Hitman Absolution director Tore Blystad faces a lot of challenges with the latest game in the franchise. The team has built an entirely new engine, Glacier, to power it. The believable environments it allows the team to craft, ironically, make the game - out later this month - more difficult to develop.

When your game is about sneaking through a realistic world, the things that aren't realistic stand out just that much more. And as all game designers know, there's another source of tension there -- because what's realistic is not necessarily fun.

In this new interview, Blystad explains how he and the team at IO Interactive have approached everything from players' expectations to be spoon-fed a completely linear play experience, how they've dealt with the challenge of making realistic but not too realistic AI, and what "immersion" really, in the end, means.

Has it been hard to keep the franchise fresh for you, as a developer, over the years? How many installments have there been? Five?

TB: This is the fifth one.

That's a lot of games.

TB: Yeah. It's been quite challenging, also, because as an artist producing the Hitman games, it's a different type from other games, because it's usually civilian scenarios of different kinds. Because military bases and these kind of very heavily guarded areas, it becomes less of a challenge in a game where you're choosing so much what to do by yourself. So coming up with these different scenarios, and different kinds of unusual ways of approaching a kind of familiar area, is something that is increasingly difficult for us, since we've gone through so many in the past.

When this franchise started there was a lower level of character nuance expected out of protagonists. At that time, you could pretty much just be a bald guy with a gun, and that would be the beginning and the end of what you needed to convey. Have you found that that's been a challenge?

TB: Yes, absolutely. That's why I would say there's probably 10 times more work going into this game than any other Hitman game in the characterization -- not only of Agent 47, but also of the other characters in the world. And since the AI is so, I would say, "complex", it has to be able to react to anything you can do as a player. Every character has to have... I don't know, there's like hundreds of different responses they need.

So just the complexity of that, and recording this, with like two thousand pages of dialogue for the game... it becomes a big logistic problem as well, because it's simply so much content you need to produce to create the diversity you need for a living, breathing world. We've seen the need, of course, to expand our skill sets, and hiring in experts in different fields -- like cinematography and film direction -- to get a more cinematic feel to it.

I wanted to talk about the AI, actually, because I spoke about this before when I interviewed your technical lead, Martin Amor. Where does the AI begin and the game design end? How do you tackle that?

TB: It's very difficult for us, because it's like the chicken and the egg in a way. Because we want a lot of diversity with the AI, and we also want a lot of responses, but we are not really sure if it's going to work until we actually try it out. And just trying something that feels believable is extremely expensive -- with all the mo-cap you need, all the voice recordings you need, all the context within the world -- to come to a point where someone can sit down that is not on the team and judge it for themselves.

So it's been a lot back and forth between level design, and game design, and technical design, to come up with something that we believe can tackle anything you do in an interesting way, and not just kill you instantly. Because it's very much about the player pushing the game, and the game pushing back, instead of just using a sledgehammer and just killing you instantly, and then restarting.

Because we really want, in Absolution, the feeling that you could try and you could fail, and then you could improvise from there and try something else. Yeah, there's some lines you cannot cross. If you start just killing everybody, of course the AI will be pretty upset, and they will start looking for you if you run away, and they will look for a really long time depending on how bad you were.

But if you're just doing simple things like trespassing into an area, they will kindly try to get you to leave the premises, and if you continue to egg them on, they will get continuously more irritated, calling on their friends. And this kind of playfulness within the AI is something that we believe is very important in a game like Hitman, where it's so much about your experience, and you trying to find your own way -- your own kind of way of solving the game.

And this is something that is also... it's quite difficult, actually, to educate players that this is what the game is trying to serve you, because people are increasingly used to games where you're told to do one thing, and if you stray from this line, there will be nothing else around. It's like, you have this experience, and that's it. So we're telling people, actually, "No, no, no. You choose by yourself."

If you want to go in here, or here, or if you want to kill them or not, it actually changes the way you play the game -- when you understand that you have the choice. So in the first couple of levels, we are continuously working [on it]. And still back in Copenhagen we're trying to find out, are we teaching the players everything that they need to understand about the gameplay and the possibilities of the game?

When it comes to the AI, there's got to be a line. You sort of talked about this. Where you're creating AIs that behave realistically, to some extent -- but then, is that actually fun?

TB: Yes.

How do you handle that?

TB: No, you have to try it out. It is very easy to fall into this "simulation trap", that it has to be realistic. You know, if you do something to this guy, he will forever remember it, and he will know you. It's like, well, that might be so in real life, but it just isn't fun to play.

You have to build a system where there is a certain forgiveness in the system, where essentially you know you're playing a game, and you can be immersed it in any kind of way, but in the end it does come down to mechanics and your understanding -- okay, how can I provoke the game, and what are the results?

