Cory Barlog: The struggle of making God of War

Cory Barlog has been on a year-long victory tour since the release of God of War, the acclaimed title that has had tremendous success on the Sony PlayStation 4. And his latest stop was at Gamelab, the game event in Barcelona this week.

Stuart Whyte, head of Sony’s London studio and creator of the virtual reality game Blood & Truth, moderated a discussion with Barlog that produced both familiar and novel stories. Barlog told tales about the hard work, self-doubt, and false starts that the project had, as he did in an earlier interview with GamesBeat.

Barlog said that early on, some members of the team wanted to get rid of Kratos. But Barlog fought to keep him. And others didn’t think that Kratos’ companion, his son Atreus, was a good addition. But Barlog wanted the game to be a father-son story.

And the game went on to greatness. The game received a rating of 94 out of 100 on review aggregator Metacritic, and it sold more than 10 million copies. It won Game of the Year at The Game Awards and many other awards.

That was a big achievement since God of War was made at Sony’s Santa Monica studio with a team of 300 people over five years. By comparison, as many as 2,000 people worked on Rockstar’s rival Red Dead Redemption 2 for as much as seven years, but God of War handily beat [the higher rated] Red Dead in awards.

During his conversation with Whyte, Barlog confessed, “I don’t know if I’ve fully processed all of it yet. It’s been a wild year afterward. It’s one of those things where–we had our heads down, working so hard for so long, and there was a point in time where, while we were happy with what we had done, a lot of us still had that possibility in our heads that this wasn’t going to be received well.”

Here’s an edited transcript of their fireside chat.

Stuart Whyte: The first question is, you were game director of God of War. Can you talk about what that role involved and your approach to that?

Cory Barlog: A game director, in my mind, is somebody who makes everybody on the team miserable for the duration of the project. Honestly, the majority of the job is about putting yourself in uncomfortable positions every single day. It’s about not knowing at all what you want to do, but also knowing exactly what you want.

You have to have this straddling balance of realizing that games are incredibly complex. You can have an idea of where you want to go with something, the structure of something, but the actual moment to moment figuring all this out–it unravels over the course of, in God of War’s case, about five years. You talk to people, get them excited. You’re sort of selling everyone a car for two years, because there’s really not much there. You’re just convincing them, “You should try this. This is going to be great. Trust me.” Even when it’s a crazy idea or something way too big initially.

The first year of the game I kept telling everyone the game was only going to be 10 hours. Don’t worry! It’s gonna be small! It’ll be fine. Then I grew into, 10 hours in the critical path and 10 hours in the exploration. Yeah, it’s 20 hours, but don’t worry, it won’t get any bigger than that. And then it was — okay, it’s going to be about 40 hours.

To be honest, everybody was always surprised when the game was finished. “We’re really surprised by how long it was!” I don’t think I’ve ever worked on anything that wasn’t way bigger than we expected. That’s all the way back to working on fighting games at Paradox. Everything seems to balloon when more and more people get involved.

From my perspective, the director is just somebody there constantly keeping an eye on everything, overlooking the entire project. At Santa Monica we do it slightly differently than other studios. I’m working with the marketing groups to make sure the game’s message is consistent throughout, and working with licensing groups to make sure there’s consistency. There’s one view over the entire thing. But I really don’t produce anything during the day, except hot air.

It’s crazy. All I do is talk. I used to animate. I started in animation, and you’d end every day with at least one substantive contribution. I created an animation and it’s awesome. Sometimes I’d create three or four in a day and feel really productive. Now all I do is go to meetings all day and beg, mostly. Please, just do this! Trust me on this! You’re constantly trying to convince people that it’s just a little bit further. Like when you’re low on gas. “Let’s pull over and ask for directions.” “No, no, it’s just up ahead!” I feel like that’s all I’m doing. It’s just up ahead, don’t worry.

Whyte: You mentioned that 30 to 40 hours of gameplay length — Blood and Truth was five to six hours, and even that was challenging to review. How the hell do you approach a 30-40 hour game, constantly reviewing it during development?

Barlog: It’s very challenging, especially when the game isn’t actually put together until that last six to eight months. A lot of the time you can’t even see the entire experience. You know what the whole experience needs to be, but it’s all in broken pieces all over the place. I think it was the Christmas before release in April where we had everything fully assembled, all together, and people could take the game home and come back and say what they thought.

That is weird, because it felt like every God of War I’ve ever made, Christmas was the time everyone would take the build home and come back right before release, and either say really good things or just have a lot of very bad things to say. This was a nice mixture of — there were a lot of bad things, because we had terrible balance. We’d just gotten the economy online for all the upgrading. It was a disaster. But that was the first pass. That’s how it’s supposed to be. It’s supposed to be terrible.

A lot of other stuff, like the camera, the no-camera-cut thing we were trying out — that was something a lot of people didn’t believe throughout the duration of the project. It was a lot of work. To be totally respectful to the team, I put a lot of work on the animation and cinematics group, even though in the beginning I said, “Aw, don’t worry, it’s gonna be fine. It’s not that big of a deal.” It was a really big deal. The AI group, the animation group, everyone had to do all this extra work. We had to invent new rules, new ways to set up cinematics, new ways for characters to navigate the world and hit their marks.

