Cuboid Sandbox – Postmortem

Posted by (twitter: @cominu)
September 10th, 2012 6:22 am

Original post is here.

I’m back with some (good?) news!

This time I’m going to talk about a little experimental game made for the 24th Ludum Dare. The game is called Cuboid Sandbox and the description in the contest page is something like this:

Half game and half simulation, this sandbox let you discover the beauty and wilderness of the different Cuboid species fighting for food and territory control in a constantly evolving environment.

Wait. Whaaat?

Well…this 3 days long experiment ended up being more like a simulation than a game. But, what is really a game? When I was a child I always found entertaining to watch ants movements and this game probably have been influenced by that experience :)

Cuboid Sandbox

So, what happened during the development? Here’s a video timelapse (3 development days compressed to 3 minutes) and little postmortem:

Cuboid Sandbox Timelapse video (youtube)

What went right

  • Familiarity with the tools
  • No time-consuming graphics and 3d models. Almost everything was made with a combination of cubes and simple texturing
  • Everything is procedural, no time wasted on making levels or story
  • Side effects are good: the “heatmap” can be used to create abstract paintings 😀
  • Making a very small warmup game the days before the competition helped me to focus more on tangible things and less on “under the hood” geeky stuff :)
  • The unity integrated profiler rulez: sometimes you have to optimize and it’s better if you know what part of the code is really slow. I can’t live without it…
  • Coroutines! I used extensively coroutines to spread the workload across different frames avoiding cpu load spikes.
  • The still-in-beta Linux exporter of unity4 worked without a single problem.

What went wrong

  • Too many hours of work. For the next Ludum Dare I’ll try making something smaller or reusing code libraries from other projects.
  • Unfortunately Marco had some problems and I worked almost solo for the 2nd and 3rd day :(
  • We didn’t really like the theme and we wasted 8 precious hours deciding what kind of game to do.
  • Handling a large numbers of dynamic objects is hard and need time to optimize the code.
  • Too many hours wasted tweaking values to balance the simulation.
  • Lack of interaction with the simulated world.


This was the first time I entered a similar competition and I learned a lot about the time needed to complete something playable and the disastrous side-effect of sleep deprivation.

I really had a good time stepping out of my comfort zone for a while and making something completely different from Blackreef Pirates and I think this little experiment will receive some extra work in the future with more emergent behaviours spawning from different cuboid AI rules and more ways to influence the simulation.


Still awake? What are you waiting? Go play and rate the game! 😀


Tags: , ,

Leave a Reply

You must be logged in to post a comment.

[cache: storing page]