When working on large open-world games, the huge number of people required often means there are few devs with intimate knowledge of what the whole game looks like.

In a lead-up to his talk at GDC 2018, we at Gamasutra were joined by Guerrilla Games lead designer Eric Boltjes to discuss his work on Horizon Zero Dawn, a game that had to lead players into the wilderness of a post-apocalyptic world without getting them lost in the woods (metaphorically speaking). 

Since several of us at Gamasutra were fans of Guerrilla's science fiction adventure, we were eager to ask Boltjes about a few topics that wouldn't quite make it into his GDC talk. You can watch our full conversation up above, but in the meantime we've got a few nuggets about the game's enemy encounters, stealth, and item economy for your perusal down below. 

Stream Participants

Bryant Francis, editor at Gamasutra
Alissa McAloon, contributing editor at Gamasutra
Eric Boltjes, lead core designer of Horizon Zero Dawn at Guerrilla Games

From Helghast to Thunderjaws: Making enemy encounters matter

McAloon: You have a lot of experience with Killzone, and you said that this was kind of like the opposite of what you see on paper for Killzone. Was there any specific design that you did with the Killzone series that was translated into this game in a subtle way?

Boltjes: I don't think anything directly got translated into Horizon. You can maybe, I don't know if you've played Shadowfall, but there was already [something] up there, that we wanted to, instead of being a very linear experience, to have more choices and options for the player, and it was going towards that feeling of, in Horizon, there's no one solution that's the key solution, the main solution.

There's lots of options, and depending on what kind of player you are you can tackle these options in different ways. We started experimenting with that in the last Killzone, but we really went all-in in Horizon in the end.

Francis: What has your experience been designing encounters in Horizon that are more open? Say an encounter is a circle on the map, and the player can literally discover the encounter from any direction. What's the difference between that and designing encounters for Killzone, which put players in a first-person perspective on these more linear maps, that gets fixed perspective on an encounter, versus this which is more open?

Boltjes: It's a very different approach. One thing is understanding how the robots work and how they react. In Horizon all the robots have different functions. Some of the robots are what we call transport robots, they transport resources throughout the world. They behave differently depending on what the player does, like run away for example. So depending on the type of robot you have in an encounter, the encounter space needs to be set up differently.

For example, the Trampler, they need open spaces and get-away areas, whereas if you fight something like a Thunderjaw, it's set up differently because that thing isn't going to run away, that thing is going to try to kill you. Also, the size of the robot determines the kind of encounter space we have to design.

I think the second part that's really important is that we have to look at all the options that the player has, depending on which phase of the game they're in. Early on there's some tactical options, but not all yet. In the middle of the game you have more, and at the end of the game you have all of them. So the encounter spaces, as you can tell if you've played the game, they're set up so that they develop over time alongside the player.

You'll notice, for example, with the Tripcaster, as a weapon, you can experiment with that early on in the game, but later in the game there's a lot more space in which to tie down the robots and actually have an effect on it. Another example would be having more height variance. So in the beginning the encounters are much more flat than they are later on the game, with the height variance it takes more skill to use. Designing encounters, you have to take into account a lot of different aspects for it to work well, and in the end if you look at players playing through them, there's still surprises sometimes. "Wah! That's a solution!" That was the goal, to give the player the power to choose for themselves which solution they like the best.

McAloon: One specific thing I noticed when I was playing, there are hunting lodges placed throughout, where there are different challenges there. Using that as a unique way to teach you different ways to use this equipment that you're unlocking, and to try new strategies with the machines and teaching you about the different behaviors they have. Can you talk a little about those, as a tutorial?

Boltjes: Sure, yeah. Early on we were trying to figure out, so we have all these tactical choices the player can do. How, in a fun way, can we expose them? We looked at the lore, together with the narrative writer and the writing team, and asked, what can we do? They wrote in the whole idea that some hunters need to show their skill and prove their worth. Especially with a character like Aloy, that felt really relevant, because she's an outcast, she wants to prove herself, and at the same time that kind of character and narrative progression really goes in line with the player as well.

As you unlock the hunting grounds, we thought it was a great platform to then showcase some of the deeper tactics. It wasn't always as easy though, some of the challenges we put in there are difficult on purpose, but what we ran into early on was that some of the challenges were also easily cheesed. People would interact with them in ways that we weren't really expecting and cheat through them almost. That helped to find more bugs, but it also made it more difficult to make the hunting grounds bulletproof.

Don't let players "fail" stealth, but don't let them hide again at the drop of a hat

McAloon: Can we talk about some of the challenges with the stealth-based systems? Anything really relating to enemy NPCs becoming aware of your character or anything having to do with tweaking to work with those hiding systems?

Boltjes: Stealth games, if you look at the full stealth games, there's a lot of things in there to make it a full stealth game. For example, about re-engaging, let's say you're hiding and get spotted. That's sort of a failure. In a lot of pure stealth games, it's either relatively easy or really readable to get back to re-stealth. That's what the game wants you to do, it wants you to hide again and try again.

Horizon is not a full-stealth game. We figure that part of hunting is being stealthed, but ultimately we wanted people to get into fighting the robots. So the difference there is that re-engaging with stealth is not as apparent as in other games. I wouldn't say more difficult, but it's also not... it's almost like you're preparing for an encounter, and once you're engaged with it we want you to play through or completely disengage with it.

Tweaking the values, again, I think it goes back to that Horizon wanted to be accessible, so once you're in stealth and none of the encounters have kicked off yet, then we're quite lenient with you not being spotted. If you edge it to the edge of the stealth areas, you can see that it's quite forgiving, you can have your arm sticking out and they don't see you, that kind of stuff. But at the same time, once you are in combat we're a little bit less forgiving, and re-engaging your stealth is trickier. Does that answer your question?

McAloon: I've very bad at stealth, so when I would be hunting down some of the deer-like machines, and they'd spot me, they'd run off, and that would give me a chance to re-evaluate my strategy. That feels like some of that forgiving element that you were talking about there.

Game economies (like real economies) are tricky to test

McAloon: Can you talk about maybe how the crafting system has evolved, since the early iterations versus now?

Boltjes: The problem with the economy is, usually, that it's one of those things that you can design on paper. Until the entire game is there it's too difficult to test. Even when it's there, you need players to spend thirty hours to really test the limits of your economy, if you know what I mean. It's easy to test the first ten hours of your game to see if the economy works, because everybody goes through that you have loads of data that you can see what's right and wrong about it. But when you stretch that to fifty to a hundred that becomes really difficult.

The balancing of that, especially in the last year-and-a-half, almost two years of the game, that was the tricky bit. Imagine that you're doing a playtest and people are being told, you have five days, you're going to play seven hours a day, and beat the entire game, because we want to see the quest progression and all that kind of stuff, and see how you react to the story, but at the same time you want to test the economy. So this is actually not relevant data, because people are focusing on the story and are not actually going into the world as much as you would expect.

So you're getting data about your economy, but is it correct? You don't know. So a lot of time is spent trying to analyze, were things too expensive? Too cheap? You get different feedback almost every day. People complaining, this weapon, I can't afford it, I can never afford it, then people say I got it after like three main quests, what's the problem? So I think, in terms of how it evolved, especially that end bit, the last two years, that was the balancing of it.

For more developer interviews, editor roundtables, and gameplay commentary, be sure to follow the Gamasutra Twitch channel.

And if you'd like to hear more from Boltjes, you can watch his postmortem of Horizon Zero Dawn from GDC 2018.

Gamasutra and GDC are sibling organizations under UBM Americas