But it paid off. At the end these people came back after Christmas and said, “Now I get it. I get why it was important that we did this.” It just felt different. Actually attacking something like that and testing it — I struggled. At the end I was trying to get people to — I’d say, “I can’t go to meetings right now. I need to play the game.” But you’d get maybe three hours before there was a fire. There’s a fire every five minutes on one of these projects. It’s ridiculous.

Whyte: It’s even more challenging in VR. At least when you’re playing the game, you can see in the peripheral vision someone coming over to ask a question. In VR you’re trapped.

Barlog: When you do character interactions within VR, part of the magic of that is to feel like the person is looking directly at you. That’s not on the performance, is it? You’re doing some extra stuff. What do you do to get that to work?

Whyte: I think we found — VR is such a new medium. I think one of the biggest learnings we’ve found is, there’s a design language that we as an industry have created over the last few decades. Thinking about health systems, locomotion systems, cutscenes, cutaways. Even PC control systems like WASD iterated over a very long time. Some of those things work brilliantly in VR and some of them just do not work at all.

For us it was an exciting time to try and play around with that. One thing was within our drama and our story elements, where the player is the camera. You have no control over where the player looks. You have a scene with seven or eight characters, and the player could just spend the entire time fixating on one character.

Barlog: I kept trying to look under the table when people were talking to me.

Whyte: Bizarrely enough, there was some chewing gum hidden under a table that I found one day. That’s why the game took longer. A lot of attention to detail there. But we had to create some eye- and head-tracking tech, so that when you’re moving back and forth like this, the characters’ eyes keep following you. Their heads subtly move. As soon as we didn’t have that enabled, as soon as you start moving around, you just feel the character is talking through you, rather than at you.

That kind of personal interaction, you’ve gotta have someone looking at you and paying attention to you. It was tech like that which I think you only need to do in VR, but it was super complex to get right.

Barlog: Did that come on early or late?

Whyte: It came together quite hot at the end. I think the whole — games can come together hot. There’s definitely elements where you think this is going to work really well, and then you start trying to implement it. It doesn’t work well. You think, “I’ve gotta keep the faith, gotta keep with it.” Then you try over weeks and months more, doing user testing, and it’s still coming back — but you’ve got to crack this. It’s so important.

For example, our inventory system. You wear a headset. The PlayStation can track the headset and track your left and right moves. But it has no concept of where the rest of your body is. We started to have ammo kits here and holsters down here. It worked most of the time, but if you started leaning forward and reaching down, what is the player going for? It gets very confusing.

Barlog: I found that if I was moving around, I’d think, “Wait a minute, something’s happening!” It’s interesting. All games — maybe there are people out there who do not adhere to this, but every game I’ve ever worked on, and every game I’ve ever talked about, is ugly as hell until it’s not ugly, which is usually in the last six to eight months. It’s usually a disaster. It’s all in pieces everywhere. It’s constantly about keeping the faith. That’s very hard. When you start getting teams that range anywhere from 20 to 300 people — you have 300 individual human beings with different levels of stress and anxiety who are saying, “I’m not confident in this.” That permeates the people near them. You’re constantly spinning plates.

Whyte: And even different — people like different types of games on teams. It’s impossible to make a game that can appeal to everybody on the team. Some people like certain elements more than others. I went to a conference last year, a Sony internal conference, where you talked about the challenge of keeping the faith. You talked about Atreus. Can you talk about the challenges with having the boy?

Barlog: I talk about this a lot, but that was not a popular decision in the beginning. It was, “Hey, that’s cool. Kratos is a dad.” But a lot of people on the team were the old guard. These are people who worked on God of War back in 2003 when I first started there. We were making the original God of War. To them, they were like, “There’s a kid running around? That’s really annoying.” They were remembering games that had escort missions and companions that were not so good.

For me, even though I was aware of those games, I thought there was a really great example in The Last of Us, where they actually pulled off something really amazing, and they did it in a way where — they threw out most of their work six months before ship. That’s who they are. That’s what they do. They realized it wasn’t working, and it doesn’t matter if the schedule isn’t there. Throw it out and start over.

It was actually simpler. They were making something really complicated, which is what we do. When you start at the beginning you’re going to make something that’s the most amazing PhD dissertation in companion AI. In the end it really needed to boil down to a few simple rules, to prevent the player experience from sucking. For us, we looked at that, added a few rules to the ones they had, and really made it around this idea of, let’s make it really simple.

We didn’t have it figured out in the beginning, but we at least understood right at the beginning that it should be simple. We should have a single button. Only one button dedicated. Now that we had the new camera where we could look around, we could use the camera for intent, and then the button for action. If I’m looking at a door, the action changes based on wherever I’m at. It’s contextual without putting icons or anything in there. It allows you to get more immersed in the world.

But even then, Atreus did not intelligently come online until probably September the year before we released. So much went into this. We kept kicking the ball down the field every time. “Oh, this isn’t working. Well, it’s fine, we’ll figure it out later.” That’s a huge thing in development. When something is scary or very hard, well, figure that out later. I realized that somebody has to figure that out at some point. When the people who had to figure it out started telling me they were going to figure it out later, I thought, “Oh, man, I’m screwed.” I’m the one usually saying this. If they’re saying it, it’s not gonna get done.

