Collected Works: Yu Suzuki
The he veteran developer reflects on his is pioneering arcade games, and nd the completion of Shenmue III
HANG-ON Developer/publisher Sega (AM2) Format Arcade Release 1985 SPACE HARRIER Developer/publisher Sega (AM2) Format Arcade Release 1985 OUT RUN Developer/publisher Sega (AM2) Format Arcade Release 1986 VIRTUA FIGHTER Developer/publisher Sega (AM2) Format Arcade Release 1993 SHENMUE Developer/publisher Sega (AM2) Format Dreamcast Release 1999 SHENMUE III Developer Ys Net Publisher Deep Silver Format PC, PS4 Release 2019
During the 1980s and 1990s, the commercial potential of an arcade game often relied as much on technological innovation as it did on carefully considered game design. This was an ideal situation for the man in charge of Sega’s AM2 team, Yu Suzuki, as he was given priority access to new technologies that allowed him to pursue his interests in simulation. Utilising motion cabinets and 3D graphics hardware, Suzuki created games that pulled off a tricky balancing act, providing a sense of realism without sacrificing the instant gratification so important to arcade success.
As home consoles caught up with the power of arcade boards, Suzuki developed the Shenmue games, RPGs featuring his most ambitious simulation effort to date: fully functioning towns. Though popular, the series was unable to recoup its budget and Suzuki’s focus since 2001 has primarily been on creating the third game in the Shenmue series, a challenge that had often seemed insurmountable. With Shenmue III now finally in the wild, here Suzuki takes a retrospective look at some of the highlights of his 35-year career in the game industry.
HANG-ON Developer/publisher Sega (AM2) Format Arcade Release 1985
Hang-On started with a project brought into Sega, which was the company I used to work for. The technology was a torsion bar, a metal bar that would come back to its original form if twisted. That sort of technology – what if we could utilise it to simulate a bike? If you pulled it down, it would come back up, that was the idea. That was that kind of project that was brought in, but at that point we had only considered the details of the cabinet itself and we needed to add some sort of software to it, which I was in charge of. Ultimately, that torsion-bar technology wasn’t able to be utilised for Hang-On, so in the end we opted to use springs for that motion of the bike coming up.
What I wanted to do was make the kind of game where you straddle the bike
– straddle the cabinet itself to play – rather than facing a table and playing like that. That had never been done before, that was the thinking behind it, but I remember hearing so many negative opinions being voiced. For example, some people were saying that Japanese people tend to be very shy, so they’d never want to straddle this huge cabinet in public. Girls wearing skirts, would they be able to do that? That sort of cabinet is very large, which means cost – that means higher risk, so hearing that kind of negativity being expressed is what I remember from that time. I’m glad that it succeeded, because it was so much fun to do something new, and I had a great time taking on that challenge.
For the game itself, I used 2D graphics to represent 3D calculations. When I was in university, I did research into 3D architecture, so I was very interested in graphics in the first place. I wanted to continue researching, because I was aiming much higher than what I could achieve during my time as a student. So when I joined Sega, I wanted to continue researching 3D technology at the same time I was creating games. But at the time arcade boards could not do 3D calculations, it was impossible. So internally it was 3D, but display-wise, it was conveyed in a 2D environment. We couldn’t carry out development using Sega’s hardware at the time, so we simulated it on a PC, then transferred it to the arcade hardware. Back then what I was trying to do in terms of simulating the movement was just using a line, to depict the way the bike would lean over. The evolution is such that starting with the line, it became triangular; then we could show triangles with colour inside, then the number of triangles multiplied. That’s how 3D developed over the years.
SPACE HARRIER Developer/publisher Sega (AM2) Format Arcade Release 1985
The original concept for this game was based around the Harrier, a fighter jet which can take off vertically and attack both aerial and ground targets. But that game was very difficult to realise, as simulating the way an aeroplane moves
“I’M GLAD THAT IT SUCCEEDED, BECAUSE IT WAS SO MUCH FUN TO DO SOMETHING NEW”
requires many animation patterns. In turn, we’d have to draw more graphics, and the memory would be eaten up. We needed to limit the number of sprites we used, so instead of the plane, we decided to fly people instead. But if we were to fly people, that had to be a sci-fi situation – people don’t normally fly! So what I did was rewrite it as a fantasy game and removed the aircraft completely. The only remaining part from the original concept is the word Harrier in the title. The inspirations behind the game’s visual style are firstly the art of Roger Dean, who did the album covers for Yes. I love that kind of fantasy world. Secondly there was the movie The Neverending Story, something that gives a child a dream. We wanted to pay homage to those two.
