Making a game in the world’s busiest crosswalk: The story behind 428

Spike Chunsoft’s 428: Shibuya Scramble tells a interconnected tale of six people whose lives are woven over the course of a single dramatic day. Set in the area surrounding Tokyo’s most iconic multi-directional crosswalk — made famous in films like Lost in Translation — 428 initially appears to be a police procedural starring a young detective who hopes to help thwart a case of kidnapping and ransoming.

The case goes awry, though. The day stretches on and the narrative perspective expands to bring in more and more citizens who unwittingly become the keys to stopping a conspiratorial threat with deadly global consequences. 428 starts far more straightforward than Spike Chunsoft’s other adventure games such as 9-9-9 and Danganronpa, but as the game unfolds it proves every bit their equal in terms of intricacy.

The story behind 428 involves nearly as many interesting coincidences and overlapping narratives as the game itself. 428’s history begins three and a half decades ago and connects to the present by way of the world’s most influential role-playing games. It belongs to Spike Chunsoft’s long-running Sound Novel franchise, a set of eight interactive fiction adventures that constitute some of Japan’s most critically acclaimed games of all time.

Despite these plaudits, the Sound Novels have been slow to leave Japan. 428 is the first to have made its way west for a console or PC, preceded only by the under-the-radar release of Banshee’s Last Cry (Kamitachi no Yoru) for mobile platforms in 2014. Japan’s lively interactive adventure genre has carved itself a modest niche in the west on PlayStation Vita, mostly in the form of romance-driving anime games, but 428 feels like something else altogether thanks to its live-action imagery and generally straight-laced crime drama. It’s a rare attempt within the Japanese adventure game genre to reach the much larger audiences that play on PS4 and Steam.

But then, adapting established genres to have wider appeal has always been the stock-in-trade of Sound Novel mastermind Koichi Nakamura, president of Spike Chunsoft.

A strong legacy

Koichi Nakamura has claim to one of gaming’s most impressive resumés: He entered the games industry as a student in the early ’80s, programmed the influential first three Dragon Quest games, and helped wrangle the sprawling computer-centric roguelike role-playing genre into a console-friendly format by way of the Mystery Dungeon series. Yet in speaking with Nakamura, you can’t shake the impression that he takes special pride in his work with Sound Novels. Like many programmers, Nakamura relishes smart and creative solutions to interesting problems, and Spike Chunsoft’s interactive stories definitely have those. There’s deceptive complexity about Sound Novels, and 428 in particular — they may look like simple digital storybooks, but crafting their independent, interlinking plot threads demand rigorous planning.

“With a TV show or movie, regardless of how complicated the plot is, you watch it from the beginning and only travel in one direction,” says Nakamura. “The interesting thing about games is that it’s kind of up to the player to decide how things unfold. We can decide what the plot is, but the story that the player watches — they decide the order.”

Affording the audience this degree of agency comes with many challenges. When it comes to planning Sound Novels like 428, Nakamura says, “it’s really hard to look at the scenario and its individual parts. You can’t make a decision until you see everything come together, so there’s a lot of time up-front working out the scenario.”

This complexity makes itself evident by the end of 428’s first chapter. The game doesn’t allow players to simply pursue a single narrative thread to the conclusion; instead, they can only follow a character’s progress to the point that it hits a dead end. Once that happens, they need to jump to a different story track, where they must explore a portion of an alternate character’s journey. Sometimes, simply walking through another person’s adventure proves sufficient to move beyond the original stopping point, but many times you’ll be forced to make a decision that alters the outcome of multiple tales. For example, choosing to have street tough Achi give a pair of gangsters false directions might seem needlessly rude, but doing so prevents the thugs from appearing in temp worker Tama’s story path and abducting her boss. Keeping Tama’s boss safe allows her to play her own part in Achi’s story further down the road. The game contains hundreds of plot paths and endings that weave together like this. Branching story points can pop up anywhere, even inside of seemingly innocuous lexicon entries.

Still, 428 handles its complexity well, and understandably so. After all, the game represents the culmination not only of years of work in the genre by Spike Chunsoft’s staff but also of Nakamura’s own personal interest in text-driven adventures, which stretches back to the early ’80s. Many American-made text adventures made their way to Japan at that time, but it was the first Japanese-language entry in the genre that caught Nakamura’s attention: Omotesando Adventure. Published by ASCII, Omotesando Adventure took place inside the company’s offices in Tokyo’s Omotesando neighborhood — one train stop over from 428’s Shibuya, as it happens.

