What's new

Yu Suzuki Talks Shenmue 4, Air Twister, and 40 Years of Game Development

In 1983, video game legend Yu Suzuki started his career at Sega. Now, 40 years later, his newest game, Air Twister comes to consoles and PC on November 10. The arcade-like shooter was originally released as an Apple Arcade exclusive back in June 2022.


“I look forward to more players — old and new — enjoying this classic gameplay experience. I might be one of the oldest game creators by now, but there should be some value in a retro game by a retro creator, right?” Suzuki told IGN with a laugh.

Known for games such as Space Harrier, OutRun, Virtua Fighter, and Shenmue, Suzuki was considered by many as Sega’s equivalent to Nintendo’s Shigeru Miyamoto, especially during the ’80s and ’90s, and rightfully so. He made the world’s first arcade motorcycle cabinet with Hang-On in 1985, created the first 3D fighting game with Virtua Fighter in 1993, and directed 1999’s Shenmue, a game that is seen as one of the earliest examples of a modern open-world video game. If Psy-Phi wasn’t canceled in 2006, he would have also been one of the first to bring touch controls to the arcades. How can one person contribute to the industry in so many different ways through widely different genres? To find out, we interviewed Suzuki at Ys Net, his development studio in Tokyo.


Upon us congratulating Suzuki on his 40-year career so far, he returned the favor by remembering his humble beginnings at Sega. From removing the CRT of a television to attach to a cabinet, to functioning as the weight for a load test of 1983’s Astron Belt, the young Suzuki got to experience every corner of the company. He even was tasked to develop a time card tabulation software application with one of his colleagues.

“I also remember being given a blueprint to make a circuit board,” Suzuki said with a nostalgic smile. “I had to solder it, coil it with a drill, and wire it to complete the structure. They would test the power of these circuit boards, but they threw mine straight into the garbage bin without even testing. It was a bit of a shock for me at the time.”

It didn’t take long before Sega would grant Suzuki his first chance to actually develop a game. In 1984, Champion Boxing made its way to Sega’s SG-1000. When we asked why he made his first game about boxing, Suzuki had to think for a while, but the answer he finally came up with might be a hint as to why Suzuki was able to develop so many different types of games.


“I thought boxing would be easy to achieve,” he recalled. “There wasn’t much you could do with the SG-1000’s processing power, but boxing seemed possible. The character’s straight and jab punches could be portrayed through small movements, and if the feet looked like they were jumping up and down, it would already look like boxing. For the spectators, all we had to do was change their individual color, make it look like their face was moving, and we were done. When taking the SG-1000’s power into consideration, I decided it had to be boxing.”

Deriving a game’s theme from technical limitations would become a theme for Suzuki. In 1992, Virtua Racing — Suzuki’s first full-fledged 3D polygon game — made it to the arcades. At the time, animating the joints of human characters in 3D was challenging, which was why Suzuki decided to make a racing game first. Of course, he was not the only developer to build around limitations, but his shrewd approach allowed Sega’s games to look miles ahead of its competitors’ counterparts in the arcades.

In Virtua Racing, the pit crew can be seen changing your car’s tires. For Suzuki and his team, this functioned as research for how they could animate human characters in their next game, Virtua Fighter. While its blocky polygonal characters haven’t stood the test of time, Virtua Fighter impressed the video game scene in 1993 as one of the first arcade games to boast fully animated characters in 3D. Virtua Fighter was very much a project of working around limitations in its own right. Animating multiple characters at the same time would be unrealistic, but for a fighting game, you only needed two characters on screen at a time.

“In those days, making games was about how to get the most out of the limited technology,” Suzuki said. “In that regard, things are quite different today. Since there are not so many limitations, pretty much any idea can be realized. In such a climate, the originality and creativity of a game becomes more important.”


Suzuki mentioned Vampire Survivors as an example of a video game that doesn’t necessarily make use of the newest technology, but manages to capture a wide audience anyway.

“At its core, a game is a set of rules,” he said. “Within those rules the player can compete, aim for a high score, solve a puzzle, or whatever. As long as the rules are interesting, the game is fun to play, even if it doesn’t have sound or graphics. I can only imagine how much time went into Vampire Survivors' tuning phase to make everything feel just right. It’s not the type of game you can plan. The developers must have played it countless times, adjusting something small each time to get the perfect balance.”

The process of playing and adjusting that Suzuki describes is similar to the development cycles of the addictive games he made for the arcades in his earlier days.

