How to Strip Away Cheesiness

 

The early conversations with AI at the start of this project were honestly cheesy. A memory-sharing ability, a setting where you can only love for one day, a superpower to directly feel emotions. The direction was to create a special story through special settings.

But when I dug into why those felt cheesy, the key became clear.

When Setting Gets Ahead of Emotion, It Becomes Cheesy

"Two people who share memories" is a setting. "Liking someone but time passed without ever saying it" is an emotion. A setting is interesting, but an emotion resonates.

In most cases where a visual novel feels cheesy, the setting is ahead of the emotion. Special abilities, special world-building, special events. These drive the story. The characters' emotions become byproducts of the setting.

Conversely, when emotion is at the center, the story deepens even with a simple setting. "Two people whose timing always misaligns" needs no superpowers. "A relationship where mutual consideration led to nothing" needs no fantasy.

When I asked AI why fantasy settings dominated its early scenario suggestions, it said "Because special settings easily secure narrative momentum." True. But easily secured momentum does not create emotional depth.

This connects to the media comparison we covered in Part 7. Why does the film "Christmas in August" create such deep emotion? No superpowers, no fantasy. A photo studio, an alley, light. It creates emotion through the accumulation of the everyday alone. The same goes for visual novels. It is not the setting but the density of everyday life that determines emotional depth. Eyes meeting in a school corridor, running into each other at the same vending machine, waiting for the same bus on a rainy day. As these scenes repeat and accumulate, the emotion they build is stronger than any superpower setting.

"The Accumulation of Choices," Not "Special Events"

Non-cheesy love stories share something in common. There are no big events. Instead, small choices accumulate.

Among the 10 reality-based scenarios the AI and I organized, the most powerful ones all had this structure.

"We were a great match, but our timing always misaligned." — No special event. Simply that when one was ready, the other was not.

"A consistently good person, but the relationship never becomes romantic." — No special event. Trust and comfort, but no decisive emotion.

"A relationship where mutual consideration led to nothing." — No special event. No conflict either; it just quietly ends.

"I did not realize when the other person liked me, and I found out later." — No special event. Just late recognition.

The reason these scenarios are powerful is that the player feels "Ah, I have experienced something like this." The foundation of empathy lies in reality. The interest that comes from fantasy settings is strong but short-lived. The empathy that comes from realistic emotion is subtle but long-lasting.

The AI and I dug further into this difference. "Analyze the difference in emotion that fantasy-based scenarios and reality-based scenarios give to the player." The AI's answer was crisp. Fantasy scenarios create "Wow, that's cool." Reality scenarios create "Ah, that hurts." The former is consumed; the latter lingers. Of course fantasy is not bad. But what this project pursues is a story the player keeps thinking about after turning off the game. For that, the roots of emotion must be in reality.

The "collision of good intentions" conflict covered in Part 14 is a quintessential reality-based conflict. "A relationship where mutual consideration led to nothing" can overlap with the player's own experience. That overlap turns in-game choices into real choices. "What did I do in this situation?" "Am I making the same mistake?" The moment these questions arise, the game transcends entertainment and becomes experience.

Coincidence Disguised as Convenience

The most common form of cheesiness is this. They "happen to" transfer to the same school. They "happen to" move next door. They "happen to" end up in the same class. They "happen to" join the same club.

This pattern came up frequently in my early conversations with AI. Meeting in elementary school, separated by a move, then "coincidentally" reuniting in middle school. Structurally it makes sense — you need a reunion, so you insert a reunion device. But the reason it feels cheesy is that it sacrifices realism for the scenario's convenience.

In reality, reunions aren't "coincidentally same school." They're searching a name on social media, showing up at an alumni gathering, or bumping into each other at the neighborhood convenience store. Not dramatic. But that's why it feels real.

Cheesiness comes from "forced situations." When the scenario forcibly creates the situation the story needs, the player feels the writer's hand. "Oh, the author set this up." The moment that thought surfaces, immersion breaks.

Instead, for situations to feel natural, they need to emerge from the character's actions. Not "they happen to meet" but "they wanted to meet, so they sought each other out." Or they don't meet at all. In reality, never meeting again is far more common. And a structure where "they never reunite, only memories remain" is more realistic — and more painful.

Intention Disguised as Coincidence

On the flip side of "coincidence disguised as convenience," there's a far more interesting pattern: a character deliberately manufacturing coincidence.

You like someone. You know which convenience store they go to every day. You know what time they walk down which street. So you show up at the same time. Same cafe, same bookshop, same subway car. "Oh, you're here too?" Pretending it's a coincidence, but you've actually been waiting for thirty minutes.

Is this a cliche? Yes. But it's a completely different species from "coincidence disguised as convenience." Because this isn't a situation the scenario forced into existence — it's behavior created by the character's emotions. They mapped out the person's routine because they like them. They pretend it's coincidence because they lack the courage to be direct. The behavior itself reveals the character's emotional state.