There was a point where we were going to cut Atreus completely. It was mostly just due to budget, where we were realizing — in the beginning, you look at the problem and you say, “Oh, this is very complicated,” and you pad it. You pad it tremendously when it’s a problem you haven’t solved before. Nobody at the studio had done companions at all. They ended up padding it so much that it was — 2030 was when the game would eventually release, with how much it was going to take to do this. Or we’d have to cut a ton of stuff from the game.

I said, “Guys, we can’t do that.” Finally, we realized we were struggling to hire engineers who could do this. I ended up writing a pitch for a game without Atreus, just as a backup. Thankfully I never had to use it, because we ended up hiring a few programmers. I ended up convincing some people. “Hey, don’t worry. We can figure it out. This will be fine.”

But it’s a lot of begging, being a director, and a lot of dealing with the fact that 50 percent of the time, 50 percent of the team is not supporting the idea you’re talking about. And then another 50 percent isn’t supporting a different idea. You multiply that by a different thing every day. It’s not because they’re malicious or anything. They just don’t see it, and they’re focused on another thing. They’re thinking, “I don’t know if I agree with that.” Or it’s conflicting with something that they want to do. It’s a lot about relationships.

Whyte: God of War was incredibly well-received. You won so many awards. How have you processed that? Coming up to release, it was a hard finish, but following that, how did it feel?

Barlog: Exhausting? I don’t know. It’s interesting. I don’t know if I’ve fully processed all of it yet. It’s been a wild year afterward. It’s one of those things where — we had our heads down, working so hard for so long, and there was a point in time where, while we were happy with what we had done, a lot of us still had that possibility in our heads that this wasn’t going to be received well.

Whyte: You start to lose perspective.

Barlog: Yeah. You get too close to it. You don’t have any sense — because everything is so insular, and we had to keep everything pretty secret, we only had ourselves to talk to. There’s always these champions on the team, people who are just incredibly positive all the time no matter what. They’re good people to feed in. But if you’ve heard them for five years, you start to not believe them either. You get into this negative cycle.

When the reviews came out, when people started talking about it, it was very exciting. But as it progresses on, maybe six months after release, then it’s just a lot of pressure. Then it’s just, “Oh my gosh, what am I doing here?” It’s like that story people tell about Dave Chappelle, where he had done the second season of Chappelle’s Show and it was absolutely hilarious. I think he felt such a tremendous amount of pressure — how do you top that? I guess I’m just going to play World of Warcraft a year. That’s what I feel like doing.

Whyte: God of War was a reimagining of the franchise, kind of a reinvention of the franchise. What were the core elements you wanted to keep and what was the stuff that was more on the periphery, that was less important?

Barlog: We had a lot of discussions in the beginning where we went into a conference room and just wrote every sort of load-bearing concept for the game on the wall. Every possible thing we could think of that represented the game, from Greek mythology to the Blades of Chaos to the double jump to platforming. All this stuff. We listed it out and we went through point by point and started talking about each one. If you remove this, does that make it not God of War?

Even the early discussions, people said we had to get rid of Kratos. A bunch of people were in the camp of, no more Kratos, he’s annoying, he’s done. I had decided from the beginning that I wanted to use him, because I wanted to see if we could make a turnaround with the character. I wanted to see if we could take somebody that people hate, or they like to hate….

Whyte: Less angry.

Barlog: Yeah. Figure out if we could take that backstory and change it a bit. But there were several people who said, “You gotta get rid of this. He’s not God of War. Greek mythology is God of War.” And I thought, “Really?” We had a long discussion back and forth about this. “People associate Greek mythology more with God of War than they do Kratos.” I thought that was a bummer, because it meant we could never leave Greek mythology, if we were tied to that, as opposed to the character.

I think it was just that they really did not like the character. They wanted to see a new character. I think what they really wanted, though, was to see him change. That’s a lot of convincing, to say, “No, I really think this is a good idea. We need to stick with this.”

But the jump, surprisingly, was not a hard thing to get rid of. I was nervous at first, but when I weighed it against the new camera system, the systems guys — Eric Williams is a guy I’ve worked with for more than a decade and a half, probably. He was the one that basically said, “Look, do you want to tackle the camera or tackle the jump? The jump is something you’ll figure out after you figure out the camera, and the camera is going to be really difficult. If you don’t want to cut the camera, you should cut the jump.”

As we started progressing through I would hear people publicly talking about the new God of War and saying, “I can’t believe they got rid of the jump. I’m really disappointed. #NotMyGodOfWar!” For getting rid of the jump? Really? That’s the thing that broke the deal?

Whyte: Probably best to avoid Twitter during development. [laughs] And then you added Atreus. That was quite a departure. You’ve talked in the past about God of War being your most personal experience, with being a father. What did you bring to the story with regards to that?

Barlog: That was a thing I’d learned — I’d been working with a bunch of film directors. After I left Sony in 2007, I was doing a creative walkabout, just meeting a bunch of people and trying to learn. I felt like I’d learned a lot at Sony, but I was very much an infant in my abilities. I needed to understand a lot more.

Some of the people I was talking to were echoing something that Sam Mendes once said. He talked about finding your own way into a story. If somebody else writes your story — if you’re a director who’s taking a script off the shelf, or you’re adapting a book, or you’re brought on a project a bit late — regardless of what that is, find your personal way into the story. There is a connection to every story that is personal to you. If it isn’t, it’s going to read on the screen. If it doesn’t mean something to you, if there isn’t a connection, you’re going to feel it.