“Back then, it was said that one-third of the development time of an arcade game should be purely for adjusting the balance,” he said. “But if I had one year to develop a game, I always tried to get the fundamentals done in just the first three months, so that I could use the remaining time on adjusting the game’s balance. That means the tuning phase was actually much longer than getting the game’s fundamentals done.”

Suzuki said he feels Vampire Survivors could have been made in a similar way. He looks back at that method of game development with nostalgia, but it was Suzuki himself who moved away from tightly designed gameplay experiences in favor of something on a grander scale: Shenmue.


“For a game like Shenmue, it was not possible to use the same amount of time for tuning,” he said. “My older games all had one theme that I would explore in depth. For example, OutRun is a game about driving. If you have 10 themes and want each theme to have the same depth as before, the game’s scope would have to be 10 times bigger a well. But that’s the challenge we took on for Shenmue. It was a taboo to even try something like that. That’s why we came up with technology to procedurally generate elements. In Shenmue, systems that have the word ‘magic’ in their name, such as Magic Weather and Magic Rooms, were procedurally generated.”

While common in today’s open-world games, Shenmue’s Magic Weather was one of the first systems that would allow for dynamic weather changes, ranging from snow piling up in the winter to blossoming trees during spring. The Magic Room system allowed Suzuki to make it possible for the player to explore the insides of over 1,000 rooms — in a Dreamcast game!

But according to Suzuki, the team used procedural generation in many more ways. It allowed them to automate a great deal of the tuning and debugging phase of development, hence making it possible to realize a type of game that was considered taboo.

“From the smoke of a bonfire to the change of flow of a river, we were able to procedurally generate a lot of Shenmue’s elements,” Suzuki explained. “Theoretically, this kind of technology can compress the graphics data to about one-millionth of its original size.”


At the same time, Shenmue is a game known for its human touch. The first Shenmue has hundreds of characters inhabiting its setting of Yokosuka, Japan. Each character has a unique name, backstory, and appearance. You can see these characters going about their day, all behaving in a manner that seems scripted for that character. These characters speak to the player in fully voiced dialogue, most of them having something unique to say each time the plot advances. Surely, something as human and personalized couldn’t possibly have been procedurally generated, especially for a game developed during the 1990s?

“In that department, there were things that couldn’t be procedurally generated, but also elements that could,” Suzuki said. “Character behavior can be generated by the subject, verb, and objective. For example, let’s say we have something as simple as ‘I go to Yokohama’. ‘I’ can be replaced with ‘you’, ‘he’, ‘she’, or whatever. ‘Yokohama’ can be changed to ‘Kawasaki’. ‘Go’ can be ‘eat’. Next, you can go even further by personalizing such elements. For example, if the subject is an old gentleman, he wouldn’t eat or talk in the way an impolite young man would. If you have three patterns for each type of character and mix these, that alone creates 27 variations. That is what we did for Shenmue’s characters in a nutshell.”

The world Shenmue showed gamers when it first released in 1999 — and the way it was built — was unheard of. Now, 24 years later, many of Shenmue’s features and methods are still common, especially in open-world games. While at the time players were captivated by Shenmue’s scope, compared to modern open-world games Shenmue is much more about density than size.

“I think that the size of an open world makes a game easy to advertise,” Suzuki said. “If people are amazed by a game that allows you to freely explore two square kilometers, it can be countered by making a game of four square kilometers. Next, someone makes a game with 16 square kilometers. The story goes on until someone makes a game that lets you freely explore the entire universe.”


With games like No Man’s Sky and Starfield, Suzuki’s comment perfectly describes the trend of open-world games of the last decade. Suzuki himself finds closed areas just as interesting. In one of my previous interviews with Suzuki, he mentioned the idea of making the interior of a long-distance train the main setting for a game.

“There are good movies that take place in closed areas, and there are good movies set in vast lands,” he said. “The same can be said for games. It just needs to match the story that you want to tell or the experience that you want to deliver. It’s no longer enough for a game to just be big in size or to have many characters.”

For his own works, Suzuki said it’s not as simple as prioritizing density over size. “I believe that the deep emotion and sense of wonder that entertainment can create is to a large degree determined by the amount of change. A steak doesn’t taste as good if you’ve already just eaten one. The same can be said for areas in a video game. If the player has just experienced the hustle and bustle of Hong Kong, next you want them to go to someplace quiet in the countryside,” Suzuki said, referring to the latter half of Shenmue 2.

