The Shifting Continuum: An Arc System Works Interview

When it comes to 2D fighting games, there are three big names -- Capcom, SNK, and Arc System Works. Capcom's Street Fighter is exacting and relatively precise, while its Marvel vs. Capcom series is controlled combo chaos. SNK's King of Fighters series is the anti-Street Fighter, stringing combos together with fetishistic precision, occasionally crippled by infinites and game-breaking combinations.

Arc, meanwhile, has been going in its own direction with Guilty Gear and BlazBlue, two very popular fighting game series that draw on anime for their look, and a love of the combo for its systems. It's the youngest child in the major 2D fighting game family, combining the combo-heavy nature of MvC, with the cancels and many gauges of KOF, along with its own, very distinctive flair.

Though it's easy to analyze them as building upon what has come before, an exacting attention to detail has ensured Arc's games have a style all their own. Even if the characters were reskinned, any fighting game fan would instantly know an Arc game simply by feel.

The company takes pride in its characters, making them diverse enough to appeal to different players not only in terms of their play style, but also their look and extensively detailed backgrounds. Distinctive characters -- whether you're talking look, feel, or gameplay design -- are its hallmark.

We spoke with Arc system Works programmer Tatsunori Ishikawa and CTO Hiroyuki Masuno to find out what makes the company's games so singular.

What, to you, makes a good fighting game? Guilty Gear and BlazBlue have a unique style, and you're obviously trying for a unique direction.

Hiroyuki Masuno: It's certainly the case that, with titles like Guilty Gear and BlazBlue, we have a certain Arc System Works kind of direction and we aim to pursue that throughout development.

Tatsunori Ishikawa: I think it's part of the producer's goals to take that unique sort of anime "look" and have that reflected in the games -- the idea that you have this corral of anime characters that you're able to personally control.

When it comes to the system itself, what do would you consider the most important components?

TI: In games like these, which you could call "combo" games, control response becomes very important -- that very quick and responsive feedback you get whenever you press a button. That's what enables players to do things like cancel combos and make things feel exciting and fun along the way -- that sense of excitement as you link combos together. That, plus the anime look, is what I think defines us.

With that kind of combo-heavy gameplay, it seems like the game gets increasingly complicated with each iteration. Does that concern you, or do you prefer to focus on the hardcore fighter fans?

TI: No; instead of that, I think we do aim for a system that even light users can take the reins with. "Combo" games like these look difficult at first glance, but like with Capcom's Street Fighter II, once you try them out, they're not as difficult as even some much simpler games. With BlazBlue, we have Stylish Mode, where you can just press the buttons rapidly to perform combos -- the purpose of that is to give beginners a way to better enjoy the game.

BlazBlue: Continuum Shift Extend

Capcom has implemented that, too, especially in the portable versions of their games. Do you think people who use those modes eventually graduate into really digging in to the game's more complex systems?

TI: Well, in the first [BlazBlue] game, Calamity Trigger, you had the Easy Special feature that worked along similar lines, and with [sequel] Continuum Shift, there was an Easy Mode, but with Continuum Shift 2 it's become Stylish Mode. It lets you break out combos by pressing a button repeatedly. So you have the Special button, the Normal button that you have to bash on, the Drive button, and the throw button, and that forms the control system for this mode. That then enables an assortment of moves to be pulled off with a single button.

How do you design a tutorial to get out of this easy mode and help people understand how to get deeper into the combos? That's always been a difficult thing for fighters.

TI: The home version of BlazBlue comes with a tutorial mode, the sort of thing that starts out with how to walk and continues on with how to execute moves, guard, do throws and so forth. There's a popular character named Rachel, and she walks you through it, and the way she's set up allows for her to be a good gateway character for players just getting started with the game. That, and there's also the Challenge mode which lets you test out and learn characters more readily.

So the idea is that if you're just starting and you find all of it a little difficult, you can jump right in with Stylish mode and experience how the basic flow of a combo-based fighting game like this works. That enables you to see the special moves at your disposal, and then you can use that combo mode to step up your game and get into normal play. The goal of this design is to get players up to the point where they're able to take on other real-life opponents.

What do you think of other combo-heavy games? There's Skullgirls; that's very much based on Marvel vs. Capcom 2, but it's another very combo-heavy game.

TI: With that subgenre of fighting games that's combo-heavy, it's very important that the process you go through from practice to being able to do everything is a great deal of fun -- that you're developing yourself. I've only been seriously involved with combo-oriented fighters for about six years, but for me, it's really fun and addictive to go through a well-designed curve like that until I'm at the point where I'm beating players who were at it before I was.

I think part of the fun of Continuum Shift at the start is the player having the ability to just press the buttons to get a taste of all the sorts of moves you can do in the game. He can't beat another human player that way, but against a CPU that's at the level of Stylish mode, he can. It's not fun for the player if he does nothing but lose from the point he starts the game.

With Stylish, the CPU level is down to the point where it's plausible for the player to finish the story mode. Once the player gets used to the game through that experience, then he'll be able to take on human opponents. There's always the impression that combo-based games are harder to learn because it's difficult to see how moves connect with each other, but I don't think that this game is as rigid with those rules as others.

Can you talk about your animation process? SNK does 3D models that they draw over, but Skullgirls does it all hand-drawn.

TI: We work the first way, with 3D -- well, first we come up with concepts for each of the moves the characters would be capable of, then we build those motions based off of that with 3D models. These motions get converted back to 2D, and then we engage in pixel-level cleanup and fixing to come up with what you see in the game.

Do you think that's faster or more efficient than the old hand-drawing type of way?