Despite its mundane theme of poking around a corporate office, Omotesando Adventure fired up Nakamura’s imagination. “How can I say it?” he says. “I can’t say ‘I can remember the graphics,’ because there weren’t any graphics. But I remember what I imagined it being, because you entered all the commands yourself: ‘Go up elevator.’ ‘Move ashtray.’ There were no graphics on the screen, so you had to imagine everything yourself. The quote-unquote graphics that I imagined for the game I essentially created myself, because I had to imagine everything.

“I got to thinking that it was very interesting that you had to visualize your own graphics. But what would a text adventure look like if it actually had graphics? I thought it would sell very well.”

As it happened, Nakamura wasn’t alone with that notion. “Around that time is actually when Mystery House hit Japanese stores,” he says, referring to Sierra On-Line’s groundbreaking creation for Apple II. Widely regarded as the first graphical adventure, Mystery House combined a standard text parser with crude but effective illustrations of each in-game location. “It wasn’t a localized version, strictly the English,” he says. “But just having the line graphics in a text adventure was revolutionary at the time.”

Nakamura wouldn’t have to wait long for his chance to work with graphic-driven adventure games, thanks to his relationship with publisher Enix. The company had given Nakamura his first big break in games by publishing his action platformer Door Door for Japanese PCs. Enix produced a conversion of Door Door as one of its first releases for Nintendo’s new Family Computer console (aka Famicom, which would come to the U.S. as the Nintendo Entertainment System), and Nakamura was drawn to the potential offered by that system’s broad appeal.

“I thought that an adventure game or a role-playing game would be something that users would be interested in,” he says. “There wasn’t that type of title available for the Famicom at the time. I talked to one of my producers at Enix, [Yukinobu] Chida, and pitched the idea of an RPG to them.”

Again, Nakamura wasn’t the only one to have thoughts along those lines — though unlike with Mystery House, the stars lined up in his favor in this case. “We had another guy working with us at Enix who you’ve probably heard of: Yuji Horii,” Nakamura jokes. The two men would collaborate closely on the original Dragon Quest in 1986, turning the Japanese public on to RPGs. Before Dragon Quest, however, they forged their partnership working in the adventure genre.

“Horii had created an adventure game for an older PC, the PC-6601, called Portopia Renzoku Satsujin Jiken [The Portopia Serial Murder Case],” says Nakamura. He quickly realized that Portopia, which had been quite successful on PC, could be an ideal fit for the Famicom — if they could sort out the technical details.

“The issue with the Famicom was that it didn’t have a lot of [storage] memory available,” he says. “Even if we wanted to create a text adventure with visuals, we couldn’t have that many different unique graphics. But the neat thing about Portopia was it only had something like 20 different [location] images in it. I figured we could probably fit that within the limitations of the Famicom.”

Other practical considerations gave Nakamura pause. Portopia might be able to fit onto a Famicom cart, yet there would still be the issue of the game’s interface. Having been designed for PC, Portopia used a standard keyboard-driven text parser. Though the Famicom did have an optional keyboard peripheral, that was a costly add-on designed strictly for use with a BASIC programming cart. The entire purpose of a potential console conversion of Portopia would be to reach the largest audience possible, so tying the game to a clunky device owned by a tiny fraction of the console’s user base would defeat the whole point. Fortunately, a solution presented itself there, too.

“There was another game in Horii’s adventure series called The Karuizawa Serial Murder Case,” says Nakamura. “Around that time, I think, he put in a command-based system, so the user could just select commands from an on-screen menu. It was a natural fit for the Famicom, because we had the command-based gameplay, and we also had the graphics.”

The Famicom conversion of Portopia proved to be both a creative and critical hit, inspiring a minor phenomenon of detective mystery games for the platform. Flush with success, Horii and Nakamura soon moved along to the Dragon Quest project. Those games in turn built on many of the innovations seen in Horii’s murder mystery games, so it’s only fitting that they would in turn lead Nakamura back to the adventure genre ... albeit through a somewhat unexpected vector.

