Knack is the first game that Mark Cerny has ever directed, but his industry resume is intense. In 1984, he developed the Atari classic Marble Madness, and later he went down in platformer history both at Sega, where he worked on Sonic the Hedgehog 2, and Sony, where he worked on Crash Bandicoot.

More recently, Cerny has contributed to franchises like Jak and Daxter, Ratchet & Clank, and God of War, to name a very few, as he has taken on consulting work for Sony-published projects.

For the PlayStation 4's launch, though, Cerny got extremely ambitious. He served as the system's lead architect, as detailed in Gamasutra's extensive technical interview. He also directed Knack at the same time he was overseeing the hardware's development.

The game has released to decidedly mixed reviews, but Cerny truly understands the ins and outs of game development -- and in this interview, he shares his insights into game design and production and creating a game that appeals to multiple audiences, while also discussing the soul of the PlayStation.

You've been creating character action games for a long time. Obviously, technical things have changed a lot, but has your approach to them changed over the years?

Mark Cerny: Well, one thing that's changed a lot is we used to make these games for ourselves, and now we make them for gamers. If you go back to how we made Sonic or the first Crash Bandicoot, we'd just sit in a room and build levels, and basically nobody would play the game until it was out on the market... and then we'd find out, usually, that it was so difficult as to be nearly uncompletable. The original Crash Bandicoot is legendary for that! These days we're a lot more in tune with people's play styles, and whether they enjoy themselves or not.

To be clear, I'm not talking about dumbing down a game via focus tests. I'm just saying that these days we do a lot of playtests, find out how people play these games, and then as a team we're free to use that information however we want to.

We can take the Portal approach, which is to say we make sure every challenge seamlessly leads into the next one, because we learned that's how to get people into a groove... or we can take the Braid approach, where the challenge is different every time and requires the player to be in a different mental space every time. Both approaches are great, it's just that I believe we need to make tuning decisions based on an understanding of the player's experiences.

Have you done a lot of player testing on Knack?

MC: Yes! Because games in general are very easy to make difficult, we started with the game on its easier difficulty settings, and studied how non-gamers did with it. After one or two fairly awful initial tests we managed to get to the point where yes, on an easy setting, the game was something that people with no games background whatsoever could really play and enjoy.

Something unique here is that with the testing I'd done for Crash, Spyro, Jak, Ratchet, Resistance, and the like, we could assume the players had some understanding of what a game was. So at the point in time when we had some part of the game done, we'd have a playtest and watch consumers play it.

Usually, the first levels that we work on are from the middle of the game. We try not to start by creating the very first level of the game, because we want the first level to be one of the very best, and we therefore want to create it towards the end of the project when we've learned a lot about how to make quality gameplay.

What that means is that the first playtest is therefore testing the player's experience with content from the middle of the game. All previous tests -- Crash, Spyro, and so on -- had been with gamers, and they handled that just fine. What I found out on Knack, though, is that non-gamers had great difficulty coping with starting in the middle of the game. We couldn't really understand how they would approach the game until we had our first level in place.

I notice that when you're talking about the game, it seems that the difference between easy and normal is quite big, in terms of challenge. It seems that how you're going after multiple audiences. We're perhaps more used to there being a gradation. It sounds like quite a difference. It sounds like you were more confident to tune the normal difficulty for enthusiasts.

MC: Well, we have a lot of experience with tuning a game for gamers. I've found that it's easy to make a game harder, but it's very hard to make a game easier and keep it fun -- so for the specific case of Knack that did mean that we needed to focus on the non-gamers in the early stages of the project.

Do you think it's important to represent that audience on the PS4? We all know who's going to rush out and buy it immediately, but I think taking the longer view...

MC: Well, I think the gamer audience is very well represented by the launch and launch window lineup on PlayStation 4. We have 30-odd titles and most of them are core titles. I just thought it would be nice to have something that would be for the rest of the family. That doesn't mean that Knack is only for the rest of the family -- we developed the game with the dual targets of that lighter, "beginner" player and the "nostalgia core player." I wanted to make something that would work for both.

I find it interesting the approach the game took in terms of progression. It's almost more, by way of analogy, like Uncharted. You have to make it from checkpoint to checkpoint and give it your all between checkpoints. Whereas most Japanese character action games are still in the level-style progression, with a long life bar.

MC: It's very quick. You play some segment of gameplay, and if you fail, you get reset not particularly far back. But the challenges are tough enough at times that you will see yourself giving it five or 10 tries to get past them. In terms of death counts, which I track, I've seen our core gamer playtesters dying anywhere between 200 and 400 times during a full playthrough of the game. 

What you are describing tuning by feel, right? You're trying to find an experience and finding that experience as you create it. 

MC: Absolutely. When you start making a game, you don't know what it will become. And that's why we don't make long game design documents, because we've determined that they are, in fact, worthless.

At the same time, because you are spending other people's money, you do have to have some idea of the scale of what you're creating. And you need to be very sure that the assets you're creating are usually things that you will be able to keep in the final game. 

With asset creation being such an expensive part of game development... 

MC: Ideally, you are throwing out some things, but you aren't throwing out much. If you're not throwing out anything, that means that you aren't course-correcting as you make your game. 

Now, you've worked with Western game development teams and you've worked with Japanese game development teams. Have you found that you are cross-pollinating practices as you move back and forth? How are people responding? 

MC: A lot of our production methodologies for Knack came from Sony Santa Monica. I'd had a chance to work with them on God of War III and wow, do they know how to structure and execute on a project. 

Japan Studio has some very talented individuals, but is a group that had been unable to make a million-selling PlayStation 3 title during the entire hardware cycle. I thought that taking a more structured approach would allow the talent to come forward, so we tried to think a bit about how Sony Santa Monica would approach the development of this kind of game. 