You can make a lot of really cool things, but the thing that we aspire to as creative people is to connect to the audience. Make them feel something. That lesson took a while for me to get. It took a few years for me to really sink in. When I came back to Sony I realized that the biggest thing that happened in my life — I was shipping Tomb Raider, and my son was born. I was either going to make a ripoff of Tomb Raider or incorporate my son into this story.

As I started thinking about it, I realized that there’s a lot that changes. Everybody says a kid is life-changing, but your perspective on things changes. It made me think about this idea that — okay, here’s a character that everybody dislikes. We built him as an anti-hero. The whole goal — at the time there were not a lot of anti-heroes in games, so that’s what we leaned into. But it was this idea of giving him a reason to want to change.

He already has a reason, of course. Spoilers, but at the end of God of War III he doesn’t get to die. He’s kind of tormented. He’s cursed. He’s a guy who’s going to walk the earth forever. He started the first game, way back in 2005, jumping off a cliff and not being able to succeed. Spoilers again, sorry. And he ends God of War III not being able to kill himself with the most powerful weapon that’s supposed to be a god-killer. Clearly he’s cursed to walk the earth forever.

Giving him this external motivation to change was very interesting, because I was aspiring to the same thing. At the time I thought, “I’m gonna work a bit less. I’ll have a bit more focus on my home life over my work life.” He was this reminder, as I was working, that we were both struggling with our own lives. We were both motivated by similar things.

That idea, bringing that personal aspect into the game, meant that we could start pulling in from everyone in the studio. A lot of people that I worked with had worked on the earlier God of War games, and were all old now. We all had families. We were all in a similar place. When we made the first God of War we were in our college years, right? We were like, “Screw you, we’ll top everything. We’re better than everyone else.” Now we’re more like, “We want to make something we’re proud of. We want to make something we can tell our families about.”

That gave us the opportunity — so many moments throughout the game are from interactions with various people on the team and their kids, or people on the team and their parents, or people on the team and their significant others. Everyone has these personal stories. As I walk into the studio and have a random conversation, I’m realizing, “I’ll steal this. I’ll steal that. We can incorporate that in the game.” Being open to that is a huge aspect of this job, just being able to say, “This is very interesting. We should use this in some way.”

Whyte: Am I right in thinking that the novelization was written by your father as well? It’s like a full-on….

Barlog: Yeah, it’s a family affair. It’s like the Trump administration. But basically, for me, there was this idea that God of War II — my father and I broke that story first. We were originally writing the script together. The direction I was steering him in was a really bad direction. We kept pushing toward this thing until the story just was not working. It was collapsing under its own weight. We realized that it just wasn’t there. We were fighting, because it’s family. When you write a script with your dad, you’re gonna argue a lot.

Whyte: Did he call you “boy”?

Barlog: [laughs] “Boy” came about because I didn’t have a name for him. That was all because, in the beginning of the thing — I’m very obsessive about names. I can’t put in fake names. A lot of writers will just put in Mr. White and Mr. Pink and Mr. Brown. They’ll write in random names and move on. I can’t do that. I have to have the actual name. Otherwise I don’t want to put it in there, because then I’ll become distracted by the fact that it’s not the right name, and I’ll obsess over that one detail.

I thought, “Okay, we’ll just call him ‘boy.'” Then Chris really got into it. He started throwing “boy” everywhere. Literally every scene we shot was ended with “Come, boy.” To a point where we had to tell him, “Hey, don’t say that anymore. I think we’ve got a lot of takes of that, and it feels like every time we leave, every scene ends with that.”

Whyte: He much get that all the time, walking down the street. People shouting that at him.

Barlog: I’m pretty sure he gets to say it like 10 times a day.

Whyte: You directed the motion capture shoots on God of War. How many days did you guys do that?

Barlog: Oh, wow. It was a big shoot, but we do shoots differently in games. I’m hoping we can figure out a way to make shoots different, because I think shooting a nice chunk, doing a 15- or 30-day shoot to get all of your mocap work done, is a really good way to do it. We shot one day at a time, spread out over like four years. We’d do one day, two days, or three days. It was hard, because by the time you get to a three-day shoot, by the third day you’re really into it. You’re ready to go. Then you stop and you don’t see the actors for six months. It’s also very hard to book everyone. If you have seven people, trying to get all of them aligned on the same three days is a nightmare. That was a logistical nightmare.

I was hoping to figure out better ways to do that, but for me that’s as close as I can stay connected to everything. We did the audio and mocap separately in the early God of War games. I would go into the V/O booth and direct the actors there, and then we’d go and shoot with somebody else, a stunt actor. Then we’d assemble all of it, because most of it was done with hand animation in the game.

Now, in order to get what we wanted from the scenes, to get the characters to actually connect, we wanted to do everything on the set, plus with no cuts, which was way harder than I imagined on the set. I had to deal with it. It wasn’t just this big idea where I said, “Somebody go figure this out!” We had to go on the set and do these seven-minute single-take shots with a nine-year-old. We’re giving him three pieces of his direction and he’s saying, “Uh, okay?” And then I’m here saying, “Do something different at 1:03 of this scene.”

