Combat game development: The fight to finish Galak-Z

When I called 17-Bit CEO Jake Kazdal last week to talk about his game, he sounded cheerful and excited about how development was going. He also sounded a little desperate. "We're running out of time and money," said Kazdal, whose studio also developed the strategy game Skulls of the Shogun. "All the internal pressure is on beating the holiday wave." Kazdal doesn't want to change the "sometime in late 2014" release target 17-Bit has set for the PlayStation 4, PS Vita and PC game Galak-Z: The Dimensional. The trouble with sticking to that 2014 release is that the game itself has changed so drastically. Kazdal likes to say it's "gone full roguelike," a concise statement that too neatly sums up what sounds like the staggering amount of work involved in tuning up this Unity-powered 2D space shoot-'em-up initially designed to be what Kazdal calls "a traditional Metroidvania action-adventure," with a suite of procedural generation systems. I wound up speaking with Kazdal at length about the process in order to learn more about the development of Galak-Z, and why 17-Bit decided to switch gears and embrace procedural generation, throwing out months of work in the process. Kazdal says he has "zero regrets" about making the switch, and in this edited transcript of our conversation he explains at length why he's confident the late-game switch will be good for Galak-Z, no matter the cost.

So how's development of Galak-Z coming along? I confess, I don't really know when you plan to ship.

Yeah you know it's funny, 'cause we don't either! We're trying to hit a certain date and we don't know if it's going to happen or not, so it's kind of a big mystery around here right now.

That's cool, I...

Nah it sucks, it's terrifying and we're all running for our lives.

Yeah, you've been talking about a lot of systems and features that aren't implemented yet, but the game itself looks complete.

It's not been the most thought-out production. We spent a lot of time on our core aesthetic and making sure the combat is awesome, the look is developed...all that stuff has been online for a while, and then there's a bunch of other stuff that's coming on real late in development that's going to need a lot of playtesting, like a lot of the different missions and a lot of the different systems we're introducing pretty late in the game. We should be at alpha by now, but we're not content-complete yet. It's unusual for a project to go this late before we hit alpha -- it's going to hit alpha and it's going to roll right into beta real quick.

What have you learned from that experience?

Get your senior producer and schedule guy on the project sooner than seven months before you're planning to ship, that's my advice to you.

So you all didn't have a schedule, or a master design document with milestones?

No, fuck, c'mon man! You're talking to an indie studio, we don't have time for that.

Fair enough! Tell me a bit more about what inspired you to create Galak-Z.

I wanted to do something that looked like a classic old arcade game but had the best of modern gaming AI and physics in it. I really like old anime and stuff -- I like '60s, '70s, '80s anime. I don't like the modern stuff, I don't really have time for it, but I love the classics -- Starblazers, Macross, stuff like that. I prefer 2D art over 3D art -- the idea was to have really beautiful hand-drawn effects, so I was inspired to look at some of that classic, Macross-style anime where they have those cool crescent-shaped explosions. Looking at that era of stuff, there's a whole generation of dudes who grew up watching what I did and there just aren't really any games out there that look like that, you know? I think the Super Famicon is the greatest game machine ever. Beautiful colors, cool scaling and sprites, perfect 2D gameplay -- I just love that era of gaming, and I want to recreate the best of that era with the best of modern tech.

How did you come to switch over to pure procedural generation, when the game started out as a hand-built game?

Yeah we actually built only one level, but we built it over and over to fine-tune the hell out of it. We'd always planned to do a traditional Metroidvania action-adventure thing, and I realized we were spending all our effort on the encounters -- the AI, the combat physics, that kind of stuff. And even in the one level we had, it was constantly surprising, because enemies could get into different kinds of fights and the player could sneak past guys so they could lead patrols into a different kind of fight, which could create different outcomes, and we really didn't have any impetus to create more custom content. I realized the game was really more about the interactions, the combat, and noodling around in the world, instead of experiencing tons and tons of custom content. But we were like 'okay, we gotta get on this,' and the way I was doing it was really tedious -- it was a lot of busywork to build levels out of all these individual rock pieces. We started building these procedural level generation tools that would help me bang out big chunks of space faster -- it would kind of build out levels based on a sketch I could draw in the tool. It was a lot lighter on the data side too, because it could just sort of build the level in real time as the game was loading. And there were a bunch of other factors that came into play too. The game was having a real identity crisis for a long time, and then at one point I was just like 'listen, we're trying to solve all these weird problems, we're trying to make it a roguelike and not make it a roguelike...let's just cut all that crap, and try making it a roguelike and see how it goes.' And all of a sudden, we had limitless new content! I'd built all these rooms, and the way the game mixes and stitches them together organically, I mean...all of a sudden we were just having way more fun playing the game. As soon as we had it working it was like, 'huh, there's no going back. This is awesome.'

It sounds like you've switched over to spending the lion's share of your development time and resources on creating systems that generate content, rather than the content itself. How do you feel about that?

It's been amazing! It's of ultimate interest to me now, and I think -- I don't want to go crazy -- but I think this is the future of games. I think procedurally-generated stuff, where you design these algorithms by hand up front, is so interesting from a creative standpoint because you have to break down your material and understand it completely. I can sit down and play for five minutes or five hours and be constantly entertained, and I think it's so much more interesting. I've gotten really into this way of playing games -- it gives you a real appreciation of death. It ties into my nostalgia for the games I played as a kid, where I only had a handful of quarters and if I died too much, that was it -- game over.

How did your experience making Skulls influence the development of Galak-Z?

I mean, we learned a lot about indie development on Skulls of the Shogun. Having been in triple-A for many years, I know that you often don't have a lot of responsibility in those positions -- I mean you have a lot of responsibility for your role, but not the project as a whole. So when I left I thought I knew a lot about making games, but I didn't -- I just knew a lot about my role in the greater production process. So I feel like production has been a lot smoother this time around...

Well, but it doesn't really sound like it -- sounds like you guys are scrambling to get a bunch of stuff working at the last minute.

Eh, that's kinda par for course though. I was talking to a bunch of developer friends of mine and I was saying "oh man, this one's coming in hot" and they were like "man, when is it ever not coming in hot?" I do wish we had another six months on Galak-Z to tune and polish, but if it does well I think we're gonna bunker down and continue working on DLC, extra features, that kind of stuff. The multiplayer stuff just isn't gonna make it for the first release, you know, because we just don't have the manpower on deck, but we really want to do it, so if the game does well we'd love to do it, maybe even as free DLC. We'll see how that goes. But I think overall we understand the space a lot better, we understand how to work with publishers better, that part of the development process -- working with Sony -- has just been super smooth.

So how are you feeling, personally? It sounds like development has been pretty tumultuous.

I see game development a lot like combat. You're in the trenches and you have plans, but some things work and some things don't work. I'm good friends with the guys from Bastion, we all worked together at EA, and they had such great success with their first game that they were able to take their time on the second one. That sounds awesome. I don't want to spend forever on a game, but I do want to take as long as I want, but with a small team and a small budget we just don't have that luxury. We're getting better with every project, and the fact that we'll be able to recycle 90 percent of our toolset and pipeline will make that even easier. Hopefully Galak-Z will do well, and we'll have enough cash to expand.