Ironically, during the middle of the project, Allan Becker -- who founded the Sony Santa Monica studio -- came over to Tokyo and became the head of the Japan Studio. 

I always felt like you couldn't have PlayStation without Japanese games. It doesn't mean that every game has to be Japanese, but it's not PlayStation without Japanese-developed games. There was something you said at Gamescom: We want to bring you the kind of games and experiences you remember from PlayStation. 

MC: Well, one of the themes of PlayStation has been the incredible variety of experiences on the platform. I'm sure there are better examples, but I always think back to PaRappa, Devil Dice, and I.Q. on the original PlayStation. 

As for whether having Japanese games adds variety, I agree that part of the diversity comes from people from many different cultures making games for PlayStation. I love that we have our heavy-duty, story-based game coming from Quantic Dream in France, and I love that we have our mascotty game coming out of the Japan Studio. 

If you look at the original PlayStation, it's hard to think of a system with a more diverse lineup, right? 

MC: A lot of that was that games were so much cheaper to make back then, that you weren't risking very much. For many of those games, the budgets were just a few hundred thousand dollars. Even Crash Bandicoot, which was a very expensive project by the standards of those days, cost less than $2 million to make. Now today, if you think about what you can do for a couple hundred thousand dollars or $2 million, it's not that much. 

Over the course of this project, did you feel that the developers there were able to adapt and understand the value of working in that way?

MC: I think, structurally speaking, it went very well. The project is the largest that the studio has created for a home console since, I believe, [2005's] Shadow of the Colossus, and it shipped on time. I think, as we go forward, the question becomes: How can we tap better into the creativity of the Japan Studio?

There's no doubt that Japanese game developers are very creative. What has been, I think, difficult has been keeping pace with technology and production, and also market taste, maybe.

MC: I agree completely. With Knack, because we decided to make "Crash Bandicoot for the 21st century," we were able to have a highly structured project. I now wonder what these same individuals can do with this new better set of production methodologies.

This is the first game you've directed, or have had main creative control over, in a long time. 

MC: It is the first game that I've directed. 

Ever? 

MC: Well, yes. We didn't have directors back in the day, because if you have a three-person project you don't worry about calling somebody a "director." And as team sizes got larger, I tended to be in a very secondary role supporting creative directors such as Ted Price or Jason Rubin. I was happy to do it, because they are brilliant. 

But were you happy, also, to take the reins?

MC: I took the reins on Knack not because I had a burning desire to be director but that I had come to the limits of how I could work as a consultant in video games. The team sizes grew dramatically over the years… 

Crash Bandicoot was created by a seven-person team, Crash Bandicoot 2 was closer to 20 people, and by the time when I was working on Killzone or Uncharted the teams were well over 100 people in size, and I was still this part-time consultant. What can I even do with that time slice? 

So with Knack I had the idea that if it was a smaller title -- "Crash Bandicoot for the 21^st century," for example -- and if I restricted my role to, say, being a producer, that I could still do something with that smaller time slice and add a lot of value to the game. 

Now what happened is [laughs] I ended up being director rather than producer, and the game stopped being a very small title, both of which dramatically increased the amount of time I need to spend on the game. 

Was that because it felt right? 

MC: What we found is that we needed to make a more complex and involved game than Crash Bandicoot if we wanted it to be a success in the 21st century. Crash Bandicoot had only two minutes and 15 seconds of narrative, and it used just two buttons. Even if the vision is to be in that sort of genre, to try to speak to the nostalgia that people have for the experiences of years gone by, it turns out you need to do five or 10 times as much today. 

When you tune the game -- you talked about doing playtests, and you talked about the game changing as you get a sense of what it is -- do you still do that more by feel, or do you do analysis, metrics, or anything more crunching? 

MC: We try to make a game that we believe in. We ask ourselves, what is the world, what is the gameplay, what is the story? But at the same time we want to make sure that it's playable and enjoyable for a very broad variety of people. So during the tuning of the game we try to reach out as much as possible to the game playing audience. 

At a playtest, I think if you ask the question "what should the storyline be," you probably won't get much of an answer... Frankly if you ask that question today, they'll describe Grand Theft Auto V to you, because that's what they're playing right now. So we have to make the game we believe in, but we also have to focus on how it is played.

How do you personally feel about the game and what you and your team achieved, and why?  

MC: I'm very proud of the team for creating an original launch title. It takes extraordinary effort to be out on the same day as the hardware and to do with a new brand speaks to the dedication of everyone at the Japan Studio. 

Were you anticipating the kind of mixed critical response you got?  

MC: There was definitely a very mixed response to the title. On the plus side, there is some appreciation of the core concept of a creature made from parts -- at the same time of course one can see the desire that this concept had been explored further. 

What do you think lead to it?  

MC: Expectations are very high for next-gen titles, and rightfully so! Everyone is looking to see how that factor of ten improvement in performance, or the more connected nature of gaming, is going to change or enhance the core gaming experience.  At the same time, it's pretty clear that the titles that unmistakably take advantage of that performance or connectivity are still well in the future. 

Have the challenges inherent in launching a new IP, and anticipating what audiences want, intensified in recent years? 

MC: Definitely the bar is very high, which speaks to the incredible talent in the games industry. Existing brands such as Grand Theft Auto continue to impress -- and though with new IP you have more of an ability to surprise people with the direction you take, you still will be launching alongside the latest version of Assassin's Creed or the like. 

Personally I'm looking forward to Ubisoft's new two IPs, Watch Dogs and The Crew.  I think the drop-in drop-out nature of The Crew will really work well with the social aspects of next-gen.