Monday, January 14, 2019

Interview with Yu Suzuki on Shenmue 3 [Gamersky]: Full Transcript - Part One

We recently posted a translation by yuc02 of a report summarizing an interview held with Shenmue III director Yu Suzuki and animation producer Hiroaki Takeuchi at the recent G-Fusion Tour event in Xiamen, China.
    Today's post, kindly translated once again by yuc02, is a full interview transcript, carried out by

    The translation starts below. Part Two to follow in a separate post.

    Shenmue 3 Director Yu Suzuki Interview - Fighting System’s Complete Revamp [Part 1]

    Original text from

    Following WeGame’s invitation, the author had the privilege to meet with Shenmue 3 director Yu Suzuki (YS) and animation producer Hiroaki Takeuchi (HT) in Xiamen, and with fellow journalists conducted an interview about the state of the game.

    Shenmue 3 director Yu Suzuki (left) and animation producer Hiroaki Takeuchi (right)
    Shenmue 3 director Yu Suzuki (left) and animation producer Hiroaki Takeuchi (right)
    Q: Apart from the location of Shenmue 3 being set in Guilin, what other aspects of Chinese culture will be represented? The first two games in the series contained the Chinese martial art Bajiquan; will Shenmue 3 introduce more martial arts styles?

    YS: In my early game development years I visited China numerous times, and learned Bajiquan from Master Wu [Master Wu Lian Zhi], with whom I still keep in touch and meet every year, although not on this occasion. I am quite familiar with China, and have visited Xiamen’s Kulangsu for cultural studies, and learned about Chinese architectural styles for the game. We did not carry out any studies on new martial arts, but have incorporated other Chinese styles such as Xingyiquan in the game, although these will not be part of Ryo’s repertoire; rather they’ll be styles owned by other characters.

    Q: Shenmue 1 and 2 already had a great battle system, with many moves being faithful representations of real martial arts. So how will Shenmue 3’s system improve on this?

    YS: Shenmue 1 and 2’s battle system used data from the Virtua Fighter series. However the battle system of an RPG such as Shenmue was still quite different to that of a pure fighting game. Pure fighting games require precise timing and reactions, but for Shenmue we used a more basic system than Virtua Fighter, such that the controls were not too difficult, which was more suited to a story-driven game.

    For Shenmue 3 we have moved away from that system, and have created a brand-new battle engine. The system is very different; richer and more playable than before.

    HT: The battle system is completely different. For fighting games (such as Virtua Fighter), satisfying hardcore gamers very proficient in reactive control can be achieved simply by setting the controls' difficulty to a particular narrow range. However for games that are aimed at a broader player base (such as Shenmue 3), this difficulty range has to be widened, such that experts as well as more novice players can enjoy the gameplay. This aspect requires much care and balance during design, and we have spent a lot of effort on this.

    Q: Previously it has been said that Shenmue will include “Advanced AI Battling”; how will this system handle the difficulty during a particularly-tough boss battle for example? Will this system aid those who are not very good at fighting games in order to defeat this boss?

    YS: As we have said, the previous battle system could only satisfy players of a particular proficiency, because the difficulty was set to a certain range. With this new AI assistance, more players will be able to match this difficulty level, so even those who are not good at fighting games can succeed.

    Q: Shenmue 1 and 2’s battle scenarios are generally born out of plot events, and many gamers feel there aren’t enough of them. Will Shenmue 3 have more battle scenarios?

    YS: In my opinion I don’t think it will be better simply by having more battle sequences. I am striving for better playability, not more fights.

    Q: What mini games will be featured in Shenmue 3? Previous games had forklift driving and capsule toys, which were very addictive.

    YS: Shenmue 3 will have more fun mini games [laughs], including the return of capsule toys, pachinko (Lucky Hit), and dice games. Some previous mini games have been removed, but there will be new ones too.

    HT: [Laughs] I was actually one of the forklift drivers in Shenmue 1.

    Regarding Shenmue 3, we also have updated the mini games such that one does not simply play them in isolation, but there will be some interesting effects upon winning prizes, which are more integrated with the rest of the game.

    Some of the mini games, when completed, will also have consequences on the main story. Shenmue 3’s mini games will every now and again throw up “pleasant surprises”, for those who persist in playing them for a long time. I hope players look forward to these in anticipation, and I very much look forward to hearing their feedback.

    Hiroaki Takeuchi (left) is represented in-game in Shenmue I as a forklift operator at New Yokosuka Harbor
    Hiroaki Takeuchi (left) is represented in-game in Shenmue I as a forklift operator at New Yokosuka Harbor.
    Q: In last year’s (2017) interview you mentioned that Shenmue 3 contains about 70% of design ideas from the previous games, with 30% new content. What do you feel are the main differences or highlights of these new ideas when compared to Shenmue 1 and 2’s? Do you think that this game’s system, pacing or feel is significantly different to the earlier games?

    YS: If we talk about the content split, with the development process we are now roughly at 50% new content. Shenmue 3 is now nearly 20 years ahead of the first game, and in this period hardware performances have leaped forward immensely. To adjust for this advancement we have made many updates, especially in the visuals, which have progressed significantly.

    Shenmue 1 and 2 were adventure games having a main storyline with various scenarios and battles inserted, and this structure remains the same with Shenmue 3. However, the mini games and side quests etc. did not influence the main story much, although at the time we wanted to strengthen this influence but could not achieve it. So for Shenmue 3 we hope to achieve this stronger link.

    For example, move scrolls and mini games such as capsule toys will more closely tie in with the economic system of the game - be more integrated - and we have spent many hours designing this with a lot of success.

    Shenmue 3 is a more in-depth game compared to the first two, with more replayable elements. One may be able to 100% complete the first two games in one playthrough, but as Shenmue 3’s replayability and freedom has increased, one may need to do multiple playthroughs.

    Q: You mentioned multiple playthroughs, can you further elaborate on this?

    YS: What I mean is that after the main story has been completed, the other content can continue to be played. There are many parts of this open world to be explored besides the main story.

    Q: The Shenmue series is renowned for its recreation of reality, and Chinese gamers really enjoyed locations such as Hong Kong’s Kowloon. How did you go about recreating a location such as Kowloon? It’s now been more than a decade between the original series and Shenmue 3 and there this game has new locations; did you face any new challenges when recreating them, and have you used any new techniques?

    YS: Recreating a location doesn’t mean copying it as-is. Kowloon at the time had a lot of slum housing, which wouldn’t be aesthetically-pleasing if we had recreated all of it as it was. What I wanted was a “reconstruction”, such that players can instantly recognize the location style to be Kowloon.

    We used advanced technologies from other industries while developing Shenmue. One of the major breakthroughs of Shenmue was that it was possible to enter every room, which we called “Magic Rooms”. There are a vast number of rooms, however due to technical limitations at the time, we couldn’t pre-save all the data into memory, otherwise just this data would take many discs to store. To solve this problem, we introduced the “auto self-generation” concept which using an AI system based on cutting-edge technology from the construction industry, and could create 1200 rooms in 0.4 milliseconds. Although each room’s materials, layout, wall design and lighting etc. are all generated in real-time, they need to follow common sense and logic, therefore we found help from many experts in the fields of architecture design and construction to create this system.

    After so many years the world has changed a lot, and now game data sizes have expanded massively, along with much better computing technology and hardware. Although many of the limiting factors of the past have been removed, I still like to challenge myself and experiment with different technologies.

    To be continued in Part 2, coming soon. Many thanks to yuc02 for the translation. 

      Become a Patron!


      1. Replies
        1. Cheers Jake! Yuc02's translation of the second part is on its way :)