Designing Kinect Games: Design Principles from Mini Ninjas Adventures

 A while back I wrote a blog post about game design for motion games. That post covered some general ideas. Now that Mini Ninjas Adventures, the Kinect game I was lead designer on is out on XBLA, I can talk more specifically about what we at Side-Kick Games wanted to achieve and how we went about getting there.

My first post will discuss the design principles behind Mini Ninjas Adventures and how they are reflected in the final game. The second post will talk about the challenges of working with Kinect. I warn you in advance that the posts will contain a fair amount of both nuts and bolts. Now that that's out of the way, let's get to it.

When I write an initial game design document for a project, I always put the guiding design principles for that game on a post-it note. This note becomes my guiding light throughout the project. Game production is hectic and stressful and that humble piece of paper can remind you of what's important and what's not. What can be tossed aside and what's worth fighting for.

In the case of Mini Ninjas Adventures, these were the three principles:

1) Be the Ninja

2) Player Connection

3) Freedom

Let's go into each one:

1) BE THE NINJA

Probably the most self-explanatory of the three. The player should feel like a kick-ass Ninja at all times (That last sentence is taken from the pitch document to Square Enix, the publisher. This is why I love this industry so much).

What that meant is that although the game has a huge number of gestures for a Kinect game (We'll get to that) none of them should just be functional, replacing a button press. They all need to make the player feel powerful, like, well, a kick-ass ninja.

Some of the gestures we were using are inherently powerful. You SLASH with your sword and KICK enemies away. The challenge was to make every gesture in the game feel as powerful as those ones. For this, I looked to the movement principals I talked about in my previous post. Stretching your limbs and doing big movements with them makes you feel strong. Folding your limbs and being static make you feel weaker.

So for instance, for our main character's magic attacks we chose three big, powerful gestures. You PUSH your hands forward to unleash a thunder attack, You DRAG your hand from the heavens to the earth to bring down lightning and you JUMP to create a massive earthquake.

To switch between weapons in the game, you pull the weapon from behind your back. The aim was to replicate that classic movie moment of the hero pulling his sword from its scabbard. This takes a functional need of the game design, switching weapons, and turns it into a movie-like, dramatic moment.

When you want your Ninja friends to jump into battle and demolish your enemies, you shout NINJA! at the top of your voice. This is because shouting NINJA! is really, really fun.

So the criteria for every gesture in the game was simple: If it doesn't make you feel awesome, it's out.

2) PLAYER CONNECTION

When I design games, I tend to think about the concept of Flow a lot. That magical moment where the player is not aware of their surroundings and doesn't feel the controller in their hand. They are in the game and nothing else exists. This is much harder to achieve in motion games since players tend to be more self conscious when they are flailing around. Any little thing can yank the player's consciousness right out of the game and back into his living room where he is not a Ninja but is in fact, a dude playing Kinect in his underwear. Your job as a motion game designer is to try and overcome those points of disconnect that pull the player out of the game.

Since our game is a sequel to IO Interactive's original MINI NINJAS there was no question that it was going to be a third person action game. That meant a one-to-one connection between the player's actions and the actions of his on-screen avatar. This may seem like an easy choice to make in the bigger world of games but if you look at the Kinect's best reviewed and received games, most do not use a player controlled avatar.

Fruit Ninja has no visual representation of the player or his weapon, just his actions. Dance Central uses a canned animation and highlights the areas where the player's action don't match the intended dance move. In Gunstringer, you use your arms to control a puppet on a string (A truly inspired design choice). And in Double Fine Happy Action Theatre it really doesn't matter what you do. It still looks cool.

Using a player controlled avatar is a challenging choice to make in Kinect development because it exposes the weak points of the hardware to the player. There is an inherent lag that becomes more evident to the player once his avatar's limbs are connected to his real ones. There is also the sensor's tendency to freak out every now and then and for instance, make your left elbow suddenly appear below your right hip. This is what causes these weird moments when your Kinect Avatar suddenly seems to suffer from broken limbs.

The next big issue we faced is the one that plagues most fighting games. Big, dramatic animations vs. responsiveness to player actions. Do you lock the player into that gorgeous two-second animation of his character flying into the air and doing a devastating spin-kick, or do you let him break out of it at any moment and lose the drama. In the first option, you can create beautiful, dramatic game moments but you risk infuriating the player since he can't respond to what's happening on the screen now.

This is even more pronounced in motion games. If the player is already doing their next move, a high kick and their avatar hasn't finished doing the previous punch move, it creates a huge disconnect. During production, we went through many iterations of the fighting animations. What we ended up with was a system in which the avatar's limbs follow the player's movement until the very end of the gesture and then transition into a big, fast animation that gives the movement a dramatic oomph. We also let the player break out of the animation if they already started on a new gesture. It was a delicate balancing act, trying to maintain the drama and excitement without sacrificing too much player connection.

3) FREEDOM

This was the big one for me and it was certainly the most challenging. Games to me are all about player autonomy. I want your playthrough of the game to be different than mine. I want you to be free to make different choices than I make.

However, looking at the best received games for the hardware, most of them coped with the inherent detection problems by limiting player choice. In Fruit Ninja the player can do one thing, slash. Dance Central checks the player's movements against a perfect version of the dance move they are trying to perform.

Another solution is a staged one. The player is able to perform a variety of actions, but each is specific to a certain predetermined time. The upcoming Wreckateer uses this to great effect. There are a variety of different activities but they happen one after the other so the Kinect is only looking for one or a couple of non-conflicting gestures at a time.

These are great solutions but they didn't fit the vision for Mini Ninjas Adventures. I wanted the player to face a battlefield full of enemies and to let them dispose of the enemies in whatever order and method they chose. If they wanted to slash a nearby Samurai with their sword, that's fine. If they wanted to kick him away then shoot him with their bow, it's all good. Again, this is very common in Non-Kinect games but not something I've seen done in motion games.

To achieve player freedom, something had to go. That was free movement in a 3D space. The player had to be immersed in the game, moving and switching between weapons without thinking about it. So the complex challenge of navigating a 3D space with motion controls had to go.

What I ended up with was a scheme where the player moves left and right in front of a battlefield that is divided into columns and rows. Enemies will move freely between places and the player will aways face one column. A reviewer compared it to Midway's Tapper and that's as good a description as any. This simplification of player navigation let us put a huge amount of freedom of choice for the player in his attacks. The player can at all times: Use his sword, throwing stars and bow and arrow to target enemies at different ranges, kick, block, do three kinds of magical attacks and call his Ninja friends for help.

There is no way for Kinect to detect all of those motions at the same time. For instance: There's no way to differentiate between a sword slice and a throwing star throw since they are almost identical. The way I approached this problem is by quarantining conflicting gestures behind a transitional gesture. In this case, pulling your weapon from behind your back to switch weapons. This results in a tree of gestures where all of them are available at all times but some demand the player do another gesture before accessing them. And of course, all gestures need to make the player feel awesome.

This process of figuring out what gesture clashes with another and how to solve it was the most involved one of this production. As a game designer you end up with crazy notes about body parts used, angles of movement and distance from body that look like they were done by Kevin Spacey's character in Seven. The game's control scheme was done and redone until the very end to get the best balance we could achieve.

And that's it about the design principles of Mini Ninjas Adventures. Feel free to ask me questions in the comments. Next post will cover the challenges of working with Kinect and the lessons I've learned over the past two years of working with the technology.

Ehud Lavski