Witchfire GGU Diary 1: Level Art and the Sunken Cost Fallacy

You know what is the common biggest mistake of a successful game studio?

They start making multiple projects.

I did it back in 1995. We – I co-owned a studio called Metropolis – just released Teenagent, that made us a lot of money, relatively speaking, of course. So we started the work on:

• Katharsis: a Virtua Fighter/Tekken style beat’em-up
• Haunted City: a photorealistic 2D side scrolling beat’em-up
• Tower of the World: a big RPG game
• Aquafight: a shoot’em-up (later named Katharsis)
• The Prince and the Sad Boy: ultimately titled The Prince and the Coward, a 2D point’n’click adventure

It was sheer idiocy, but I guess this is something a lot of us need to go through before reaching adulthood. I’ve seen this with CD Projekt (you would not believe how many Witchers they had in development at one point) and tons of other studios.

It’s not that taking a studio from being a one-project one to being a multiple-projects one is a bad thing. You just need to take it slow and be super smart about it. Like Remedy, for example, although I’m sure they had their own share of troubles – but who doesn’t?

With The Astronauts, we’ve decided to take the opposite route. One project, super lean team, and everything that can be outsourced will be outsourced. For our philosophy we’ve chosen the one I mentioned in the prologue post, something I call Valve Light: you need to be able to self-organize, seek work if you’ve done yours, and generally be super invested in making the best game you can make.

The plan was never to stay tiny. But we would hire only when we’re sure we can afford it and only when we find the best people who will mesh well with our philosophy.

A month or so before the launch of the Witchfire Early Access, we’ve hired a QA guy – Damian – because we knew we could no longer rely purely on outsourcing here.

After the release, we looked at the road ahead and asked ourselves: what is our biggest weakness? If we are late with stuff, if we cannot work fast enough on a certain area of the game, what is it?

This really wasn’t hard to answer. It was level art.

We literally have just two guys working on the levels: Adam and Andrew. If you played Witchfire, you know how huge and beautiful the current two maps are. It’s incredibly impressive that just two guys did most of it …but also one of the reasons why Witchfire is taking us as so long to make.

This absolutely could not continue so we’ve decided to hire not one but two level artists.

We announced this on a well know Polish website dedicated to gaming jobs – named Skillshot, after a Bulletstorm feature no less – and a couple of other places…

…and the flood gates have opened.

One thing I always did when looking for new Astronauts was to always reply to whoever applied. It quickly became clear that this would not be possible this time around. I had to edit the announcement and add a disclaimer that we would only write back to people who made the shortlist. This kind of sucks but was the only way to keep my sanity.

Most of the applications were bad, very bad, extremely bad, or plain wrong.

The wrong ones I can live with, oddly enough. “I know you’re looking for level artists but maybe you need a character artist as well?” – if you’re really good, yeah, maybe we do. It’s called “opportunity hire” – when an incredible talent contacts you out of the blue and you just can’t pass on such an opportunity.

But most of the applications weren’t this, they were on the subject, just bad. We’ve attached Witchfire screenshots to the text and said: “If you feel you can help us build worlds of this quality, please contact us”. And yet you get emails from people who clearly have no experience, no talent, no taste and no skill. It’s depressing.

But also expected. 99% of great developers are already working for another studio. They are most likely well paid, are working on prestigious titles like The Witcher IV, that RPG from Techland, etc. It’s people you want the most but are hardest to contact, let alone convince to switch projects.

But it’s not all bad, of course. Every now and then, you get an email from people who are talented but not too experienced yet and somehow slipped through the net, or great, experienced devs who by pure chance are between jobs or for whatever reason are on the lookout for a new one. When you’re good, money in game dev is more or less good everywhere but the people, projects and studios differ. You might want to, for example, leave a corporation with all its internal politics infighting behind and join a smaller studio where your voice matters more and where you don’t need to watch every word.

Anyway, a keen eye has probably noticed that I said we were looking for level artists, not level designers.

Yep, we’re making Witchfire the same way we’ve done Painkiller: with no whiteboxes, no blockouts, almost straight from zero to final.

