Interviews// Peter Molyneux

SPOnG: I think one of the problems with Kinect in terms of core games is that there's not an obvious way to navigate game worlds like you have with analogue sticks. Was there any temptation to have something - maybe just a nunchuk-type peripheral that you could use in conjunction with the rest of the hands-free tech - that would bring Kinect games closer to 'core' game mechanics?

Peter Molyneux: Well, to start off with I think you fall into the same trap that I fell into when I first got Kinect... I thought, 'I wish there was a nunchuk involved in this'. Now, actually thinking it through, I realise that actually I was just trying to fit a round peg into a square hole.

And wouldn't it be better to think of what a core game experience actually is and what navigation is. If I describe navigation the vast majority of games, in one way it sounds like the most tedious thing in the world. If I said, 'Right, in most games 70% of the time your interaction is about pushing the thumbstick forward and doing nothing else', and you'd never seen a controller before, you'd say 'oh, that sounds rubbish! I'd never want that.' So, maybe as far as core gamers and navigating worlds is concerned there's a different way of doing that, and it's less about, you know, you move the thumbstick slightly from left to right and it's quite rare you pull it back, so there are ways that you can get round that, absolutely.


SPOnG: With Fable 3, was there ever any temptation to just go all-in with Kinect and ditch the controller?

Peter Molyneux: No, not really. Fable was very much crafted to be a controller experience. That being said, the Kinect stuff that we are working on is pretty cool.


SPOnG: One of the things you mentioned in your talk that was pretty interesting was how Kinect can affect the narrative of a game. Now, from what we've seen the games for it seem to very much be focused on the actions you're performing rather than a narrative. Could you expand on how Kinect affects narrative? Maybe that's a slightly nebulous question...

Peter Molyneux: It's not a nebulous question. It's actually quite a specific question and there's a very easy, specific answer that is, really, that it's one of the surprises of the TED talk that I'm giving is how that connection is made and how, if you can see a person, how then you feel that you're closer to the experience, and most crucially that the characters in your experience can meet the person interacting with them, which is the opposite way of thinking about it. That affects the narrative enormously.

But, you know, narrative is very hard to get across if I'm standing and jumping and being very active. If you and I were jogging together then I probably wouldn't be telling you a story. We'd probably be just listening to music, because human beings find it very hard to multi-task, and they actually find it quite hard to stand up and listen to a story and things like that. So, again, I think that's something I'll probably cover in the TED talk.


SPOnG: Well, that brings us back around to Milo. With what you've just said in mind, will Microsoft be sharing some of the technology behind Milo to facilitate other developers making those sorts of games?

Peter Molyneux: Yeah, for sure. And, you know, there's some of the technology in the animals game and the sports is actually fed from Milo already. Little tiny bits of technology are scattered in other games.


SPOnG: One of the aspects of Kinect that's not been focused on quite so much is the voice recognition. On a granular note, will it recognise regional accents? Is there any training that will be required?

Peter Molyneux: So, I'm not a really big fan of training for voice. Our experience in the world about voice and technology is appalling, and actually the worst thing on my iPhone is the voice recognition. It just doesn't work. You say 'phone' and it comes up with playing a piece of music.

So, what's crucial to me with voice is that you feel like it works from the first second that you use it. That is crucially important. And solving that problem has been a really tricky [thing to do]. That said, even on the 2009 demo, when you used your voice it worked straight away. We showed it to 50 groups of people, we used people from France and Belgium and America and, you know, it worked. You do need slightly different vocabulary files, because the tonation of peoples' voice is very different regionally. But you don't need an exact vocabulary to get to where we want to be.


SPOnG: Thanks for your time!