It’s very interesting. My favorite times are on the set, but it’s also the more stressful time. You have to get everything. On the floor, when you’re working with everyone, you try something and it doesn’t work? All right, that’s cool. Come back tomorrow and we’ll mess with it again. You can tinker and finesse things for days. Whereas on the set, just to get everybody in there and turn the lights on for that one day is a cost. If you screw up and you don’t get everything that you need, you have to do all that over again and bring everybody back in. That’s not good. I tried to not screw up as much as possible.

Whyte: From a casting perspective, how did you approach casting?

Barlog: Poorly, I think, mostly. I struggled with that. I had specific people in mind that I wanted to work with, but in the beginning we wanted to see if we could figure out how to do the mocap the old way. We hadn’t decided whether we were going to have a new actor for Kratos. We were going to have performance capture, but we needed a bigger actor in the suit to perform. We were looking into tests as far as how we could modify the animation and mocap data, as well as bring in the audio. Those tests didn’t work out. Then we said, “All right, now we have to find a very large actor that has the voice of Kratos.”

That took two years. I had almost lost faith. When we found Chris Judge, I was getting meeting requests from Shannon and Yumi on a daily basis to say, “All right, let’s talk about this. If it doesn’t work out we’ll have to go back to this. We’ll have to scale this thing down. We’ll have to cut back.” I was very anxious about that. We found the kid so easily. Atreus was found on the second set of auditions. I think he was the first kid that came in on the second set of auditions. I thought, “Oh, this is great, this will be so easy.” And then it was not easy after that.

Whyte: He thought he was auditioning for a movie, right?

Barlog: I think they all did. Danielle, who played Freya, she thought she was auditioning for Game of Thrones. She tried not to be disappointed when she realized it wasn’t Game of Thrones. Chris was the same. He said, “This isn’t a game. This is a movie, right?” I thought, “Should I tell him, or is he not going to want to do it?” How long can I hold the lie until I have to tell him, “By the way, Chris, this is a game and we just shipped it”? But he was very happy. I think his son was a God of War fan, so he was able to get into it.

Baldur was a very hard character to cast. I did a lot of director letters initially. Jeremy Davies, the guy who ended up playing Baldur, was someone I thought about in the beginning, because I’d just finished watching Justified, and I thought he was absolutely brilliant. I thought, “Aw, we should get this guy. But he’s not gonna talk to us.” We had bad experiences with actors on the earlier games. We had contacted several actors that are fairly well-known who came back to us and said they wanted $10 million. Our budget’s not even $10 million, what are you talking about? It was just that they’d had a bad experience on a game, so they didn’t want to do it. That was their way of saying no, just asking for a lot of money. Other people flat-out said they’d never work on a video game.

Other people who I fought really hard for ended up not working out. You get really excited about bringing in somebody where you’re familiar with their work, only to find out that they have some baggage that comes with that. It makes it hard to get what you want. Your experience — you’re bringing in actors that are fairly well-known as well, right?

Whyte: We’ve historically — I guess like a lot of people in the industry, you just hire V/O artists. You get the right voice, put them in a booth, and it’s all fine. With Blood and Truth we wanted to try and scan in the artists and try photogrammetry. We suddenly created a much higher bar from a casting perspective, because we had to find something that obviously had the right voice, but also looked the part and could physically act. Particularly within VR and the motion capture there — because the player is the camera, it’s much more of an ensemble feeling. It’s almost like theater direction.

Trying to get that combination where — again, we’d had some show reels coming in where they had great acting chops, and they definitely sounded the part, but they didn’t look like the kind of character we wanted to put in the game. Trying to those three combinations made it more challenging, definitely.

Barlog: It’s great, because I think VR is helping to usher in more comfort on the game side. I look at actors to be in the game. For a long time people said, “Oh, we don’t need to have the likeness of the actor in the game. We don’t need celebrities. Just bring in voice actors.” I like to have a good balance. There are certain performers where this person is just perfect. I get that they’re a name, but they’ll be perfect for this. You end up battling a lot of game understanding. “That’s not necessary!”

But clearly it’s becoming more accepted when you see something like Keanu Reeves playing Johnny Silverhand in Cyberpunk. The entire cast of Death Stranding is amazing. I think that’s a really exciting aspect. We’re starting to see more of a connection between these industries. There was pretty much a wall for a while. I felt like in the beginning there was not this connection between these two industries, but we’re seeing a lot of younger directors on the film side that play games. They’re more aware of this. A lot more actors are fully aware of what this is. We’re able to give them meaty performances, as opposed to just doing voice.

Great voice actors are a treasure. When you find an incredible voice actor, it’s absolutely amazing. But also being able to have someone who can perform completely, who can take over that character, even in those ridiculous helmets and unitards — it’s very difficult. I don’t think I could do that.

Whyte: Similarly, we went into the Hollywood route with our performance director, Rick Porras. Really early on, we knew we wanted to have a solid story for our game. Typically I would say, as an industry, we’re not great at stories. We partnered with someone who had been a producer and director, worked on Lord of the Rings movies, Contact, Forrest Gump, to come and help with character development story development, which was super useful.

Then, when we came to do our performance shoot, we were like, “We need someone to direct this stuff.” We didn’t feel super comfortable directing ourselves. We had to try and findthe director for our mocap shoot. Then we thought, “Rick knows the characters really well. He directed the second unit scenes in Lord of the Rings. He knows his stuff.” We brought him on board for that. But I think for him as well — within VR, because you have no closeups, no cutaways, it’s definitely challenging. It was great to have him on board and helping bring out the performances from our cast that I think we might not have gotten otherwise.