Systemically, this is a concrete example of Part 13's "discrepancy between action and inner self." The inner state is "I want to see them," but the action is "pretending we bumped into each other." A textbook avoidance behavior for a character with low expression_skill.

And this can be applied to both the player and NPCs. A structure where the player can choose specific locations and times to engineer "chance encounters." Or a structure where the NPC changes their routine to meet the player, and the player discovers this later. When the line "That time we met at the shop — I was actually waiting for you" drops later — every past "coincidence" gets reinterpreted. The memory system's delayed consequence at work once again.

Coincidence created by the scenario is cheesy. Coincidence created by a character is endearing. Same coincidence, but who initiates it changes everything.

Concrete Methods for Stripping Away Cheesiness

A checklist organized with AI.

If "coincidence" appears more than twice, be suspicious. One coincidence is realistic. Two is convenient. Three reveals the author's hand. If characters need to meet, make them meet through their own will, not luck.

Compress your setting into one sentence. If that sentence includes a special ability or world-building, be suspicious. Ask "Does this story hold up without superpowers?"

Explain "why these two people drift apart" through personality, not events. Conflict must be created by internal reasons, not external events. Not "because the parents objected" but "because their expression styles differ." The action-inner self dissonance structure designed for characters in Part 13 becomes a natural seed of conflict here. Two characters with different expression_skill levels drift apart without any special setting.

Read the character's dialogue and judge "Does anyone actually talk like this in real life?" If AI-written dialogue reads like a novel, filter it out. People do not speak that cleanly. Words break off, they speak indirectly, they dodge the point, and they regret it later.

Break the "good choice = good result" structure. This is the most important one. The right choice always producing a good result is not reality. A structure where caring becomes hurt, honesty becomes burden, and patience becomes missing out — that is reality.

Raise the resolution of emotion. Do not stop at "likes them." Express complex emotions like "likes them but feels anxious," "likes them but feels burdened," "used to like them but is not sure anymore." This is why we built the three-layer emotion system. The disposition layer, momentary emotion layer, and relationship layer designed in Part 9 systematically guarantee this resolution. "Likes them but feels anxious" is a state where the relationship layer's affection value is high but the momentary emotion layer's anxiety value is also high. Because this compound state can be expressed through variables, dialogue and actions can also be made compound.

When I asked AI to "show me the concrete difference between a cheesy line and a realistic line," an interesting comparison came out. Cheesy line: "I can't live without you." Realistic line: "I can live without you, and I kind of hate that." The second is more complex and more real. Not a perfect declaration of love, but something honest yet somehow incomplete. This incompleteness creates realism. AI is good at this kind of comparative analysis, but when you actually ask it to "write me a realistic line," it reverts to clean sentences. A gap exists between analysis and generation.

The One-Line Takeaway

Uncontrollable Timing

The accumulation of choices alone isn't enough. In reality, there exists timing that the player cannot control no matter how determined they are.

After school, you've decided to confess to that person walking home alone. You leave the classroom, match their pace, you're about to speak — and another friend calls out "Hey, let's walk together!" and cuts in. Timing gone. You didn't do anything wrong, but the opportunity vanished.

You're walking together, the mood is right, this is the moment — and suddenly a downpour. You're running for cover and the confession evaporates. Or an earthquake alert goes off, or a siren blares, or a phone rings. The world intervenes.

This follows the same principle as the weather system. The reason Part 11 designed weather as an uncontrollable variable is exactly this. Not just weather — third parties appearing, sudden events, every external element that disrupts timing falls into the same category.

Systemically, this can be implemented as follows: the player selects "say it now," but depending on certain probabilities or conditions, an external event intervenes. An experience where timing collapses regardless of the player's will. This is the most natural way to break the "good choice = good result" structure. Nothing went wrong, but it didn't work out. In reality, that's what hurts most.

The Core Line

A non-cheesy love story comes not from "special events" but from "the accumulation of unspecial choices." And uncontrollable timing intervenes along the way.

The direction of the visual novel this project aims to create was finalized here. No superpowers, no fantasy, no dramatic events. Instead, daily small choices build a relationship, the world occasionally disrupts the timing, and the accumulation of choices and chance becomes love, or regret, or something that ends as nothing at all.

The time structure designed in Part 15 supports this direction. In a timeline stretching from elementary school to university, small choices accumulate in every period, and the meaning of those choices shifts as time passes. A structure where failing to say "Let's walk together" determines the direction of a relationship 10 years later. This is the power of narrative built through "the accumulation of choices." More realistic and deeper than dramatic twists created by fantasy settings — the gradual change made by everyday choices.

In the next part, we design endings. Not success and failure, but outcomes determined by combinations of states.

---

Next: Endings Are Not Success or Failure — State-Combination Outcomes