• scroller_01_southpark.jpg
  • scroller_02_thecrew.jpg
  • scroller_03_sanfrancisco.png

Level Design


Enhancing the gameplay



The Level Design is the pratical expression of the Game Design. It is directly dependent on the gameplay that it has to enhance. This is the reason why I always take my time before designing levels to accurately analyze the gameplay and to resolve any remaining issues.


This is also a great opportunity to get a first overview of all the possibilities I could exploit. Indeed, I use to draft as many gameplay workshops as possible - starting with very simple ones and iterating it by gradually implementing new elements.


You can see for example few gameplay workshops for a platformer personnal project on the opposite sheets of paper.



Defining strong visual identities



Once the gameplay clarified, I like to talk with artists about the environments where each level should take place. We talk about overall themes (if no scenario has been designed yet), look for visual references on Internet, define color palettes... While artists pursue the concepts, I wonder how to smartly integrate the gameplay elements. For example, I could associate an obstacle in a forest environment with a stump, a platform with a branch, a slowing surface with mud, a dangerous one with bramble...


The relevance of this step consists in defining strong visual identities and creating consistent connections between the art and the gameplay. Indeed, the player should be able to quickly identity a specific level and to anticipate each element behavior according to its aspect.




Using Rational Level Design



The Rational Level Design (RLD) is a mathematical approach of the Level Design which consists in allocating difficulties values to each gameplay scenarios in order to better regulate the difficulty curve. Here the steps I usually follow:

  1. Allocating a difficulty value to the gameplay scenarios according to specific standards
  2. Designing the aspect of the game difficulty curve
  3. Deducing the overall difficulty of each level
  4. Associating gameplay scenarios to each level in order to match its targeted difficulty and duration
  5. Creating the level flowchart / layout to regulate the inner difficulty; alternating challenging and resting areas.

Of course, the Rational Level Designer is only a guidance: the game feel has an absolute priority over any mathematical value; only realtime tests can define how difficult a level is. For further information about the RLD, I invite you to read this exceptional article the Gamasutra community.


Playing on instincts



The greatest challenge of the Level Design is to instinctively guide the player in the targeted direction (especially in 3D environments). Of course, signs can easily indicate the good way but playing on instincts greatly improves the game flow and gives to the player the illusion he masters the level.



The most famous method consists in spreading coins / stars like Super Mario Bros. does. Indeed, the player is basically cupid and likes gaining a lot of points. Spreading coins / stars at the specific positions such as a cavern entrance or stairs so greatly encourages the player entering or going upstairs.



Many other methods exist. You can see below an excellent video shot by Clement Melendez who analyzes how F.E.A.R. 2 plays on lights, colors and movements to point up specific elements or directions. A detailed article is available on Clement's website.





Choosing the right tool is a decisive step to efficiently model the game levels. For FPS projects, I like to prototype levels on the Unreal Development Kit. As UDK is based on the FPS gameplay, the levels can be directly tested in grey boxing and exported as OBJ or FBX files once polished (these formats can be imported in Autodesk 3D Studio Max). UDK also includes a Terrain Editor which is very useful to shape large outside levels.

For other 3D projects, I may use Autodesk 3D Studio Max or Google SketchUp. Artists usually work on 3D Studio Max; using this tool could consequently prevent export or compatibility issues. SketchUp is also quite interesting thanks to its intuitive ergonomics and the high constrast between models / faces.

For 2D projects, I often use Microsoft Visio that I particularly like. Indeed, the tool offers useful layer and snapping systems which help Level Designers to quickly model tidy levels. I have also developed a macro exporting specific data to a XML file. It allows to quickly design, test and correct any level without the help of programmers.