“I was involved in creating Dragon Quest 1, 2 and 3 at Enix, and they were mega-hits,” he says. “It was like a social movement, in a way.

“I was dating a girl at the time, so I showed her Dragon Quest and said, ‘I made this thing everybody’s talking about. I was involved in making it! Isn’t that great?’ But she wasn’t a gamer, so she tried getting into it and said, ‘I don’t really get what you’re supposed to do. I don’t get what is fun about this.’

“So I started to think, ‘Well, maybe I need to make a game for people who haven’t played games before. An opportunity for somebody to get a controller in their hands and become acclimated to video games.’ I thought about the old text adventure games that I’d played and realized they’re pretty simple. But even then, they require a bit of knowledge about games. If a game is command-based, you as a player have to control the game and tell your characters to do things. Even that’s a bit complicated.

“I thought that maybe I could even simplify it further by having it be decision-based, where you’re just reading the story and it will come to a branching point where it’ll give you a choice: The character does A, B, or C. It’s very simple, but it also gives the player some level of interaction with the game. I figured something very simple like this would be something anybody could pick up, and maybe it would also lead them to playing other games in the future.”

Nakamura took up the director’s chair for this first attempt at simplifying the adventure genre. Called Otogirisou (“St. John’s Wort”), it was developed in tandem with Dragon Quest 5, which would turn out to be the company’s final contribution to the core Dragon Quest series before development duties were taken up by Heartbeat. In contrast to the lighthearted fantasy adventures and comical action games Nakamura had typically worked on to that point, Otogirisou was set in the real world and fell clearly into the horror genre: The tale of a young couple who, stranded in the woods by dark of night, take refuge in an old mansion that turns out to be more than it appears. Consisting entirely of text overlaid atop dark, brooding visuals and accompanied by music to match, Otogirisou could perhaps be written off as little more than a digital version of the old Choose Your Own Adventure books of the ’80s. However, the mature writing combined with the eerie atmosphere created by the graphics and music set the game apart from anything that had come before.

Nakamura feels the ominous horror vibe was a natural extension of the direction he took with the game’s design. “At core, sound novels are a type of visual combined with realistic sound and music,” he says. “What can you do with visuals and sound? What kind of experience or emotion can you make the player feel? I think these two things combine best for a sense of fear. So our early sound novels were horror based.”

Once again, Nakamura found his own inspiration paralleled by others’ work. “At the time [we began development], there weren’t any real horror games. But right around the time I was thinking of making Otogirisou, Capcom created Sweet Home [a horror RPG for Famicom that heavily inspired Resident Evil]. The thing that was really interesting about Sweet Home was that it so scary that you didn’t want to continue playing. I wanted to create an experience where the user would be too afraid to press the button to continue the story, too.”

Otogirisou sold well, and Chunsoft followed up on its popularity with a string of Sound Novels for Super Famicom and PlayStation over the following decade. As technology advanced, the games evolved to match. Eventually, the series broke away from its horror roots with Machi, the direct predecessor to 428. Like 428, Machi was a suspense story set in modern-day Tokyo.

“Machi, in a lot of ways, was the blueprint for 428,” says Nakamura. “It’s a similar set up. There’s a bunch of different people in Shibuya whose stories overlap. The original plan was to have 100 different people walk through the Shibuya Scramble and explore how their lives interact with each other. Where do these people end up going? Of course, that wasn’t very realistic, so we narrowed it down to 10 different characters. It covers a five-day period, broken down into time slots for morning, afternoon, and night. We wrote each character independently — we finished their scenario so each person has their own self-contained story with ups and downs. Once we had each character’s story figured out, then we would decide how the different narrative threads kind of entangled with one another.”

By the time of Machi’s development, Chunsoft had dropped the digitized 16-bit illustrations of its Super Famicom titles in favor of live-action photography. Nakamura stresses that despite the use of live footage and photography, neither Machi nor 428 have anything to do with the failed full-motion video adventure trend of the early ’90s. “When games shifted to become CD-based, the biggest boon was an improvement in the amount of memory to play with,” Nakamura says. “You could have realistic sound, voices, and movie files. I’m sure that a lot of people thought, ‘Wow, this is the future.’ But the thing is, the actual hardware specs didn’t really improve that greatly, so in truth it was more a question of, ‘We have a lot of extra memory to use. What are we going to do with this memory?’ Some people decided to pursue [FMV-based] interactive fiction, and that definitely crossed my mind, too. But ultimately, that’s not where I wanted to be.