Normally, a level designer designs the playing space and the combat arenas. This is usually called a whitebox and looks something like this:

Then, after many hours of testing, when the whitebox is fun and approved, it’s given to the level art department and they make their magic, turning a series of white blocks into a breathing living game world. Something like this:

This works and is actually a necessity for game in which the player movement is restricted, in a way. Pixel perfect. Like a jumping platformer game like Uncharted…

…or a cover shooter like Gears, where the distances between covers need to fit the cover-switching animation:

But for Witchfire we wanted a world that feels way more natural and real.

I mean, level artists can really do wonders, I never forget how a series of unidentifiable blocks turned into a believable shopping mall area in Bulletstorm. But still, it’s very hard to maintain the sense of truth in such environments, but also …they’re too perfect.

Here’s something I already used in my post back in 2014. Nintendo 64’s Goldeneye, 1997. Martin Hollis, the director of the game, explains the design of the game’s world:

The level creators, or architects were working without much level design, by which I mean often they had no player start points or exits in mind. Certainly they didn’t think about enemy positions or object positions. Their job was simply to produce an interesting space. After the levels were made, Dave or sometimes Duncan would be faced with filling them with objectives, enemies, and stuff. The benefit of this sloppy unplanned approach was that many of the levels in the game have a realistic and non-linear feel. There are rooms with no direct relevance to the level. There are multiple routes across the level. This is an anti-game design approach, frankly. It is inefficient because much of the level is unnecessary to the gameplay. But it contributes to a greater sense of freedom, and also realism. And in turn this sense of freedom and realism contributed enormously to the success of the game.

Obviously, it’s not that we’re ignoring the rules of play and just yolo the maps. No, we had many discussions of what makes a proper combat area in Witchfire, and Karol (design) spends extra time with Adam (level art) reviewing the levels. But we do aim for a mix of proper and real, calculated and accidental.

What this all means is: we were not only looking for level artists who can mesh a level. We were looking for level artists who could design and mesh a truly three dimensional level. A bonus would be if they could create a good 3D asset themselves as well, but the most important things was the level design and meshing.

Ladies and gentlemen, we got’em.

Actually, despite the vast majority of applications being bad, at the end of the day, our shortlist was almost ten people, and then two promising candidates applied after we’ve already made our choice. This is pretty big for us, all things considered.

When it comes to art, our hiring process is fast as hell. We know what we’re looking at when we see you ArtStation portfolio, then we have a chat and voila, let’s go.

I wish we could hire more level artists but as I said, we’re taking things slow, in opposition to when I wanted the studio to go from one to five projects those nearly thirty years ago.

But here’s the funny bit: only one person to join The Astronauts was from the job advertisement. Literally a day before, a great level artist contacted us asking if we might be interested. His portfolio included games like Call of Duty and Escape from Tarkov. I asked Andrew and Adam if they mind that their department – such a big word for a two men team – will have to switch to English. They didn’t. All they care is for the game to be great, as they should.

So we’ve just hired the man and along with the person we found through the job ad, we’ve now basically doubled the team responsible for creating levels. That’s not the end of that team’s expansion but more hires will happen next year.

This is great news for us …and for the players as well. We have neutralized our biggest weakness, we hope.

But …this is GGU Making Of diary. What does it all have to do with GGU?

Nothing.

Andrew and Adam and the new guys are not working on the GGU. They are working on the stuff that will come after GGU. The new levels, basically.

But I wanted to tell you all of the above to make you realize that in Early Access, you don’t just work on the next thing. You work on at least two things in parallel: the next big update, and the ones after that. And this is how it’s going to be until we hit 1.0.

So… Do I have anything to report directly from the GGU front?

I do but this post is already long, isn’t it? I literally wrote about a certain thing and just deleted it and will keep for the next week’s update. As a teaser of what’s it about, look at the title of this post again.

Other than this, it’s good old regular work. Testing new weapons, creating new enemies, designing and implementing the GG part of the GGU… It’s going to be funny when you look back on these posts knowing what GG means. It’s going to be super funny when we reveal that someone on our Discord was uncomfortably close to deciphering what GG means …without even realizing it.

---

Question of the Week

Just tried Witchfire with #UEVR and hooked up some 6DOF controls in <1min and feels awesome! #flat2VR #Quest3 @TheAstroCrew🎃👻🧙 pic.twitter.com/C5SVKiiO6q

— アレツクス (@waifu_enjoyer) October 27, 2023

…how?!