In a 3D shooting game like Space Harrier distant targets obviously appear smaller, and it’s difficult to successfully shoot that kind of small target in the distance. That’s why competing games weren’t often very successful. What I thought of was using homing-missile technology, where you lock on, and even if you roughly direct the shot it will hit anyway. If you create that kind of algorithm, the player can shoot much better than they could even in a 2D environment. That’s one of the reasons why it was popular.
OUT RUN Developer/publisher Sega (AM2) Format Arcade Release 1986
Before Out Run came onto the market, the majority of games with cars were racing games where you competed against each other. What I wanted to focus on was making a driving game in such a way that you could enjoy driving as an experience in itself. That meant open-top cars driving in nice scenery, such as Monaco and the south of France. We did sort of follow that racing format, but the objective isn’t getting to number one. In this case you’re the handsdown number one, the second car is way behind and you’re the only one going through the course – maybe you’re only driving with one hand, you’ve got the beautiful girl next to you. That sort of sense of pride and superiority, where you’re number one in the long run, is what I wanted to create.
When I was researching locations for Out Run, I went on a road trip across Europe that was full of unusual incidents. When I was driving in the south of France, I found that at that time the people in the countryside couldn’t speak English. We needed to eat so we dropped by a restaurant, but it’s not like a Japanese restaurant where the menus have pictures, and it was in a language I didn’t understand. But we needed to order something, so I pointed at the menu and said “this, this, this and this,” and selected some at random. The waiter looked kind of puzzled, and gave us a look as if to say “are you sure?” When the dish came, it was just four different kinds of soup. What I learnt from that is that when you order from a menu in a language you don’t understand, you need to space out your selections – choose something from this side and something from that side, otherwise everything will be from the same section. When I got slightly more used to the trip, further down the road, the next time I went into a restaurant I got the waiter to walk with me through the restaurant. I secretly pointed at some of the dishes that the other customers were eating to show him what I wanted. That’s how I found a sure way to order something that I’d really like.
I can’t remember whether it was in France or not, but definitely in a country where I didn’t understand the language,
I really wanted to go to the toilet. So we found a public toilet, but the problem was I couldn’t tell which one was male and which one was female. I was so desperate, but I decided to wait until someone came along so I could tell which one was which – so of course in that moment of desperation, nobody came along for a long time. But that kind of real experience is so valuable when it comes to creating a game such as Shenmue.
Going back to that sense of feeling superior to others, that sort of sense of superiority, the epitome of that is Ferrari. In terms of the brand image, it’s something that all the boys would aspire to, and that’s why I thought it was symbolic in that sense. The Ferrari Testarossa that is used in Out Run doesn’t actually have an opentop version in reality. But when I saw an open-top Testarossa in front of the casino in Monaco, probably one that someone had illegally customised, I said, “Oh, that’s what I need.” VIRTUA FIGHTER Developer/publisher Sega (AM2) Format Arcade Release 1993
Back when we started Virtua Fighter, the president of Sega was always asking the people around him, “Why does our company not have a strong fighting game? What’s going on, why don’t we have such a thing?” At the time I thought that as long as I made one, that would take care of something the company required – I recognised it as something that needed to be done. That’s the real story! The technology at the core of this product was made solely by myself – carrying out the testing using computers and talking to the hardware engineers in charge of 3D boards at Sega, remaining in close communication with them. Initially we could only achieve 300 polygons per frame, but we managed to improve that to 3,000 per frame, or 180,000 polygons per second, by talking to the hardware designers and improving the Model 1 design.
When trying to figure out how collision would work when the 3D fighters made contact with one another, it was all about trying out different algorithms available. When something hits another thing, it’s not a simple matter of two things colliding – one object might get pushed aside, an arm might be knocked that way; so many different types of complicated calculations had to be taken into account, so it was all about experimentation.
