Wild Fibers, a Postmortem

A long project ends...

After more than fifteen months, 982 commits, 605 issues and eight versions of this prototype we stop its development. It is the biggest project by far created in the club (Wild Box Studio) and it comes at an end.

But first, let's rewind time a bit!

Summary

At first, the project was centered around four principles:

• Tower Defense, you can create towers, upgrade them and buy powers to kill enemies on a strict path.
• A squad system like Into The Breach, each squad of towers has their own capabilities and limitations. Mastering them all will take time.
• Bugs, they are everywhere and we want to make them more fantasy-like in our world. Bugs are the enemies in this game, chewing the forest apart.
• Plants, they are the defenders of the protected land.

We were two for this projects, with occasional help from other members of the club sometimes by the way: Valentin Dremeaux (Sydiko) in charge of design, drawing and animating when needed and Gabriel Jouet (WoodBear) in charge of programming, balancing and additional drawing if needed.

We created a first version back in May 2020. The prototype back then only has a level, four towers and two enemies. But, unlike most of our previous prototypes, it was pretty fun. So, we start to iterate on that. We build level after level, enemies after enemies and so on.

What went right?

Actually, a lot of things went right and was fun to pull off. The prototype was fun and easy to explain, not something all prototypes can be proud of. Ideas for towers were just flooding all over our heads and we were super excited to implement them. Levels and enemies were a bit trickier but they were created, animated, balanced and implemented rather quickly. Pathfinding was really easy to get working, by using multiple Bezier curves hand-crafted, we were able to convey randomness in the paths. Making really cool looking levels without high memory usage. The use of Scriptable Object in Unity lets us gain several hours by holding enemies or tower upgrades information.

What went wrong?

As hobbyist, we are not weathered enough to see these problems coming early on. We had to much ideas, each time we have a new idea we write it down on GitHub as an issue. But as we progress in the game, our backlog (all issues) was getting bigger and bigger. Some issues were always pushed back, some issues from June 2020 are still un started at this day! And they were too much of them, in the end, we find ourselves with more then 70 issues un started or unfinished after some clean ups where we decided if we keep them or not.

Another problem was balancing enemies, levels and towers all together. We are no experts in game design nor level design, so, we tested levels, again and again. But with only two of us, we clearly see a bias in the balancing. We buffed towers we like and nerfed towers we don't see useful. Making them more useless. Each time we added another enemy, tower, skill or level, we need to retest from the beginning to see if the balancing is still good. It takes too much time for a hobby and we eventually stop testing this much to focus on implementation. But implementing without testing is worse.

The final problem was the different of pace in work. We were not working at the same pace and eventually this causes arguments or delays in the project. Even if we are friends, working with somebody with a different background can be tricky or complex, do not hesitate to talk with them if needed, you will work better after.

These opportunities and problems leaves us to change the scope and schedule of our project: At first we were aiming for 8 big updates from 0.0 to 1.0 but we stopped at 0.4, leaving three updates unfinished. We were aiming to launch a version every two months to, but ends up impossible.

Within the project, we can see some changes of scope too:

• We went from 10 levels to 7 and finally we shipped 3
• We stop developing the "squad" based system to focus on only one squad: Crazy Foliage
• We went from 21 enemies in 3 families, to 14 in 2 families to 8 in 2 families
• We wanted music and sounds, it was not done due to skills and time

Again, even if this seems pretty grim, the project was really fun to pull off and to publish it on Itch.io, and again:

Thanks for playing our game!