We were trying to ape an action movie. The whole thing is being an action hero, John Wick, James Bond, John McClane. You don’t need to have amazing complex stories with an action, but they have to be solid. The characters need to be solid and believable to bring you into their world and immerse you. That was super important.

Barlog: Some of the best stories are simple. For me it’s this adage of simple stories with complex characters. Your characters having layers and depth and dimension. They feel like they’re having an interesting arc that they’re going through in this. It doesn’t necessarily need to be a complicated story. The story I kept pitching people in the beginning was, Kratos and Atreus go to the mountain. Spoilers, to scatter mom’s ashes.

Whyte: That’s pretty much the opening cutscene.

Barlog: That actually wasn’t even the first story we did. The first story we did, we spent a year on it. I was trying out something different, where I was going to say, “You know what, I’ll let the writers run with this.” I had a high level idea of what I wanted, but I was going to give them a lot more control. I realized that I’m terrible at giving up control. What it did was it made me farther and farther away from connecting with the story.

They were doing a great job. No judgment against whether the story they were coming up with was a good story. It was just that it happened to be better suited for a later story. We were assuming the relationship between the two characters was already what it was, and we were focusing more on the interesting elements of Norse mythology. It was one of those situations where it’s very easy, I think, as creators to go, “Wow, there’s all this cool backstory and this cool world-building we can do.” You lose sight of the core of what you wanted to do. I had to shelve that story and figure it out, get deep into it myself, so at least I could be connected to it, and then start working with them again.

It was definitely, for me, a huge learning experience. At first I felt like, “Oh, I can let other people take care of this stuff.” But the reality is that I needed to be involved. They were still the ones doing it, but I needed to be, on a day to day basis, getting involved. The connection between the play experience and the story experience is so intertwined. It’s like a soup. You can’t separate the elements out of it. They’re all one thing together.

This idea of someone writing a story over here and somebody making the game over here is such an old way of thinking. It all really needs to fit together in the sense that — it’s not just questioning every decision and saying, “Should this be interactive?” It’s really trying to find the things that drive the character forward in a way that feels like it’s cohesive with the interactive experience.

One thing I was amazed about while playing your game was trying to understand — how do you test for timing? For us, we can understand the pacing between walking and combat and talking and being in a menu. We understand these limits. You don’t want to go over this too much. You want to integrate these things to ensure that you’re getting a good variety and a good pacing. But from a VR perspective, how do you find people’s attention span for story and interaction and combat?

Whyte: One of the things that we’ve — VR is super intense. VR is a really young medium. Compared to the 90 million PlayStation 4s, there are 4 million PlayStation VRs out there. It’s still really new. One thing that we found early on was that a gunfight in a standard console game, which can be intense — if you put that in a headset, it can be crazy intense. And it works really well, but we also found that an intense fatigue can creep in.

My example is, if you’re in a VR experience and you’re escaping a building that’s collapsing around you, and it’s a 30-second sequence, that’s awesome. It’s super intense. It’s a really good payoff at the end when you escape. But if you do that same thing for five minutes, about three minutes into it your brain is just screaming to get out of here. It’s too much sensory overload.

We had to break our combat up. Combat is a key component of being an action hero, but we had to put drama scenes in there. We had to put in object interaction stuff and more peripheral things, which were much lower-intensity, but also really worked well in VR. Just picking up objects and interacting….

Barlog: The ladder stuff.

Whyte: Yeah, yeah. We knew from PlayStation VR Worlds: The London Heist, which was kind of the spiritual predecessor to Blood and Trust, that that stuff was really powerful. But even in the development in Blood and Truth we found problems there. We loved how, in VR Worlds — there was a scene in a pub where you had Mickey giving you a lighter and a phone and a cigar. You could light the cigar. It was really cool.

We kind of think of those objects as being toys. When we were in development with Blood and Truth, we put toys everywhere. You could reach into the environment and pick stuff up and play around. It was great.

Barlog: Were people in playtests doing unexpected things?

Whyte: Well, this is what started to happen. Our third level in the game is a casino. There’s a really light tutorial for a CCTV mini-game. We found that players had found some scrunched-up balls of paper, and there was a basket in the corner. They were getting the paper and trying to throw in the basket over and over again. By the time the tutorial had finished, they’d paid zero attention. Similarly, we had these really intense drama sequences where people would just spend the entire time trying to juggle or throw things all over the place.

Barlog: Yeah, I was throwing the papers back and forth the first time in that interrogation.

Whyte: We had to pare that back, but we knew we couldn’t delete it, because there was an expectation and it was cool. But we definitely — if we’re ever teaching the player something, we’d make sure there was nothing to distract them. You pay attention. Similarly, when we’re in some of the drama scenes or whatever, we try to keep the props relevant to the scene. The opening of the game has a clipboard with the character’s photo on it. You can look at it or whatever and get that vibe. It’s much more in keeping with the scene.

There’s definitely something that became emergent during development. What did emerge for you during development? Was there any kind of feature that you didn’t have planned early on that came out as you were developing the game over five years?

Barlog: Throughout the development of any game, there are always these sparks, things you’re not expecting. The size of the exploration spaces definitely became….

Whyte: It opens up when you get the boat, yeah.