TI: I think there are cases where doing it all hand-drawn might wind up being faster in the end. Using 3D models, however, makes it easier to maintain an even visual balance across every move from every character. The backgrounds are 3D, too, and thus it's important the characters seem natural against those backdrops -- something that the 3D model approach also helps with. So it may take more time, but I think it's still a more efficient approach to getting better results.

I like 2D art a lot, so I'm happy you've made the choice, but why do you choose to have it be 2D in the end instead of just polishing up those 3D models you make?

TI: I think that's just been a part of the flow of our game history, starting with the first Guilty Gear. It's something that people expect from us at this point. There are tons of 3D fighters these days, but it's not as if the market is demanding every one of them to be 2D. That's the philosophy our producer takes to it, and one of the aims of BlazBlue was to retain the nice things about pixel art while taking the whole package to the next level.

Well, I'm glad that you're protecting 2D.

TI: (laughs) Well, a lot of players still enjoy that style. If you went fully 3D -- like with Street Fighter IV, presenting a 2D style with 3D graphics -- then that's inherently not going to be the same. That same animation style won't work, for example -- there are players who really enjoy looking at each individual frame of animation.

The UI in your games have always been excellent and stylish. It's been the case all the way back to Arc's visual novel games -- is that something that naturally evolved from that era?

HM: That's something we've almost forgotten about at that point. (laughs) The visual novel era. I think there are fewer people who know about that than don't.

TI: It's true that we pay special attention to the UI, to the point where we have artists specializing in that sort of thing. We do make an effort to not just take the simple approach, but to really pay attention to the transfers from section to section and make everything look nice. It starts with the graphic team's requests for the programmers, and then [BlazBlue series director Toshimichi] Mori, at the top of the project, checks that and decides how much of it is practical to implement.

How do you decide when it's too much stuff on the screen? There are lots of gauges and so forth that must be demonstrated, but it could easily become too visually complicated.

TI: There are times when there are limits to what the programmers can do with the art team's requests, but most times they work around it, coming up with special shaders for the gauges and so on. People praised the UI parts of the first game, as well, and implementing all of that was really a major challenge for the program side, but it's something we provide for in the schedule. We try to do whatever we can do, and even if it takes a little more time, we like to see the artists' wishes take physical form without making compromises.

I think one of the things that may intimidate new players is the fact you have so much onscreen to pay attention to besides the actual characters fighting. A newbie could look at that and have no idea what it all means.

TI: I think it's true that you run the risk of making things too complex on the programming level, but on the other hand, if you cut out too much of that, you may wind up making the game less fun as a result. It's true that our game is definitely noted for having a lot of gauges, but I think the prevailing attitude here is that we'd like light users to try their best here. (laughs)

If we got rid of them, then we would get comments from series veterans about how much less data is at their disposal. I do think we're at about the upper limit with the current system, though, so if we add anything else, then something will have to be cut out.

My favorite fighting game ever only has two gauges, Asuka 120%. The life and power gauges.

TI: (laughs) That is a simpler fighter, that's for sure.

Out of all the fighter developers, you seem to be the most interested in creating a story. Why do you put so much story in the games?

TI: Well, one thing we try to aim for in our games is to make the characters as interesting and engaging for players to control as possible, and one way we try to achieve that is by providing a solid background for them. We start with a world setting, create interesting characters to populate that world with, and then we get to designing how they will work in battle. We try to make story mode seem like a sort of adventure game-within-the-game, so that's what led to having so much volume in it and introducing voice acting and so on.

I notice not many people complain about your netcode, unlike with other fighters. How much attention do you pay to this? Is it a plug-and-play system, or does it get coded separately?

TI: With regards to the netcode, we really haven't touched it very much since it was originally developed for Calamity Trigger. We've been able to use that code pretty much verbatim in CS; Arcana Heart uses the same system as well. We spent the time way back then to create a really solid system, and it was our aim from the start for it to be as flexible with different types of fighters as we could make it.

Nurarihyon no Mago

Do you think you'd ever try another four-player simultaneous title?

HM: Like with Nurarihyon no Mago, which we put out on the 360?

TI: Indeed; that's a fighting game that runs on the same basic code as our others. As a result, there's no technical reason why we can't make a 2-on-2 game like that, even though BlazBlue hasn't gone that way yet...

HM: "Coming Soon". (laughs)

[Ed. note: Nurarihyon no Mago is a licensed game published by Konami, and is based on a manga franchise known as Nura: Rise of the Yokai Clan in the U.S.]

When a character changes position in a four-player game, will you do auto-facing, or do you have to do it manually?

TI: Well, with Guilty Gear Isuka, which we released a while ago, there was a button you had to press to manually change your facing. That approach was not particularly well-received (laughs), which shows how important an issue this becomes in a four-player fighting game -- it's much less intuitive than if you only have one opponent.

How has publishing been going for you guys? Do you want to do more of it?

HM: In terms of Nintendo systems, or digital downloads, that's something we try to manage entirely from the Japan office. For Sony systems, there needs to be a local company branch for whichever region you want to sell a game in, which is something we can't do directly. XBLA games can be published directly by us, but the retail software situation is a lot more complex than that.

If it were possible we'd like to be able to directly handle all forms of publishing from the Japan office, but the worldwide console scene is still built out of individual regions, so that's why publishing is instead divided between three companies.

Do you foresee a lot fewer packaged games and much more download stuff in the future?

HM: Yes. I think that's what we'll see, and I think... Well, my fighting game skill has been at the "mid-punch/mid-kick" level for about 20 years or so now (laughs), but I think especially when it comes to the simpler fighters or music games we've been making, the shift is already very prevalent, and that's only going to proceed further.