This is a post-mortem for "Claude Van Clam Presents: Roberto Fishermun's Horrific Fishing Adventure 2019", a fishing/ tower defence game

Click here to try our game.

Our Previous Game

Feature creep is a big problem with our previous game jam entries, they usually end up being much larger in scope than we can make and so they end up unpolished. At the end of the previous Ludum Dare, where we made a game about removing feature creep from game development, I wrote a post-mortem which suggested how to remedy this problem.

“Next time we will concentrate on a single aspect of the game, polish that and build a game around that instead of focusing on too much at once.”

So, our intention was to make a small game based on a single mechanic. And then the theme was announced…

The Team

We Are Not Spies have entered Ludum Dare four times, but the team members have swapped in and out. We had quite a big team working on this jam:

Myself, @soldierbear and @david-mason have worked on all previous WANS games. Soldier does the music and David does most of the art. @ttocs167 and @stuple joined for Silicon Boulevard last year, and papasee and @carvster were new for this game.

We are all programmers, and while some of us did have experience with source control, we had never had good experiences with Github and Unity. This would prove to be a continuing source of pain.

Theme Announcement

We were not a fan of the theme that was chosen, it basically forced us to ditch focusing on a single mechanic, which was not a promising start. No-one had expected to get it because it was basically a less inclusive version of “Two things that don’t mix”, so for a while we were clueless. Then we just started naming things. Things like ”Horror sports” and “FPS Retirement home simulation”. Eventually we decided on the idea of a stealth tower defence, where you sneak into places drop turrets and activate them, so they kill all the enemies in an area. Then Carvster said ‘Tamagotchi’ and the idea was forgotten for a while. Then it was settled again, it was definitely a stealth tower defence, so the team started going to bed. Then the idea changed to what would eventually become “Claude Van Clam Presents: Roberto Fishermuns Horrific Fishing Adventure 2019”.

Final Idea

The idea was a tower defence mixed with fishing game, where you build turrets that defend the player character from zombies, which allows him to fish to get money for more turrets. This would require working on two separate mechanics, which worked out well despite wanting to only work on one because it was segmented so allows our team to focus on different things, so people are waiting for work other people are doing.

Development

https://www.youtube.com/watch?v=35AZBtgm6D0

(Timelapse created expertly in movie maker by @david-mason)

Once the final final idea was decided I went to bed, in the morning I woke up to (once it had been uploaded to GIT) to a game that allowed the player to move around, and turrets that aimed and fired at targets. This was the work of ttocs and since the game was his idea he became the unofficial leader – it later transpired that the stealth game was his idea and the fishing idea was carvsters. Ttocs said shaders might be nice to show the range of the turrets, so despite having very little experience with shaders I decided to give it a go.

Since I had not done much work with shaders before I loaded up a tutorial and managed to get enough knowledge to work out how to increase the size of an image, blur it and give it a hue, which was the effect we wanted for the turret. To begin with we were unhappy with it, but once we used a higher resolution image for the turret base it turned out to work quite well. Another problem with this shader was linking it up with the in-game range, this was solved by multiplying the bounds of the image by 2.5 for whatever reason.

The shader turned out to be quite inefficient because we weren’t sure when we wanted it to appear if it appeared only on hover over the turret or if it was a toggle on-off button so it was applied to the object instead of the camera to allow for this flexibility and was not optimised once we decided to leave it at toggle on-off. Additionally, since the original intention was for the player to see the base of the turret and the outline of this base on hover over (although this was never added because turrets have no colliders and you never have to interact with a single turret), so the shader does three passes over the drawing to draw an outline, an original and a large version.

There are also shaders on the temporary towers that make them opaque, and as a hacky method for a day-night cycle on a sprite in front of that turns on with a dark hue at night to make it seem darker. Although this worked well it caused problems with certain builds of the game.

Problems with the Development

Our main trouble was with having little experience using Git and Unity. These problems started when, although we decided to check we were on the latest version of Unity, the repo was created in an old version. This did not cause anything major. The biggest problem was that in our first repo we put the git ignore in the wrong folder, which meant it would not properly ignore the files, which meant levels changed and prefabs broke etc. Eventually we had to ditch this Git and move to a new one (hence the name).

Even after this we had trouble when people would forget to pull before committing or people would not save their prefabs before committing. This primarily meant that the fishing game Carvster was working on which was playable when he left at night was routinely broken by the morning.

As alluded to before shaders were also a problem in building the game, this is because for whatever reason WebGL 2.0 does not support all the shader tools in unity, so in loading up a web version it would just show a pink screen because of the day-night sprite. This was fixed without the need for removing all the shaders on the web version by building using WebGL 1.0.

The Good

While it is not universally popular I am happy with the mood, music and graphics. I am particularly happy with the day night cycle that I worked on and the text that pops up next to characters.

While the design of the fishing game and the tower defence section of the game are both quite simple, they work together well. The game is probably the funnest, most polished game jam I’ve worked on. Although I think the fishing game would work better if you could fish from any position on the beach, making the player choose a specific point was a hacky solution to the problem of people being able to leave if and catch any fish that they missed by quickly running downstream.

The Bad

The main issue with the game from what I understand in the comments so far is that we have done a poor job sign posting the game (a problem that was also noted in Crunchtime). For example, people thinking the game is bugged because you can only place turrets near the player character, that you can’t build in fight phase and at its most extreme one person not realising there was a fishing component to the game at all.

Another big issue was game balance. For many of the players the game was too hard because the early waves can be very brutal if you don’t buy the right things in the right way. However, once you get past this by buying the right upgrades the game no-longer really represents any challenge. This should be able to be fixed by tweaking the prices of upgrades and towers.

While it’s not a design problem it’s also worth noting that the control scheme and particularly the buggy inputs for leaving the fishing minigame are also a cause of frustration.

How to Improve for Next Time

Again, I think we need to focus on a single gameplay element. If we choose to have an economy, it should be polished. Everything we work on should have more polish. This should be done by not building and putting together everything everyone has minutes before the deadline.