Barlog: There was always a temptation, especially for the Lake of the Nine, for you to really not understand what type of game you were playing for a while. So much so that we didn’t even show it to the press. There was an initial plan to show the press the Lake of the Nine for the demo right before the game was going to release, but then I pulled back and said, “I don’t think we should show them. It needs to be a surprise when they play it for the first time. If we let them play it and show it to everyone, everyone will expect it, and they’ll be playing the game wondering when it’s going to happen.”

That always annoyed me about games, when you feel like you know exactly what’s coming up. Well, when is that coming up? That’s why the Blades were such a big deal, to try to keep that secret. If everybody thought they weren’t in there, they’d let it go from the beginning. They’d embrace what was happening. Then, when it pops up, it’s a great surprise.

Throwing the axe was the thing that was probably the most surprising. We didn’t plan on that from the beginning. I wanted to have the axe in the hand. We’d do a bit of throwing, but it wouldn’t be the full-on hunky Australian throwing things. But as we were getting the core down and I was not letting anybody on the combat team do any of the crazy — I said, “Look, I want to feel like, when you impact somebody, it feels like it sticks.”

As opposed to when you’re playing fighting games — this is a technique we used throughout all the other God of War games and all the fighting games I worked on. You do an attack, like if you’re gonna punch somebody, you’re basically following through. When the hitbox is detected you continue doing the follow-through. There’s a bit of a pause, and then they react. But there’s not a real sense of connection, pushing through, and a change in the animation.

With an axe we had a good opportunity to do something nobody had done before. We invested eight or nine months into really prototyping that. Everybody on the combat team was getting so angry. They were like, “We want to do all the cool stuff! You’re not letting us do the cool stuff! You’re shackling us!” This was just the core of it, keeping the axe in his hands. I knew we could do all the crazy throwing and the magic stuff, but let’s nail this and move on to the next thing.

I think that frustration boiled over until a few of the guys basically said, “Screw it, I’m gonna do this.” They ended up prototyping — George Mawle and Vincent Napoli prototyped the initial axe system, where you could throw it anywhere. It would stick. You could recall it from anywhere. It was a 100 percent programmatic solution. That was probably the coolest surprise of anything.

I would spend every review — we review the levels and walk through the levels. I’d just be throwing the axe. Can I clear this tree? Is there collision on those rocks up ahead? I really annoyed people, big time, by throwing the axe constantly. At one point they said, “We need to take this out, because it’s really getting on my nerves.” When I was testing the earlier games, God of War and God of War II, I’d just jump everywhere. I’d go somewhere and jump, jump, jump. The combat guys were like, “Stop jumping! You’re annoying me!” And the new jump was the axe throw.

Whyte: You found a new way to annoy your team.

Barlog: I looked for ways to annoy different groups on the team in different ways. I don’t want everyone to be annoyed by the same thing. I want to switch it up.

Whyte: The emergent stuff is interesting there. During development of Blood of Truth — we showed the game at Paris Games Week, which was our announcement. We had this cinematic slow-mo happening a couple of times. It was all completely scripted by us. It was when you jumped out of a window, to give you a chance to see what was going on. Also, if you shot red things, red barrels or red fire extinguishers, they blew up and we slowed stuff down.

We saw people playing it at Paris Games Week, and we saw they were really loving that cinematic slow motion. We went back to the office and started trying to iterate as to whether we could make it player-controlled, which is on the two Move buttons in the final game. Once we had that in, it was thinking about, well, what else can we do with this system?

We started with some of the bigger enemies or some of the vehicles in the game. You press down the precision mode, as we call it, to slow down the game world, and you get targets to aim at, weak points. You can even chain to extend the thing. It’s fun how you can explore that idea space and pull stuff out.

Barlog: The awesome thing about games is very similar to television in the way that they’re doing several episodes and then testing what the audience thinks. Getting the real audience, not just a focus group. Everybody’s watching it and they say, “I’m not liking this. I’m liking that.” They’re able to react more. We’re able to do the same thing, getting people to play and then react.

The hardest thing to do in making video games is to know what to listen to and what not to listen to when you’re playtesting. That’s probably the most difficult thing that I’ve had to figure out over the years. You can go very wrong by listening to the wrong advice. Being able to discern between subjective and objective. Understanding that somebody’s not having a good experience and I need to improve the experience, versus somebody just playing this game and they want to shove this thing they really like in another game into our game.

Figuring out how not to get caught in that loop is very hard. I still get caught in it when we do playtests. You’ll hear something that just gets under your skin, because they’re pretty brutal. When people playtest they’re very honest.

Whyte: And also, you’re crafting a story, a story where you get to know characters. Then suddenly you drop people in midway through the game, and they haven’t got that emotional connection.

Barlog: Or if it’s the writers and the animators doing the voices temporarily and the animations are really bad, then everyone says, “This story is terrible!” Well, are they reacting to the fact that it’s presented terribly, because they’re seeing it in a very rudimentary form, or the sound is dropping in and out? One person could not get over the fact that Kratos was played by a different person.

Throughout the experience we were testing in probably four- or five-hour segments. We just had temp voices. Sometimes it was one of the animators. Sometimes it was the writer. It’s difficult to test something that’s not done and fight that urge that says, “Just relax, it’s not done!” Instead of listening and saying, “I get it. They don’t like that. I’m not going to listen to that. But they don’t like this here, and that’s really important. Fundamentally they’re not going to have fun if we don’t fix this.”

