Jenova Chen, the co-founder of Thatgamecompany and creative director of Journey, played a lot of World of Warcraft during grad school. And he always knew that he wanted to make an MMO one day - a form of games that are synonymous, rightly or wrongly, with scope and scale.
And yet when Chen started to make games, the games his studio turned out tended to be small - or at least they seemed small, before you got properly into them. In Flow, you are a tiny amoeba or some such, swimming about in the watery deep. In Flower, you are a handful of petals riding the winds. These games are beautiful, but, they remain compact - nothing like the sprawl of a Warcraft.
Scale is only one aspect of an MMO, though. "What we were taught in school is to push the boundary," says Chen. "Everyone was saying that the future was social games, but the games weren't really social." Chen had seen a few Zynga money-spinners, for example, but while he grasped the game part, the social aspect of something like Farmville didn't seem to move beyond the purely mechanical. You go to your friend's farm to click something, but so what?
And what about Journey? How was a game so sparse - and yet somehow so luxurious - born from a desire to make an MMO in the first place? How did its simple narrative of a desert crossing - enlivened, if you are lucky, by the random players who join your game for one section or another - emerge from the busy factions and cities and battle-plains of Azeroth?
"I wanted to show the world that it's possible to have a game where you are truly emotionally engaged and connected to another person," explains Chen. "That's the beginning. Can we do a Thatgamecompany spin - change the emotional feel - of a multiplayer game? That's how we started."
So how do you get people to engage emotionally with other players in a multiplayer game? This would be the defining question for Journey, from the prototype through to the final release. And the answer, surprisingly, has more to do with what you take out than what you put in.
The first prototypes came very early. "When I went to visit [the studio], people were working on a top-down 2D version of a little game which four people could play at once," remembers Robin Hunicke, who would soon join the team as a producer. The prototype she saw was fairly basic, but playtests were already on the way. And they were already revealing interesting things about the ways that multiplayer games work.
"There were just a lot of dynamics with four [players]," explains Kellee Santiago, co-founder of Thatgamecompany and the studio head during the production of Journey. "It seems obvious that the more people you have, the number of interactions you can have as a group increases." Interactions between players sound like just the kind of things that multiplayer game designers are interested in, but with Journey it was never so simple. Were they the right kind of interactions for the kind of game Thatgamecompany was seeking to make? "It was not leading to that feeling of connection," Santiago says. "Connection and also giving the player space to experience what they were going through with the game."
Competition, or at least playing at cross-purposes, was an immediately obvious problem, "That [early] playtest informed us of the fact that having four people play at once introduces a lot of dynamics, like three-against-one or two-against-two," says Hunicke. Incompatibility came to the surface quickly. "[One of our playtesters] said she felt like she was a slow player and she wanted to explore, and other people she was playing with were achievers and they had wanted to pester her into moving forward a lot faster than she felt comfortable with. And she was like, 'I hope you don't make a game that makes me feel like a slowpoke.'"
Strange as it sounds, maybe there were simply too many players for the game to be truly social? "As a player, you might be experiencing what you are going through, but then you are also experiencing the ways in which you are interacting or not interacting with the group," says Santiago. Something had to give - and a reduction from four players to just two was an obvious starting point.
But once Thatgamecompany started cutting things, it was hard to stop. Hearing the team talk about it, it seems that Journey only started to truly emerge once things were being lopped off all over the place.
Take communication: a necessity for a social game, surely? "We want people to trust, befriend, fall in love and rely on each other in this game," says Chen, transformed briefly, by the act of remembering, back into the pitchman with the PowerPoint deck. "When we first started, our gamer instinct kicked in - we supported chat, we supported thumbs up/thumbs down. We supported all the conventional multiplayer game stuff.
"When we played it, we saw that people started to use thumbs down more often than thumbs up," he continues. "It started to get toxic. When we tried four players, people started to create situations where three player were heading out and leaving the other players behind. That player felt socially left alone. There was a lot of disturbing experiences coming from the playtest so we knew that this wasn't the emotion we were going after. We were trimming off the weed that went away from our goal."
So how about text communication? "The biggest problem about text chat was that consoles don't have a keyboard and to use a controller to type 'Hi, how are you doing?' takes a long time," says Chen. "Voice chat? People hated to hear a teenage boy cursing at them and blaming them for not doing well. Those are the things that were in our way of connecting players."
The solution was an abstraction - something that stood in for communication while allowing none of the difficult anxiety that the online space often creates. "Instead of [all that other stuff], we just turned communication into a ping," says Chen. "When you ping very quickly, you come across as quite urgent. When you ping large, it seems like you are calling. In prototyping and playtesting we found that was kind of ambiguous. People might know you are mad, but they don't hear you cursing at them."
If anything, this ambiguity actually fed into the fun - with a ping, Journey became a game about actively interpreting the player you had been thrown in with rather than simply following orders or giving up and muting them.
"Immediately after the very first playtest, [one of our testers] was saying to other players, 'Were you the blue player? Because you seemed this way,'" says Hunicke. "Like she had opinions of how people had been playing just from watching them move around and call to each other. So we knew that, okay, this happens if you remove all the communication and it's just a kind of puppeteering experience. People do develop ideas about the other person and they feel feelings about this cube that's moving around on screen. Once we get a real character in there, they'll definitely have feelings and thoughts."
With progress visible, the trimming of weeds continued, moving from the realm of communication to interaction. Journey's early prototypes seem to have included a lot of classic co-op material - doors that only open if another player pulls a lever, say. But guess what?
"It wasn't getting to the feeling of really being connected with another person," says Santiago. "So, that certainly led to stripping out the things to do with the other player." What the team ended up with, in fact, was a single-player game that you can simply experience with another player when they drop in alongside you. And that was enough. "It just feels different," says Santiago. "Even when you take away nearly every game mechanic you can to validate having another person there. It still feels different to have another person there."
