Get In, Get Out

Project Outline

• Timeframe: 5 weeks

• Engine: Unity

• Roles: Level Designer, Sole Developer

Responsibilities

• Concept and design a game in my choice of genre through documentation

• Pitch and present final designed concept to clientele for feedback and approval

• Develop the approved concept as a polished prototype

• Source, create and implement art assets

• Create AI enemies and script level events with text based C#

• Design and greybox level layouts

• Playtest and iteration

This project was done as an assessable piece for my Game Design course. For the assessment we were given a list of game genre that we could choose from. We then had to concept and create design documentation for a game idea in that genre, before finally pitching it to faculty for approval. I chose the stealth genre for this assignment. I further twisted the genre by making it a local two-player split screen game. The main objective for the players was to race each other to an escape zone while using stealth and parkour to slip past patrolling AI.

My main goal for this project was to create a level with PvP and PvE elements, that supported third-person parkour movement and stealth.

As I was the only developer on this project, a lot of work had to go into making a reasonably polished experience within the time frame. To aid in this goal, I made use of art asset packs for set dressing and AI models, Mixamo for animations, and the third person parkour package controller from DiasGames. Opting for a low poly, stylised aesthetic allowed me to easily source a lot of free assets.

The Design Process

After having my concept approved for development, I set myself a few design goals for this project. These included:

• Designing and balancing a PvP race level with asymmetry

• Giving the players choice and agency in what route to take to the finish line

• Creating engaging spaces that facilitate stealth and parkour

• Creating at least one scripted event

• Use AI and other environmental elements to create a strong feeling of tension

To design a level capable of reaching these goal I had to take a few things into account such as:

• The movement capabilities of the players (ledge climbing, wall running, crawling etc.)

• The movement constraints of AI

• AI placement and the placement of other elements hazardous to the player, including lasers and security cameras

The first order of business was making some code adjustments to the parkour package I was using. The package was developed for singleplayer experiences, so there was no real support for a local split screen game. This also meant there was no foundation for interactions between the players other than bumping into each other.

Without adding too much to my workload and increasing the scope, the solution I had to this problem was to code in a form of pseudo-combat that the player could use on each other. The package I was using came with a model and a catalogue of movement animations. One of the movement capabilities of the package is a combat roll. For combat, I simply added a sphere collider to the player that activated whenever they performed the combat roll. If they collide with the other player, they are are sent flying and knocked down for a period of time leaving them vulnerable to the security elements. A couple of my friends who I got to playtest this early found this very satisfying and so did I.

With that out of the way I could move onto prototyping the other elements I wanted for the project as well as greyboxing. The theme I landed on for the level was a bank, with the two players being rival phantom thieves. It lends itself well to the stealth aspect, as well as being an environment commonly associated with strong security features. The level begins with a camera fly through to give players a rough idea of what to expect. Players then start at the big vault door housing the money just as the alarm gets triggered. This calls the players’ to action as well as places both of them in proximity of each other, this area also provides a safe space for the players to interact with each other and learn the basic controls. From here I give the players two options, take the vent on the left or the stair well on the right. The players can then split off from each other taking their own path, or stick to the other player and try their best to undermine them.

So far the level design was holding up and providing an experience that I wanted for players, the only thing hindering it was the AI.

The AI I programmed has two states, pursuing and patrolling. Originally, if the AI caught a player they would be knocked down similar to if they were hit by the opposing player’s combat roll. This however felt weak and severely lessened the impact of the AI’s presence as my intention was for the combat to potentially leave the players in a state where the AI could catch them. This meant the consequence for being caught by the AI should be more severe. To overcome this problem I changed some of the AI’s code so that now when they collide with the player, the player is transported to a sort of holding cell within the level. The holding cells were dotted through the level in proximity to the start of a stealth encounter, and had a simple means of escape. This made the AI feel a lot more threatening as it was a strong set back in the race if the player was caught. This however did cause a problem for scenarios where a skill gap was present between players. If one player found any of the stealth encounters particularly challenging and could not overcome them after multiple attempts, it resulted in an overwhelming victory for the other player. This did not feel good for either player.

I wanted to create more opportunities like this later in the level for the players to contest each other. To support this I created small landing areas where the paths available to the players converge, giving them the chance to run into each other if they split up earlier. By placing these areas in the down beats between stealth encounters, it gave the leading player a chance to observe the upcoming area area and the trailing player a chance to cover some of the gap between them. I found this to be a good opportunity to include the small scripted event I had planned at the start of production. Here the player would be hard gated for a small amount of time before they could proceed. I created an elevator that would open through a trigger before the player could move to the finale of the level. The elevator’s arrival also knocks a vent down off the wall allowing a means for the trailing player to catch up.

My solution for this problem was to add a shortcut system into the holding cells. As part of the level I was already working with a modular set of vents to create crawl spaces and ways around guard patrols, I used these vents to insert shortcut routes into the holding cells that passed through the level. Shortcuts were covered by a grate, and only opened when one of the players reaches a milestone point in the level. The trailing player could use this vent if they were caught by the AI again and could safely navigate past the stealth encounter they were struggling with. This worked really well for keeping the race close and the tension high. The victories felt a lot sweeter and the defeats not so bitter.

Before Get In, Get Out I had never designed a level for a stealth game, let alone one with parkour movement and multiplayer. It was a great learning experience that really pushed my level design skills. I would love to do something similar like this again in the future.