the single parts can—kind of “complex behavior emerging from simple rules” approach. 

We’ve talked about it like our constructs are, in a way, similar to a Magic card-combo. They also just need to look like what they do. You have to be able to tell them apart, and they need to communicate how heavy or breakable they are. The target blocks mainly need to look good next to each other, be modular, and have good segments that they can break into. They also need to be super highly optimized, because there’s just going to be so many of them at once. They’re super low poly and most of them even share the same material to minimize draw calls. 

Also, what ideas go into making something that is satisfying to break? 

Knop: I think one cool effect that helps with that is that we have shells that generate around the whole outer area of the targets - small white panels that start to fall off even if the target isn’t completely destroyed. The targets need to have a good place to fall on once they fall down or tip over, and they need to fall from enough height to actually break. So, a lot of the time, we’ll place some cliffs or pillars or walls where the targets would fall so they crash into something in the environment instead of just vanishing into the void or landing on the ground in one piece. 

When building a target, what's most important is the placement of the main targets - the red special blocks that you need to destroy to win (not all of the targets have these). They’re extra satisfying to hit, because they cause a small explosion, which often also means the rest of the target is destroyed. The structures themselves will be constructed from easily breakable blocks and harder to break blocks, so there are some predetermined breaking points in the whole target. Targets can also have armor plates outside to protect them from guns or bombs. They can also hover, which means that they will usually fall even if there’s just a small impact. 

Usually, we start by copying another target and then change it up until it looks completely different. If we have several targets in one level, we can also bake placeholders we call “Skins” that look exactly like the target but are just one object instead of hundreds that replace themselves with the actual blocks on impact that keeps the overall number of blocks at once in game down. Generally, we have to be careful with how many blocks we use, so we will sometimes make big targets hollow. That sometimes makes impacts more spectacular, too, because the collapse is more complex. 

Did ABRISS change much over the course of development? Why do you think it changed, or that it didn't change? 

Knop: Yes, it changed a lot since our first prototype in 2020, and is really a different game from ANTITECT now. We started out with something we thought had pretty cool visuals and effects at the time, and if we now look at it we really don’t like it anymore. We almost named the game “I BUILD TO DESTROY”, because we were scared the German name would deter people, but it seems like it made the game even more interesting. We kept the “build to destroy” slogan from that name, because it’s nice that the title already explains the main mechanic. 

We released several demos and did a lot of testing, and every time we found that something that seemed important to us was something that players really didn’t care about, other things we hadn’t even thought about players really wanted, or we didn’t have activatable objects, at all for example. It’s actually very reassuring to me once you have released something, because you can just look at what people are doing and the problems are immediate and obvious. Game crashes in level 5? Ok, fix that. You don’t have to guess anymore if something is important or not. 

It changed a lot because we tested it a lot, and I think we could probably have done even more. That works the other way around, too. We never really changed the art style too much because we saw that people on social media really liked it. So, we just added fine details and tried to make that style even more defined.

This game, an IGF 2022 finalist, is featured as part of the IGF Awards ceremony, taking place at the Game Developers Conference on Wednesday, March 23 (with a simultaneous broadcast on GDC Twitch).