For example, if I swing my arm from left to right and there’s a solid object at the central point, there’s a frame before where no contact is made between the two models. If that momentum carried on with two humans, there would be a definite hit. But by the time the next frame is drawn, there’s a chance that the arm has already passed through the other character model and again, there’s no contact. So what I would focus on is the space the arm would have to move through to get from the first frame to the second frame, and checking if an object is present there. Then, even though there’s no collision in realtime, an impact is made so the object must move in the direction of the arm. It wasn’t like it is now. There were no physics engines so I had to manually calculate that sort of thing – everything.
Despite the fact that the character models were very primitive, their movement was very smooth, and I think that gap was part of the reason for the game’s success. From everybody else’s point of view, I think people saw it as something very innovative. That kind of 3D imaging was only really achievable by military simulation companies – at the time, I believe there were only three such companies. That’s why it had a huge impact on the industry.
“THAT KIND OF 3D IMAGING WAS ONLY REALLY ACHIEVABLE BY MILITARY SIMULATION COMPANIES”
SHENMUE Developer/publisher Sega (AM2) Format Dreamcast Release 1999
As you know, my background was in creating arcade games. In arcade games, the duration of play ideally had to be about three minutes, because what you do is you put in your coin – 100 yen in the case of Japan, or quarters in America. If that coin takes you over that three-minute threshold, that means the game’s income will decline, so we were supposed to conclude the game within that amount of time. In terms of monetising the game effectively, there was a need to adopt that kind of shortterm mindset. In consumer games, you don’t have to worry about that sort of time limitation, so no matter how long it takes for the player to play the game, it will be okay. On a home console you can enjoy games that take plenty of time, and maybe games that carry strong messages. That’s why I decided to make a consumer RPG.
It’s difficult to say how I dealt with the pressure of Shenmue’s importance to the Dreamcast as a platform. All we could do was do our best. When it came to working with the huge number of staff that you have to manage in order to make a game like this, because they’re human beings, sometimes it’s very difficult to control them. That definitely presented me with some challenges. But beyond that, it’s difficult to talk about.
I don’t think Shenmue’s Yokosuka is particularly small, but when it comes to designing an open world, to me one of the important keywords is ‘density.’ What’s most important, as opposed to how big an area is, is how dense it is. Of course, if you can pursue the greatness of being big that’s good, but in some cases being relatively small is beneficial – as long as the area is utilised in an effective manner. The best size for an open world is one that’s the right size for the game, that’s what I believe. I applied the same concept of density in Shenmue III, too.
SHENMUE III Developer Ys Net Publisher Deep Silver Format PC, PS4 Release 2019
There were plans to make Shenmue III long before 2015. There were so many different methods, so many avenues we were trying to pursue, but we couldn’t really go ahead with any of them because there were certain things that were insufficient to proceed with development. When I became aware of Kickstarter, I did some research and then I realised that perhaps we could finally start this project, so that’s why I chose to use crowdfunding.
I think I understand now what crowdfunding is all about. In the past, everything about the development of my games has been very confidential, even until three to six months or so before the launch date. Even though the project itself takes three years to make, up until that point everything had to remain a secret. But when it comes to Kickstarter, we need to create games together with backers, so at each stage we needed to disclose the contents of the plan. That was something quite different.
Because we used Kickstarter, our number-one priority was to meet the demands of the backers. In order to please the fans you need to give them some sort of nostalgic element, whether it be through the characters or the gameplay, that reminds them of the existing games as part of the fan service. Having achieved that, of course it’s still a business, we need to gain new users. But if we were to carry on exactly the same way as we did 20 years ago, we cannot expect to expand our audience and gain new fans. The new kind of fan needs more stimulation, so we needed to evolve in that sense as well, in the same way as the movies of 20 years ago wouldn’t work now. We did research with regard to playability, and we found that when the player experienced stressful situations they would stop playing the game, so we needed to change structures so that they wouldn’t feel stress. That was something to enhance the usability, which was there from the basic concept stage.
I’m so pleased that Shenmue III has been completed. You know, there is a time jump system in Shenmue III, and I’ve wanted to use that in real life – I couldn’t wait for that launch to finally happen.