Week 2: Defining a Game Loop

This week we are jumping into designing our main game loop and defining what we want the experience to be. Our individual research was a good starting point, but in order to make a rap battle game, we need to know more about the inner works of rap as artform. Therefore, we’ve sent up a meeting with Kenechukwu Echezona, a student at Carnegie Mellon that teaches a Freestyle Rap course.

Design Meetings:

Level Design:

We want our game to take place in subway system and for the player to be a noive rapper battling in this underground competition. Every “subway line” would represent a different style of rap i.e. Crunk, Boombap, UK Drill, etc, and every station would be a different AI Opponent. The player would be riding the train in the opening of the game and then get off at the station where they want to battle.

Hopefully we have time to make an intro scene and tutorial level where we teach the player the main mechanics and why their character is participating in the battle.

The meat of the game is when the player starts preparing for the battle and then participates in the battle. When they get off at the battleground station they will be able to collect information about their opponent, the rap style of that station, and survey what the audience might like. The audience will be powered by an AI personality profile.

When the player starts the battle, they will begin writing their rap for that round with AI programming to help suggest what they should write. First, with the information that the player has collected they will write a short list of things they want to rap about i.e. disses and boasts. Then they will start constructing the rap. We’ve developed two writing modes; fill in the blanks and rhyme sets. The fill in the blanks mode is where the begining of a sentence will be generated and then the player will to fill in the rest of the sentence. The rhyme sets mode is where the player can come up with a collection of rhyming words and a wrap will be generated using that at the prompts from before hand.

In the end, the rap will be scored based on content, sentence structure, and delivery. Content is how much sense the rap made and if there was a story throughout. Sentence structure is more about using different rap skills like inline rhyming or double entendre and how much you’ve catered to the audience. Delivery is a loose metric as of now, but we’re thinking about it being scored for timing (how fast you wrote the rap) and performance.

Characters:

Since battle raps often have disses about their opponent’s appearance and background, we wanted to stay away from human characters and instead do anthropomorphic characters. We’re thinking about having a dog player character. The fish was the tester opponent personality we created as a demo.

Finnegan Flow is a fish that wants to make a name for themselves in the rap community and joined the underground competition to do that. She performs mumble rap because she is constantly in water and often uses B words. They have a school of fish as their audience that enjoys consistent rhyming in an AAAA format.

This character helped us understand what information needs to be given to the player to write a rap. With Finnegan’s backstory, the player knows which rhyme scheme will win over the audience and to stay away from disrespecting fish since that’s what their audience is.

Using AI:

In this design, where exactly are we using the AI? To clarify, we are using AI to power each opponent and collective audience per battleground. We are also using AI to assist the player in writing the raps.

Meet our SME, Kenechukwu:

One of our teammates is taking a Freestyle Rap class at Carnegie Mellon and put us in contact with the teacher, Kenechukwu. We met with him this week and talked about constructing the game in a way that categorizes the process of making a rap.

In our conversation, we talked about finding an easy way to onboard someone new to rap. The hardest thing is for people to get comfortable with themselves, their voice, and their ability. Small inconsequential exercises help ease this discomfort. He also helped us figure out that the audience’s perspective and the performance of the rap are what make or break the battle. In a text-based game, how do we perform?

Preparing for Quarters:

Quarters is a presentation a collection of eight 15-minute presentations about our project to the entire faculty. The goal of these presentations is to highlight our project’s direction and ask the faculty for advice as we enter the development stage.

To prepare for this presentation we are creating two posters with our project purpose/goal, research, and main game loop. As we explain the game loop we will walk to faculty through a paper prototype and have some digital sample for them to test if they have questions about the AI.

The main question we are trying to answer with Quarters is: how do we integrate musicality and delivery into our project? Since we’re making a game with text-based AI but we’re simulating a rap battle we don’t have a clear way to incorporate the act of rapping or the sense of music, flow, or rhythm that are so crucial to the experience.

Next Steps:

Next week, we will present our Quarters presentation and iterate on the feedback from the faculty. We will then outline a timeline for the first half of the semester and work towards our first major prototype.