After the End

Overview

After the End is a first-person 3D narrative prototype about sustainability, built in Unity for the Digital Games course of my MSc in Interaction Design. The player is Ulisses, the first human to return to a post-collapse Earth, and must decide the planet’s future: restore it for humans only, or seek balance with the mutant insects that now dominate it.

We designed the game so that the “best” ending requires choosing coexistence over domination — and then ran a multimodal evaluation (EGameFlow/GEQ, Self-Assessment Manikin, Geneva Emotion Wheel) to measure whether the experience actually produced the reflection we intended.

The headline finding is an honest one: it mostly didn’t. 80% of players exterminated their way to the “Human Dominion” ending, despite the insects being completely harmless. The evaluation showed why — the mechanics we gave players (a gun, monstrous creature design, a hostile soundscape) communicated “threat” louder than the narrative communicated “coexist”. That gap between intended message and designed behaviour became the most valuable lesson of the project.

My role: game and narrative design, storyboarding, UI and pixel art, Unity implementation, and the design and analysis of the emotional evaluation, working as a team of two with Leandro Francisco.

The concept

The premise: humanity fled a devastated Earth decades ago and survives in a decaying space station. Reconnaissance mission #RXT31H sends Ulisses back to the surface, where a new dominant species — giant mutant insects — has taken over.

The choice of insects wasn’t aesthetic. Insects are, in reality, biologically more resistant to radiation, pollutants and extreme environmental change than most life forms, which makes them a scientifically plausible survivor species. In the game’s world they took over not through aggression but through pure adaptability — they live in complex colonies and are passively indifferent to humans. They never attack, even when provoked.

That’s the ethical trap at the core of the game: the player can kill them, and the fiction whispers that they should — but nothing in the insects’ behaviour justifies it.

Three characters carry the narrative:

• Ulisses — the protagonist, a 28-year-old elite soldier trained to face “any adversity”. His training is part of the problem: he arrives primed for combat.
• RS1 — a friendly spherical robot, a remnant of the old human civilisation. Deliberately impartial: it presents both poles of every decision (humans-only vs. both species) without judging, so choices belong entirely to the player.
• The insects — visually monstrous, taller than an adult human, with asymmetric mutated limbs. Threatening to look at; harmless in behaviour. That tension was intentional.

Designing the journey

The game unfolds in three missions across a four-zone map — landing zone, water supply station, insect nest, and the “Elephant Graveyard” — with paths and boundaries designed to steer players through the zones in the intended order.

Two decision points determine the ending:

1. The water panel — inside the supply station, the player redirects water to one of three targets: an empty field (benefits humans only), the insect colony (benefits insects only), or both.
2. The seeds — at the Elephant Graveyard, the player plants one of two seed types: natural (humans prosper, insects are condemned) or hybrid (both species can prosper together).

The combination of these two choices produces four endings, each delivered as a narrated comic-book cutscene:

Reaching Ending 1 is the game’s true objective and its maximum reward; players who only benefit humanity are confronted with a final message that reprimands them.

Before implementation, we storyboarded the full experience in 16 frames — first on paper, then digitised in Procreate — specifying for each frame what the player sees, what it should transmit emotionally, what they can do, and the consequences of each action. This frame-by-frame emotional intent later became the baseline against which we could compare measured emotions.

Building the prototype

The prototype was built in Unity 3D, with the terrain modelled as a Unity Terrain and populated with assets from Sketchfab and the Unity Asset Store — the insect nest, ruined buildings, dead trees, elephant skulls. Smaller props (insect nests and eggs, the station’s control panel) were generated from prompts in MeshyAI; Blender handled 3D modelling and texturing; digital art composition was done in Figma and Canva.

The first-person character carries a gun that disintegrates insects, reloading after 12 shots — a deliberate design decision that would later prove decisive in the results. Movement uses WASD + Space, and the character is immortal, keeping the focus on choices rather than survival skill.

Dialogue drives the whole experience. Using a dialogue system from the Unity Asset Store (Grasshop Dev), the player presses “F” to talk to RS1 and “C” to contact it after each mission. The same system powers the mission-objective interface. To identify who’s speaking, each character has a pixel-art portrait — I designed RS1’s in Figma.