“The difference between FMV games and our Sound Novels with live-action elements is — I think a lot of it comes down to development time and budgets. With Machi [...] we did a lot of research into the best way to present the different elements within the game, both the visuals and the sound. This goes back to the beginning with Otogirisou, actually. When you play it, it seems very simple: It’s just text displayed on the screen. It seems like very natural, and you think, ‘Oh, of course the text on the screen would scroll in this way. It’s simple, and the devs didn’t have to do much.’

“But really, we put a lot of time and research into deciding how the letters should appear on the screen, things like scroll speed. Should the letters appear one at a time? Should they come out in chunks? You know, what’s the best way that we can lead the player’s eyes? If we have the text come out from the top and go down to the bottom, the player’s eyes will follow the text as it comes out on the screen so you move from the top to the bottom. When you go to the next page, the text returns to the top. That way, you’re not only looking at the text, but you’re also looking at images in the background and your eyes are constantly moving so it. Even though you’re just reading the text, there’s still a lot of action with your line of sight to prevent a player from getting bored. It also allows us to set the pace of the story. We set the speed and the style at which the characters are displayed on the screen, and that creates a tempo for the story.”

This philosophy defines 428, which creates moods and punctuates its pacing by playing with the size and speed at which text is presented. The story typically advances at a leisurely pace, but sometimes plot twists are delivered in a single burst of large text that flashes on the screen all at once, underscored by dramatic images or sudden changes in the texture of the music. A limited sort of animation — multiple consecutive still frames of photographic images — also helps create a sensation of movement and flow.

“Previous sound novels had used live action visuals before,” says Nakamura, “but I think the first one that did it right was Machi, because of all the research we put into it. Originally, we had literally just been shooting the scenes and then choosing one frame out of it. You now, we’d have the actors act out the scene, record it, then scrub through the video and choose one frame that felt right. The issue with that is that the images we ended up with really felt like you had paused a video and you were waiting for somebody to hit play for the story to continue. It didn’t look right.

“What we tried for Machi was to have the actors assume a pose, then have the photographer go and find the best angle to shoot the character from. With that method, we ended up with a style where each shot looks like a panel out of a manga — very stylized posing and composition.

“Of course, Machi was created before digital cameras were readily available, so we were shooting on film. We were shooting thousands of photos a day, and then we had to go to the lab to develop them, then go through the actual photos and see if they fit within the scene. So you see FMV and think, ‘Oh, they must have put a lot of production cost into that.’ But in many ways, this method of creating still images one-by-one is even more labor-intensive than a video shoot.

“All the time and resources we invest into creating the text and the visuals of our Sound Novels, I think, is what sets us apart from B-grade FMV games.”

The photographic approach used in Machi and 428 presents other logistical challenges, too, including legal hurdles. “Technically you’re not allowed to shoot in Shibuya,” Nakamura says. “You kind of have to go in under the assumption that you’re going to get arrested.”

He laughs. “The way the law works in Japan is that you have to be technically caught in the act, so if your camera crew can run away before the police get you, it’s technically not illegal! There’s an old TV production trick — you kind of set aside members of your staff to run interference for the police and have them get questioned and arrested while the rest of the crew run away.”

These logistical issues do add up. Nakamura admits that all the different factors involved in creating such complex story-driven games crammed with thousands of live-action photos shot illegally on location makes Sound Novels costly affairs to produce — costly enough that the series’ future is in doubt.

“We currently don’t have any plans to do another visual novel,” Nakamura says. “They just don’t make enough money to justify the risk. Take 428. We had to sell something like 200,000 copies to make back our initial investment [in Japan]. On 428, we barely made our money back. Considering the, the cost and risk involved, it’s not very realistic.”

Still, Nakamura seems too personally invested in the series to write off its future entirely. “I like the genre, and people like trilogies, and we already have Machi and 428,” he says. “Could there be a third installment in the Shibuya trilogy? Who knows. If 428 does well in the States, maybe it’ll give me the leverage to create the conclusion of the trilogy. But we’ll see.”