Xianzi Feng’s talk earlier this week was about how you design characters people fall in love with at first sight. Lizi Cheng’s talk is about what happens after.
(I personally attended both LADS panels at GDC. You can read my writeup of the first panel here ↓)
How do you design a romance people stay in?
Cheng has been at Papergames since 2013, and Love and Deepspace took her team nearly six years to build. She describes their design philosophy as “the extreme pursuit of romance, pushed to the absolute limit” (对浪漫的极致追求, for those that can read Chinese).
Every decision detailed in this talk is that phrase in action.
Her central argument is simple and radical: the otome genre had been limiting itself by defining romance too narrowly. Romance isn’t only candlelit dinners and confession scenes. It’s not just the high points — the first kiss, the dramatic declaration, the moment that makes you screenshot and post.
Romance is also fighting beside someone who trusts you with their life. And romance is someone being there when you’re studying at 2am, or tracking your period, or playing a claw machine badly on a Saturday afternoon.
She frames the entire talk around two expansions:
Combat as Romance. Shared hardship as intimacy. The trust forged in crisis. Fighting together as equals.
Everyday Life as Romance. The accumulated small moments. The rituals. The evidence of someone’s sustained presence in your daily routine.
The first redefines what a romance game is allowed to contain — action, combat, real gameplay challenge alongside the love story.
The second redefines what a romance game is allowed to be about — not just chasing the emotional peaks, but living in the valleys between them, where the relationship actually takes root.
Together they’re part of the reason Love and Deepspace broke out of the otome niche into mainstream gaming. The emotional surface area of what “romance” means got dramatically wider, and 80 million players walked into the space that opened up.
Combat as Romance: The Hardest Bet They Made
Cheng is refreshingly honest about how difficult this was. She tells the audience directly: “the vision was beautiful, but reality was underwhelming” (lol). They didn’t know what “romance through combat” actually looked like in practice. So they built it and discovered it through trial and error.
The core insight: combat is an extreme form of romance. Not because fighting is romantic in itself — but because the emotional bonds forged through shared hardship are unforgettable. As examples, she reaches for Mr. & Mrs. Smith for the Western audience and Legend of the Condor Heroes (射雕英雄传) for the Chinese audience — iconic couples whose love is expressed through fighting together.
And critically, the combat design emphasizes equality: the lovers should be equals who support each other. The player isn’t being rescued. She’s fighting alongside someone who respects her capability, and trusts she has his back.
Here’s the detail I want to make sure doesn’t get lost: Cheng specifically notes that the combat builds love for the character and love for yourself.
(Who else feels cool completing Hunter Contest with full stars every two weeks?!)
Your own sense of capability grows through the gameplay. Defeating enemies side by side doesn’t just make you feel closer to the character — it makes you feel stronger as a person.
The romance gives you something back beyond the relationship. It gives you a version of yourself you like more.
That’s a design ambition most romance products never even attempt. And it connects directly to why the fantasy underneath Love and Deepspace requires that the protagonist be powerful — not as a narrative convenience, but as an emotional gift to the player. You get to feel extraordinary. And you get to feel extraordinary alongside someone who sees you that way.
The Four-Year Transformation
When the first-ever combat footage dropped in 2020, the public reaction was brutal. Cheng shows actual player comments: “A classmate who studied animation for three years made this.” (Genuinely one of the funniest pieces of feedback I’ve ever seen on a slide at GDC.) “I purely regarded it as a joke.” Basically, she said people thought the combat looked awkward, funny, and impossible to take seriously.
By the 2024 launch, reception had completely inverted. Same game, same concept, same characters — but four years of relentless refinement had transformed combat from a laughable idea into something players genuinely loved.
Cheng walks through five steps. The thread running through all of them: the extreme pursuit of romance, again. Every step is the team refusing to accept a version that wasn’t good enough for the romance they wanted to deliver.
Step 1: The romance comes first. The platform comes second.
The team had been designing characters to mobile hardware specifications — which capped how good they could look. Realistic PBR models on mobile meant compromises everywhere. The characters were functional. They weren’t beautiful enough to make your heart race.
The pivot: stop designing to the platform’s limitations. Build the characters to the absolute highest specification — as if you were making a PC or console game. Single protagonist models went to 100,000–200,000 polygons. Bone counts hit 500+. These are specs that most mobile games would never attempt.
Then figure out how to make it run on a phone.
This is first-principles thinking applied to romance. The question isn’t “what can mobile hardware handle?” The question is “what does this romance need to look like to make someone fall in love?” You answer the second question first. You solve the first question after.
