- Role: Lead Level Designer & Gameplay Designer
- Team size: 4
- Team Members: Marguerite Dibble (Artist), Michael Hopke (Programmer), Matt Brand(Programmer)
- Date: May 2011 –
- Engine: Unity
- Description: Loc is a 3D tile based puzzle game created for a casual audience. Using three different types of pieces the player must create a path from the beginning tile to the end tile. In addition they are required to also pass through particular tiles in order to win. The puzzles compound in difficulty as the play space is the six sides of a cube
- Company: Birnamwood Games
- Available: Click Here for Web Demo also available on the iOS store: iTunes Store
The concept first began with a desire to move about on the surface of a three dimensional object. I had just finished playing Double Fine’s Psyconauts and was greatly intrigued by their menu screen. It was a brain which the player had to run across in order to load their current save and enter the options menu. The best part was that your gravity was localized to the brain, the player could never fall off even if they ran to the underside of the brain.
This concept I applied to a cube, one in which the player could freely rotate, but continue to play along its surfaces. What came next was an amalgamation of the rubric’s cube and classical sliding tile puzzles.
Each face of the cube is broken into a four by four grid, allowing for a sixteen tile playspace. This number gave me a wide variety of movement options, while not overloading the player with the number of tiles they must utilize.
Loc is based around the use of different tiles in order to solve each puzzle. Every level has a start tile, where the path begins, and an end tile where the path ends. These two tiles cannot be moved or rotated. They function as starting places for the player in their thought process for overcoming each level. Each face then has a set number of straight, corner and cross tiles which the player moves around to create a path.
What I quickly discovered was that players needed some reason to utilize all side of the cube, otherwise they would naturally take the path of least resistance. This was where the concept of gates was born. These solid green tiles cannot be moved or rotated and the path must pass through them all in order for the level to be complete.
Creating a Level
This was my work flow for creating every level in Loc. I was restricted, due to licensing to only use Photoshop, however through practice I was able to create a level in around 20 minutes. So what you are seeing here is the folded out play space of Loc. This was by far the easiest way to look at patterns and ensure that the path would correctly cross between faces. Each level was built solved first than each face was unsolved moving each tile individually in order to replicate in-game tile movement.
I would start each level with a theme or pattern in mind. In this particular instance I wanted to use the least amount of tiles possible, but maximize complexity. This was done by having many of the faces link to two or more other faces. As a general level design each level must have at the very least four open locations to move tiles around. Any less than this and the player quickly get frustrated by the limitations of the tile movement.
Placement of the green tiles, which are the loc’ed tiles, was key. These tiles along with the start and end piece functioned as a starting point for players trying to solve the puzzle. They gave direction for each face and would often indicate what pieces needed to go where. Making a level harder was sometimes as easy as removing these loc’ed tiles.
From there the unsolved and solved images of each level were translated into a series of numbers. For example a 31 represented a straight tile that is positioned vertically. This was our process as Loc unfortunately did not have a built in level editor. This number translation was the only way for each level to be put into the game. Though it had its margins of human error we were able to correct them thanks to a rigorous testing team who would play through each level.
One of the most rewarding aspects of Loc is that several of the puzzles can be completed in multiple different ways. The only requirement is that the path must trace from the start tile
through any gated tiles and reach the end tile. This allows for player creativity and a bit of ownership over their particular method. This comes with a risk; each level needs to be extensively tested to ensure that every way is indeed possible.
On top of this challenge is creating an accurate difficulty curve which challenges players, but not enough to drive them away. To track this I pushed for us to incorporate an analytics system into our game. I want to be able to track how players completed each level, the time it took,
the number of resets per level, and so forth.
Using this system I was able to track player data and shift levels around according to how difficult I wanted the game to be. The following graphs show how the difficulty curve has changed over time.
Looking at this information I realized that many things were being misrepresented. On these first four graphs the actual levels changed dramatically which means that any two graphs cannot even be compared to see how the timings of each individual changed over time. The other information that is missing is the number of sides of the cube the player is interacting. Taking this
information in I started over. This following graph is correlated from the information provided by over thirty testers. I took each time and averaged them to find the median time per level. Each color segment shows the number of faces they are playing with. I made sure to not change levels around between testing sessions this time.
This graph illustrates the near exact difficulty curve I was striving to make. Each section has a high point, which is immediately followed by an easier level. This is to help retain our player base, so they do not give up after finishing a difficult puzzle.