And it wasn't just co-op puzzles. Journey's movement is glorious stuff, whether you're sliding down dunes and threading between archways or lifting yourself through the sky on magical winds. It always feels great, and it's never particularly tricky. And this was an entirely conscious decision.
"Making the traversal itself the reward was because we didn't want to put a lot of points in or external modifiers that'd make you feel like, 'Yeah, I'm doing great', like a score or whatever," says Hunicke. "We wanted it to be the movement itself that felt delicious and juicy."
But the desire to make game-like puzzles was hard to step away from. While the team was initially working on making traversal accessible and fun, it was also blocking out levels that had a certain Zelda-ish vibe to them. "It was like, 'Okay, climb this thing and jump over this gap and grab onto this piece of cloth and use it as an elevator and it'll pull you up'", says Hunicke. "It ended up feeling very platformy.
"So we had a lot of platforming mechanics that were kind of difficulty based and a lot of it involved having to rotate the camera and move the character at the same time," she sighs. "And the more we progressed in the game and the more we playtested the game with new players and people who weren't necessarily hardcore gamers, the more we saw that kind of platforming was destroying the relaxed and introspective vibe that the rest of the game was trying to build up."
As ever, if it got in the way of connection, it had to go. "Jenova in particular had some really kind of interesting puzzle ideas," Hunicke remembers. "Maybe you're walking along and you come up to a shelter and there's a person in there and then there's only room for one more person. So you and the third person approach the shelter, do you let yourself stand out and be swept away by a sandstorm or do you push the other person out? That kind of stuff.
"But again, as we started kind of thinking about implementing those ideas, it started to feel really kind of pedantic and a little bit too much," she says. "Like they were trying to tell you how to feel. Like the game designer was putting you in a situation where you wouldn't really have a real choice, like it was a false choice. And so we ended up cutting out a lot of that stuff too because it just felt like, well, if we're building a straightforward platformer and we wanted you to feel smart, then we'd build these really hard puzzles and if we're building a narrative game that was really about communicating a really specific story, then we would put these false gates in and you just move in without feeling challenged.
"But we wanted to actually feel connected to the player which meant needing them without having to need them, right?" Hunicke continues. "Wanting them to be near you, but being able to succeed without them. Wanting them to see what you could see from a high vantage point, but not needing them to turn a switch... It really makes it about you wanting that other player there as opposed to needing them there." She pauses. "And I think that's a very, very difficult thing to design from the outset, but it's something we discovered over time."
And of course, if you don't need to talk to the other player of give them the occasional boost up over the classic co-op game ledge, do they need to even have a name?
"I play a lot of World of Warcraft and the problem is that people have very wacky names," says Chen. "If you run into someone with a weird name in the middle of the desert, where you have been travelling alone for two minutes and you finally find a person who is like you, most of the time people are so excited to see another player. And if the first thing they see is 'Ilovehitler' it immediately takes you out of the world and you don't really think about this person as a fellow traveller, you think about them as some very naive young person who doesn't care about other people's feelings.
"We had to trim all the weeds," Chen continues. "Most PlayStation usernames aren't very inspiring. So we told Sony we needed to hide all the PlayStation names and they were like: 'No, all multiplayer games sell better if you can invite your friends. You need to be able to put your friends name on your HUD'." A pause. "Eventually, they played the game, they saw why we didn't want the names."
The more you look for it, the more you see that this creation-by-omission ethos is everywhere in Journey. It's even visible in the main character, clad in brown robes, riding the breezes and slowly becoming a perfect vehicle for the player's changing emotions: stoical, fearful, weary. "There was a period where we had this very heavy kind of almost humanoid character with arms and legs," says Hunicke. "And then we removed the arms because we didn't really want to support arm-based climbing. So then it was just legs and a body, but then we added all this stuff to show you experience and we removed it and went back and forth. And the final design, which is so simple and elegant in my opinion, is a combination of all these factors we wanted to capture.
"This smooth movement of the character being something that you wanted to see when you engaged with the game," she explains. "Wanting to see the character move gracefully and do this kind of air ballet is a huge motivator for wanting to press the stick forward. Like seeing how the character reacts to different surfaces and the way it animates through space by itself. And [when you get two players] they could be with one another, almost like a pair of butterflies or kites in the sky."
On the path to its eventual release, there were many challenges. Thatgamecompany was working with Sony Santa Monica, and dealing with a bigger budget that it was used to - and the attendant pressures that come with a bigger budget. PSN was hacked during development and then there's the fact that Journey was simply hard to describe to people. It was a game about a feeling more than anything. You had to play it to understand it - and to understand, crucially, why it needed a handful of things that seemed like pretty basic elements of the multiplayer experience removed.
Yet when you do play it, when you load it up even now, with most of its players long gone and only the slightest of chances that a fellow pilgrim with drop in with you on your travels, Journey still feels special. In some almost indefinable way, you will find yourself reacting to the empty space of the game, the embrace of the things that are left after everything else was removed. Peaceful as it may seem, Journey is brisk and without filler. It's a game that wants to zip you along to its climax, carried forward not by puzzle mechanics or cutscenes or QTEs, not by boss battles or new skills, not even necessarily by the strangers you might fall in with along the way, but by the soaring score, the bright sands stretching into the distance, and that mountain.
"We really wanted Journey to feel like a place," concludes Hunicke, who admits that one of the team's stated reasons for people replaying the game would be: because it is gorgeous. "Like a place you visit, a relic, like a ruin. Like if you go see the Parthenon or something like that, we wanted it to feel like that. Like there was ghosts in there and there was a sense of respect and wonder and mystery about just being there."