And, of course, having to hide for an hour for them to stand down is not very interesting, so finding the balance between making it realistic and making it playable is something that we are constantly trying to see if it's working, and using user tests as a reference, to see are they having fun with this, or are they having fun with that? And different groups can say, "Okay, now it seems like 20 seconds is better than 50 seconds on this stand-down for these particular things."

When it comes to games that have heavy systems design -- meaty gameplay systems that operate on rules -- is that getting more and more difficult to do as games start to look more cinematic, more realistic? Is there a tension there between making systemic games and making visceral entertainment experiences?

TB: Yeah, I think you said it yourself. With the first Hitman game, you could basically just make a bald guy, and you didn't need to make anything around it, but now you have to create this enormous context, and you really have to spend a lot of effort to build up everything to a degree where you can immerse yourself more into any area of the game.

And I'm not sure. I think in some, you can say like "systemic" games, they will take more clever choices, I guess, because of their design that allows for simple systems. I remember when I played Portal 2, I was so envious of them in the design choices that they made, with having a first person game, everybody around you are robots, and the almighty voice is the one telling you the story of the game -- while we have to live with this world where everyone is right in your face, your main character is right there, everything has to be done flawlessly for you to be completely immersed in it.

And sometimes I really wish we had a game that could have done some simpler choices that would allow us to, for instance, tell the story in a simpler way than we have to live with. We have this almost mute character that is completely mechanical, so he will never express himself like Nathan Drake would, or like Kane & Lynch would, always telling you what they're feeling, and how they're doing in the game. So the only thing we can hope for is that people can then at least live themselves more into our main character, because he is such a non-described character, and because there is so much freedom of choice put in the world.

You just referenced "immersion", and this is a word we hear a lot, specifically usually with big console games, but it's not that well-defined. How do you define it from your perspective, as a developer?

TB: Well, it's about you feeling that you're really inside the experience, and that in some way this world is coherent and believable for you, and the logic of the world is not taking you out of the experience.

And we have been discussing this a lot with the heads-up display, and the intrusion into the world that it has, and how cinematic a game would be the more you remove it from the game.

But sometimes it also seems like adding these elements, because you very quickly abstract [information] from these things, that it will not necessarily take you out of the immersion. It will just suddenly help you to live yourself truly.

But it is, as you say, an undefined word. It's difficult to say what it does to you, and sometimes it can be not necessarily the drama or the animation that will make it immersive, it will be the mechanics of the game that you just start to understand, and you're looking at the clockwork more than you're looking at the production values, so to speak. So I think you can have games that are very minimalistic that are extremely immersive experiences.

I think for a game like this, honestly, having a HUD that lets you quickly check all the information that you need to know is actually more immersive than trying to take it away.

TB: Yes.

We've had games in the past where people experimented with things like more blood on the shirt, or whatever, he starts limping, that kind of stuff. That gives me another layer of decisions I have to make, another kind of information I have to process. I think a bar that just tells me how hurt I am is the quickest solution.

TB: Yes. That's why we also reverted back to a more, you could say "classic", heads-up display. While seeing games like Dead Space, of course everybody's really envious of them. They have this really cool suit, they use that actively. They have the very cool menu system. You just look at it -- it's so sexy, you know? And everybody wants it.

But it would never fit this game, right? Having his barcode [on the back of the main character's head] fill up with blood, for instance? [laughs] So we can't really use that. Everybody has to use their own way, I guess, to see what fits. You have to see what fits your game and your mechanics best.

Speaking of Dead Space brings me back to something you mentioned earlier. People are very much getting used to linear games. Dead Space is a pretty linear game. How do you let players know they have options?

TB: This is also something we've been working with. And this is also a big problem, or it's a big challenge for us, because our fan base are very diverse. Some people, they resent being told anything in the world. They just want their target, and they want to find out by themselves. Other people, they actually want to be told the options that they have and the opportunities in the world.

We found out that yeah, one thing is explaining the mechanics, so that you know that you have these different options in for your character. We have this new feature called "instinct" that's been a big debate, which seems to go down very, very well when people actually get their hands on it. But you can also use that to find things in the world that might not be directly linked to gameplay, but more kind of conceptual areas.

If I press my instinct button, you see these little sparks rippling up different places in the world, and this is just simply telling there are these points of interest that might be few spots down -- kind of subtly telling you that there are these opportunities in the world.

But this is a real tightrope for us to walk, because the players have to define themselves, in a way, as they approach the game. If they want to play on normal, they will get more information. If they play on easy, they will get even more. But we really try to restrict ourselves from telling you a solution, because we want to say, "There are these opportunities, and you always have them available, and you should find out for yourself which one is the best one."