OF SAND AND STEEL | romaintresarrieu

OF SAND AND STEEL (School project | Oct. 2017 - June 2018)

Online Boss rush | 3 players | PC

Engine: Unity

Roles: Game, Combat & AI Designer

Being my graduation project, Of Sand and Steel is the biggest game developed during my scholarship. In a team composed of 8 students, we spent 9 months of our 5th year at Supinfogame Rubika working on this online cooperative “boss rush” game for 3 players. If the concept takes its roots into Action-RPGs, we tried to make it so it requires the same approach to boss fights as in MMO raids: everybody has their role in the team, there are clear strategies or tactics to come up with, as much as execution challenges (avoiding areas of effects, for instance). Part of these principles can be found in the game’s special, unique feature: the characters can draw lines behind them on the ground, then use their abilities on these lines to create a wall protecting them from attacks, a big explosion, a healing area, and such. Still, we wanted to keep everything simple enough so the game would be accessible for newcomers to these game genres.

This project has a long story: first of all, I spent about half of my 4th year at school working on its concept. Back then, it was much more inspired by Japanese RPG, and my original goal was to create an action-oriented counterpart to the genre in the form of multiplayer boss fights and dungeon crawling. Bit by bit, the team was formed around the project, and the concept naturally evolved into what it is now: not “my” game, but a game each member of the team could relate to.

On the left is an old key art from the beginning of development. Lots of things changed since; even the game's name was different back then!

Once set, I was put in charge of many game design tasks. First of them were inherent to the game’s “3C”: aside from the many iterations for camera position and behavior, I also imagined how each character would work and synergize with others based on its fighting style and abilities. Even if we created our characters based on the “tank-heal-DPS” trio, I wanted to challenge these archetypes and make it so players would not always have to act accordingly to their assumed role for short period of time. This I kept in mind when ensuring my other main task: AI Design for the enemies.

In the final version of the games, there are only two types of enemies left from the many iterations we created. There is, obviously, one boss. For this one, I kept the phases system from MMOs, that I then split into sub-phases depending on how many health points the boss has left in its current health bar. Each phase has a main theme and the boss’ behavior is based on a part-predefined, part-random cycle based on its attacks pool for this phase. From a phase to the next, the boss can get new attacks or enhanced versions of the previous ones, but I designed the whole fight with players’ progression in mind: first, they can focus on how their character works and their role within the team, then study the boss’ more complex patterns.

The other enemies are small “trash mobs” that are used for several purposes linked to the boss’ behavior: overwhelming players, protecting the boss or sometimes acting as a way to weaken it. These little enemies have three main behaviors: when alone, when in a group, and crowd behavior. For each of them, they use variations of the same three base attacks. I also designed the whole tutorial around these enemies.

I was also in charge of the game balancing, which was a daunting task considering its many aspects coupled with its multiplayer nature. I quickly created a big spreadsheet for that, which allowed me to see the impact of each small value adjustment on the whole game. Once satisfied, I would integrate those values in the game and tweak them a bit more if it felt necessary.

Extract from the spreadsheet I created for balancing. The "Calculator boss" sheet is more complex and detailed than this one.

Finally, I also helped with the game’s trailer realization and editing. This is a version made by another team member, based on the previous version I was in charge of but with new shots that looked less “work in progress”.

We faced a lot of issues during the game development: a troublesome pre-production that slowed us down for the rest of the year, no concept artist, technical arti, experienced environment artist, lots of trouble due to online, motivation drops... This resulted in a game prototype far less polished than we would have liked it to be, despite our efforts to narrow the game’s scope down. We had realized soon enough that developing the fighting system and the enemies would already use all of our workforce, so we dropped the dungeon crawling part; due to our restricted art production force, we also decided to only make one boss, but with more phases that first intended to make the most of all combat features.

Even though, there was also a lot to learn from this project. We experienced all the game development steps for the first time. Since our game is animation-based, the Animator and I managed to establish a solid pipeline for characters and enemies’ animations: I would describe every move (attacks and such) and its duration based on my balancing work and references studies, then the Animator would create of first, rough version of this move we would tweak until the final version. We also had the opportunity to make several presentations to professionals from the games industry (Ubisoft, Sloclap, Asobo, Gameloft, Fishing Cactus…) at various development steps.

Sand (main menu) - Rapahël Batôt

00:00

Steel (all combat phases) - Raphaël Batôt

00:00

And, last but not least, we also worked with the composer Raphaël Batôt to get our very own OST for the game: one track for the main menu, plus one track per boss phase. I overviewed their composition, with the intention of having an evolutive soundtrack that got more and more intense as the fight went on.