Evaluating the experience

A single usability score couldn’t capture what we needed to know — whether the game was enjoyable, but also whether it provoked the intended emotional reflection. So we combined three validated instruments:

• EGameFlow / Game Experience Questionnaire (Fu, Su & Yu, 2008, built on Sweetser & Wyeth’s GameFlow) — enjoyment and engagement across dimensions such as concentration, goal clarity, feedback, challenge, autonomy and immersion, measured in-game and post-game.
• Self-Assessment Manikin (Bradley & Lang, 1994) — pictorial, non-verbal 9-point scales for pleasure, arousal and dominance, bypassing the limits of verbal self-report.
• Geneva Emotion Wheel (Scherer et al.) — frequency and intensity of 20 distinct emotion families, mixing dimensional and discrete approaches to emotion.

Five participants played the prototype and answered an anonymous Google Forms questionnaire (GDPR-compliant under Portuguese Law 58/2019). The sample was young (100% aged 18–25) and predominantly male (80%), with diverse play habits — from frequent players (40%) to one participant who never plays. I analysed the data in Python with Pandas.

Findings

The game worked as a game. In-game GEQ scores showed strong engagement: Competence reached the highest mean (μ = 3.20 on a 4-point scale) with Sensory immersion right behind (μ = 3.18), while Tension scored lowest (μ = 1.60) — players felt capable and immersed, not stressed. Post-game, Positive Experience led (μ = 3.06) and Tiredness was minimal (μ = 1.75).

The SAM confirmed a comfortable, controlled experience: high dominance (μ = 7.20/9, median 7) and high pleasure (μ = 6.20/9, median 6), but low arousal (μ = 1.80/9) — engaging without ever being alarming. Dominance also showed the widest spread (σ = 1.10, range 6–9), meaning some players felt far more in command than others.

Emotionally, the reception was overwhelmingly positive. On the Geneva Emotion Wheel, every single participant reported Interest (100%), with Contentment and Fun at 80%. Joy was the most intense emotion (μ = 3.67 on the 4-point intensity scale), followed by Interest (3.60), Contentment and Fun (3.50 each). Compassion appeared in 60% of participants (μ = 3.33) — the narrative’s empathetic side did land for some. Hostile emotions (Hatred, Contempt, Shame, Anger) were entirely absent.

But then came the endings. Four out of five players (80%) reached Ending 3 — Human Dominion — by redirecting water away from the insects. Only one player made the choices leading to Ending 1, Balanced Hope, the outcome the entire game was built to reward.

Why? The design told them to. The insects never attack — but they look like they will, the soundscape feels hostile, and the player is holding a gun from minute one. One participant put it perfectly: “The bugs don’t kill, but I do!” (“Os bichos não matam, mas eu mato!”). Player typology (Bartle) likely amplified the effect, but the core mechanism was ours: we equipped players for extermination and then asked them, narratively, to choose peace. Mechanics won.

Learnings

• Design communicates louder than narrative. Every affordance is an argument. Giving the player a gun tells them what the game is about more convincingly than any dialogue line — if the intended message is coexistence, the mechanics must make coexistence feel like the natural verb, not just the virtuous one.
• Measuring emotions exposed what playtesting alone wouldn’t. The triangulation mattered: EGameFlow said “the game is enjoyable”, SAM said “players felt safe and in control”, the Geneva Wheel said “no hostility was felt” — and the endings distribution said “they exterminated anyway”. Only the combination reveals that the aggression was structural, not emotional.
• A reprimand is not a redesign. Scolding players in the final cutscene for human-centric choices doesn’t counteract an entire playthrough of mechanical incentives. The honest fix is upstream: make non-violent traversal mechanically interesting (the stealth section pointed in the right direction), and question whether the weapon should exist at all.
• Known limitations, stated plainly: n = 5, drawn from a homogeneous age group, is enough to surface directional insights but not to generalise — future iterations need a larger, more diverse sample and a harder difficulty curve to test whether challenge changes moral behaviour.