SeaScale

Overview

This game was my entry for the GMTK 2024 gamejam. https://itch.io/jam/gmtk-2024
The theme for this global jam was Built to Scale .
The game idea came from the word scales as for fish's skin and managing the fishes to scale up their numbers.
The project is a cozy, top-down view, underwater micro-managent game.
Many iterations of AI behaviours and parameters were required to balance out the ecosystem.
Development took 4 days as given by the deadline of the gamejam. This project was my first short jam entry.

Project Details

Designed and developed in less than 4 days
Created in UE 5.4
Scripting, Feature Development, Game Design, AI Design
Solo work
Assets from Epic Marketplace and the web (see sources on the download page)

CFO - Chief Fish Orchestrator

For the gameplay loop I wanted the player mostly to observe nature doing its part: fish swimming around, looking for food when hungry and grow in numbers when happy. I intentionally left a few tools for the player to affect the outcome of fish populations, still without the player's interaction the populations cannot grow and one of them will most likely perish ending the game. As the development time was short I had to dump many of the effects outside of the natural ecosystem, eg. fishing, rehabitation. So I ended up with two internal ecosystem tools coming from the bottom and top of the food chain:

Player is able to spawn worms to feed the small fishes. This allows the player to build the lowest lifeform at their will and support the growth of the next step easily.
Player can drag'n'drop the sharks. The player can contain these top predators as they can hunt down all the big fishes. Managing two sharks at the same time can be challenging as the player needs to follow both of their movement and act quickly by moving them away. This is the most skillful part of the game. Still letting sharks go off is a must, as they also need to increase their numbers, and they can also aid the player to control the population of the big fish.

Anything in between these food chain steps is taken care of by the fish AI.

Tip the scales

As the game had constrained the player inclusion to certain parts of the ecosystem, the running AI had to be crisp and balanced in-between. By default the fish regardless of their type were always looking for food unless their happiness meter was full. This caused high uncertainty for the player to actually reach the winning condition.

After several iterations of the actor AI behaviour and parameter tweaking, I introduced two big changes:

The fish ended up switching between idle and hunting state. The hunting state comes more often while the fish is not happy, eg not fully satisfied. Once their happiness meter is filled, they will stop hunting for a longer time. Over time happiness will deteoriate and once again the fish needs to eat again. This loop is broken, only by the death of the fish, eaten by a bigger one, or by the end of the game.
The implemented hunting state isn't a 100% success for the fish to get the food. As fishes move all the time, they can avoid the bites of the bigger fish, resulting in an organic random factor of a successful hunt. The only exception is the small fish are always catching the worms put down by the player.

Worms Armageddon

As placing the worms is the only tool for the player to manage the small fishes, I did many iterations them. Eventually the result was a static worm actor with built-in food capacity. The non-mobility was required so the small fish can easily find them so one less variable of an unsuccessful hunt was removed, making the predictability of the game play higher. One more ease to this was that the worm could provide enough food for more than one small fish, and the worm actor would only disappear after the small fish ate all the food stored.

With these factors combined, few playtesters realized they can just massively put down worms all around the scene to have an almost unlimited food source for the small fish. I kept this strategy in the final version, as I've found it funny, although it would be easy to cancel this tactic by putting a lifetime on the worms, or delaying the place action in general.

Post Mortem

As this was my first gamejam with a short deadline, I had learned how to design, develop, iterate and test efficiently with a brief turnaround time. Although the start of the project went slowly, once I got the main concept sketched in my mind, the production went smooth and fast. Most of my time was taken by the iteration of the game ecosystem loop, how often do the fish need to eat, how quick they swim, how much food they need, etc. I had to cut corners in order to finish in time and few major bugs existed in the game jam version as well.

I had planned the following from the start and can be added to a later version:

Player could put down algae and corals to attract diverse smaller fish and crabs
Different bigger fish and marine mammals with various needs increase in numbers