After a drought of 18 years, Shenmue 3 was finally released in 2019. Due to its nature as a crowdfunded project, Suzuki said that Shenmue 3 was developed to please the fans who made it possible. However, if he gets to make Shenmue 4 (which he very much hopes to), he wants to do things differently.

I want Shenmue 4 to be enjoyable for newcomers.

“I want Shenmue 4 to be enjoyable for newcomers,” he said. “To make that possible, the most important thing is to make it enjoyable without knowing previous events in the story. I don’t think that a new player wants to know 100% of the story. 20 or 30% could be enough. In Shenmue 3, we implemented a digest movie that teaches the player the main events of the previous games, but for Shenmue 4 I want to integrate that part into the main game. It would be great if the player could naturally learn about previous events just by playing the game. For example, rather than watching a separate movie, having playable flashbacks could be a way to do it.”

For a long-running game series, making a sequel inviting for inexperienced players is always a big task, especially for a story-focused series like Shenmue. Once in a similar position, the Yakuza series made a shift by releasing Yakuza 0 in 2015. This prequel became a turning point for the franchise, especially in the West. We asked if Suzuki had thought of a similar approach for Shenmue.

“Yes, I have. I cannot go into detail at this point, but it is something I have been thinking about,” he said. “I think recreating the streets of Dobuita with modern visuals on a new engine alone would make it worth doing. It also connects to the topic of not simply expanding in size. Making an even more detailed Dobuita than the original Shenmue is an interesting idea, especially if it’s not a remake but a prequel with a new story.”


While it doesn’t seem like Shenmue 4 or a Shenmue prequel is officially in development, the possibility alone should be exciting news for Shenmue fans. But Shenmue is not the only thing Suzuki has been busy with.

“Besides Shenmue, there are various ideas for other games that I would like to make,” he said. “And while I cannot go into detail about it today, we are working on a new game right now.”

Air Twister became Suzuki’s first new work after Shenmue 3. While it was undoubtedly a spiritual successor to his own classic arcade shooter Space Harrier, Suzuki said he was delighted to work on a new IP for the first time in years.

From a gameplay perspective, Air Twister plays similarly to Space Harrier. By combining this classic gameplay with visuals inspired by The NeverEnding Story and a soundtrack by Dutch musician Valensia that simply screams Queen, Air Twister is a combination of ’80s and ’90s nostalgia we haven’t quite seen before.

“When developing Space Harrier, we tried to make the game look like 3D as much as possible, but we were actually working with 2D sprites,” said Suzuki. “For Air Twister, it was the opposite. The game is in full 3D, but we strived to make it control and move like a 2D game. In the bonus stage, for example, we completely ignored the gravity and inertial force of a 3D environment. It’s interesting that we are doing the exact opposite to create an experience with a similar feeling.”


Originally released for Apple Arcade, Air Twister was developed with touchscreen controls in mind, and the Nintendo Switch version can be played with touch controls as well. Suzuki told us that the development of Psy-Phi, his canceled touch-panel game that was planned for the arcades in the mid-2000s, helped Ys Net solidify the game’s touch controls.

“The biggest difference with Air Twister is that Psy-Phi was a competitive game.” Suzuki said. “Psy-Phi was designed for a big screen, but in location tests, some players burned the top of their fingers from the friction when sliding their fingers over it. The problem could have been fixed with gloves or a stylus for players to use, but in the end players burning their fingers became the reason for it to be canceled. For Air Twister, both the iPhone and Nintendo Switch have much smaller screens, so please feel at ease,” Suzuki laughed.

Almost 20 years after the cancellation of Psy-Phi, it is interesting to see Suzuki finally deliver a game with touch controls. Perhaps even more fascinating is the fact that many of the rhythm games you see in Japan’s arcade scene today use touch controls on big screens, such as Sega’s own Maimai music game, similar to what Suzuki originally attempted with Psy-Phi. Suzuki said he still sees appeal in competitive games with touch controls.

“I like the intuitive nature of touch-panel games to play against others,” he said. “Actually, for the original Virtua Fighter, I had the idea of a control scheme with many buttons, similar to a keyboard. The player could slide over the buttons all at once with their palm to attack. At the time, touch controls didn’t exist, but the thought process for this control scheme was similar.”

While Suzuki’s influence admittedly isn’t as big as it once was, seeing the legendary creator bring out a new game in 2023 should be pleasing for any old-school Sega fan. Air Twister releases on all console platforms and PC on November 10.

Continue reading...
 
Top