Or they don’t understand something. That’s the other thing. We found playtesting helps us a ton in finding areas that we took for granted. We had the idea of Freya giving herself up to Baldur, to say, “Look, if you want to kill me, kill me.” We took that for granted, because it’s a small group of people who thought, “Yeah, for my kid I would do that.” But a lot of the people who were playtesting were 18 or 19. They didn’t get it.

We thought that maybe we’d failed. Maybe this element of the story sucks? But then we playtested it again and it got even worse. People were really angry, capital letters in the feedback angry, about how we’d ruined it. She was a good character. We’d screwed it all up. I thought, “What is happening?” I ended up talking to one of them and explained what Kratos ended up saying in the game, in a conversation. We realized that we’d never explained it. We took it for granted. We assumed everyone would understand that she would rather lose her life and let her son live.

For people who’ve not been in that place in their life, they didn’t understand that. They needed it explained. When we explained it they said, “Okay, I totally get it. That makes sense.” We ended up having a great opportunity for Kratos to open up to Atreus, which we didn’t have at that time, all because we needed to plug a hole that we had left open.

Whyte: I watched the documentary last week. That was no holds barred. You guys were pretty much showing how it was. That’s a brave move. What was it like being followed by a camera crew for five years?

Barlog: It’s weird. It’s weird. Initially we had talked about — I wanted to shoot a bunch of footage when we first started, back in 2013. Let’s just shoot something. We didn’t know if we’d use any of it, but I think it’s cool. I love watching behind the scenes documentaries. I love to see what it’s like in the inception of creating things. The making of The Dark Crystal, if anybody’s seen that movie — there’s an amazing 60-minute documentary that shows Frank Oz and Jim Henson and all of them on these sets using these puppets. It’s just brilliant. To me it’s almost better than the movie. It’s such an incredible slice of that creative life.

Initially, when we were talking about it, I said, “If we do this, we should show the reality of it. Which is something nobody’s gonna let us do. They’ll never let us do this. It’s going to be crazy.” I thought that if we just did a bunch of vignettes behind the scenes, it wouldn’t be worth it. We had worked with Brendan Aiken from creative services at Sony, and he said, “I want to do this. I want to make a full feature.” And I said, “Yes, let’s do this!”

Because our job isn’t hard enough trying to make a game. Let’s make a feature-length documentary at the same time, with a bunch of people who don’t want cameras following them around. There was a lot of resistance from people on the team. They didn’t want cameras all over the place. It took years to get people comfortable. Some people were just never that comfortable with it.

But for me I thought it was important, because I think — it’s an interesting thing to look back on. It’s an interesting way for audience members to understand what it takes to make a game. As much as it seems like it’s about playing games and having fun, it’s not. It’s exhausting. It’s stressful. There’s a lot of doubt and a lot of worry. That’s every project. Everything I’ve ever worked on is always this soup of doubt. You’re just wondering — this isn’t going to be good enough.

I was surprised that we were able to put it out. I was surprised that we finished it. But I’m surprised every day when we finish something. I’m easily surprised.

Whyte: I totally get the doubt thing. I said earlier that creating in VR is such a new medium that doubt is even more accelerated. You make decisions about how you approach your locomotion or your AI or other design issues, and you know that you feel that you’re making the right decision, making it right for the experience you’re making, but you have no idea whether people, when you release it, are going to feel the same way.

Barlog: The pool of expertise is a lot smaller, too. For making a game on a console, you bring people on and they have experience. Even when you’re trying something new, people have experience. But from a VR perspective, everyone’s still figuring it out.

Whyte: One of the things that we knew really early on was that — we definitely didn’t want to make the game realistic. But we came up with this phrase of “reality through an action movie lens.” The whole thing was about — running out of ammo isn’t fun. Having a gun that runs out of bullets really quickly isn’t fun. We increased that, because when you’re watching a movie — you see Die Hard, he’s just shooting the gun forever, but it doesn’t matter. It just works.

We actually partnered with — we found an ex-SAS soldier that came in and really helped us with some of our dialogue. Certainly between the soldiers in the game. Our cast, particularly Amy Bailey, who plays Kayla, who’s a psychotic kind of highly-trained individual — how she would carry herself in a situation so Kevin, our SAS guy, would know not to try and mess with her. He also showed our mocap lead seven different ways that he could disarm a gangster with a gun pointed at him. It was pretty intense.

That was really cool, to bring that onboard. But it was more about authenticity than realism. We didn’t want to have realistic levels of gunplay or whatever. We wanted it to be a game fundamentally. It’s a fun experience.

Barlog: Realism is great. It’s not that realism is bad. But you can have realism and authenticity. To have realism without authenticity — the authenticity is such a huge part of it, getting that feeling of legitimacy. We took a lot of the actors out on some survival stuff. We taught people how to carry swords and studied medieval warfare and stuff like that, just to get a bit of extra information. Even if it wasn’t exactly what they were using, it was this idea that they’re getting little slice of what it’s like. It isn’t so crazy that we’re all learning how to fight with swords and stab each other. But being able to at least understand — even when it’s not the same fighting style, it gives you a better sense of grounding in the world.

Disclosure: The organizers of Gamelab paid my way to Barcelona. Our coverage remains objective.