Fickle Forest was the first time I had tried Ludum Dare, one of the oldest game jams which has been going since 2002. There are two different options for participating. The one we did was 'Jam' which is similar to other game jams where you have a team. The other one, which I would like to try some day, is 'Compo'. This is where all the game design, code, art and sound is created by one person over the weekend.

The theme for this jam was 'Unstable'. We ideated on a bunch of physics ideas but ended up with an ecosystem. A cool example of this is from the Ars Technica War Stories Ultima Online, where they made an intricate ecosystem and players did not interact how they expected.

We ended up basing our wilderness on the Russian continent as it had a different set of animals than we had normally seen. The Siberian Musk Deer with its fangs was a new animal for me and it was interesting to have both Tigers and Bears together.

We started off working out the detection and interaction which was pretty easy using Godot's layers and collision shapes. This let us have a priority queue of actions for the animals to perform. Animals would look at what was around them then, run from danger, move towards a mate, move towards food or if there was nothing interesting then explore.

One interesting thing with this type of game is the emergent behaviour that results. One bug was that we made predators not collide with trees so it was easier for them to hunt. Bears broke this by being able to eat berries. They could see the berry tree so it became a goal but when they moved onto it they never got the event. This meant that bears would hang around berry bushes, trying to eat them until they starved to death. We also had to slow down the spawning after the game lagged out with exponentially breeding rabbits.

It was really great that both the artists could match the art style, it all looks really consistent and they both drew animals and plants. I was also impressed with how well the 'stamp' look worked. Rather than spending time making heaps of frames they just have a move, eat and dead frame. This meant that more time could be spent on adding different types of animals and the motion brought them to life.

When we had everything working it was a sandbox that was interesting to experiment with but not really a game. To try to turn it into a game with a challenge mode. You have 8 clicks to set up what is going into your ecosystem then your score is the highest number of generations it achieves. This ... kinda works. It definitely results in unstable ecosystems meeting the theme but I think to end up with a reasonable chance of a stable state we would need a more rigorous approach to the balancing.

One of the cool things about making a simulation game is you can go infinitely deep. The gold standard for this is the incredibly deep Dwarf Fortress. Reading a lets play such as Boatmurdered or the type of bugs they find to get an idea of just how much detail and intricacy its simulation has. Everyone could see all the extra detail we could add in to expand it and it would be great to do another simulation game in the future.

Lessons Learnt

Emergent gameplay is cool.
Think about how it will be a game with a challenge.
Minimalist art can work well.