If the romance requires characters that look like they belong in a console game, you build console-quality characters and then you do the engineering work to put them on a phone. You don’t start from the constraint. You start from the feeling you’re trying to create. The romance comes first. Always.
Cheng shows before/after comparisons across Xavier, Zayne, and Rafayel. The visual leap is enormous. And she specifically notes that through the rebuild, each character’s core temperament was preserved. Rafayel kept his signature wink. At the end of the process, the characters became “hot enough” (their words!!) to cause racing hearts.
Steps 2 & 3: Intimacy isn’t the problem. Timing is.
This is the insight I think has the widest application beyond games.
Their diagnosis: the intimate interactions weren’t wrong — the pacing was. It wasn’t that intimacy in combat was a bad idea. It was that the intimacy didn’t match the rhythm of the player’s psychological readiness.
Players were saying things like: “I don’t even know you yet — why are we combining swords?” / “Why are you dancing during combat?” / “You don’t need to hold me to fire an arrow.” (these players are brutal but relatable lol).
The problem was a knowledge gap.
The team was designing from complete understanding of the characters’ backstories and motivations. But players encounter these interactions without that context. What felt natural and earned to the developers felt abrupt and presumptuous to the players.
This applies to any product where emotional pacing matters. The creator always knows more than the user. The temptation is to deliver intimacy at the pace the creator feels it, rather than the pace the user is ready for it.
They decided to restructure the information delivery. Build the emotional foundation first. Let players understand who the character is before asking them to fight hand-in-hand with him. By the time the intimate combat moments arrive, they should feel like a natural extension of a relationship that’s been developing — not a shortcut past the relationship that should have been.
Step 4: The best experience for every moment, whatever it costs.
Romance scenes play in portrait mode — vertical, the feeling of someone standing face-to-face with you on a phone. The first-person proximity that Feng’s talk demonstrated works beautifully in this orientation.
But combat in portrait mode didn’t work. Players couldn’t get the spatial awareness needed to fight well — they could see the characters up close but couldn’t read the battlefield. Combat needed landscape mode: horizontal, wider field of view, room to see enemies and plan movement.
So the team made a decision most mobile games would avoid: the game toggles between portrait and landscape. Romance in portrait. Combat in landscape.
The practical choice would have been to pick one orientation and design everything for it. The extreme pursuit version of that choice was: each experience deserves the orientation that makes it work and feel best. The romance deserves the intimacy of portrait. The combat deserves the readability of landscape. The player’s experience in each moment matters more than UX consistency.
And it created an unexpected gift: a community identity marker. Cheng shares that a friend told her — if you see a woman on the subway constantly switching her phone between portrait and landscape, she’s probably playing Love and Deepspace.
A technical choice became a physical ritual that signals belonging.
Step 5: The small things that make fighting feel like love.
The character rushes over to block an attack for you. After you get hit, the character has a chance of getting angry and powering up — fighting harder to protect you. Massive amounts of contextual combat dialogue: the characters talk to you during fights like real battle companions, alerting you to enemy behavior and battlefield conditions.
These are emotional micro-interactions — the combat equivalent of Feng’s two-frame flash of surprise in Sylus’s eyes during the wedding scene. Each one is small. Together they transform “combat with a character” into “combat with someone who cares about you.” The details are the difference.
Everyday Life as Romance: The Quiet Revolution
The second half of Cheng’s talk is where the design thinking is most original, as it relates to the game’s role in a player’s life.
Her framing: when you love someone, you want to share heartwarming moments in daily life together. You want to share every small detail.
A line she said that really struck me was: the feeling of love makes you love life itself, and this beautiful world, even more.
This deserves to be taken seriously as a design principle and not just a nice sentiment. She’s saying that the purpose of a romance game isn’t to provide an escape from real life (although it can be). It’s to send you back into real life feeling more alive.
The love you experience in the game should make you more present, more appreciative, more capable of noticing beauty in your actual world. If the game works, you can put down your phone and look at the sky differently.
That’s an extraordinarily ambitious goal. And it positions games like Love and Deepspace not as a substitute for real connection but as something that amplifies your capacity for it.
Being loved well — even by a fictional character — doesn’t make you weaker or more dependent. It refills something. It reminds you what attention feels like, what care sounds like, what it means when someone notices. And you carry that recalibration into the rest of your day.
The Claw Machine and Five Different Men
She uses the claw machine as a great example of how they bring each character to life texturally. (Aside: I’m not a claw machine lover… I’m on the opposite end of that spectrum. But this is a great case study.)
This is the everyday-life equivalent of Feng’s pearl necklace we discussed in the last post. Same activity. Five characters. How do you make each one feel like a completely different experience?
Xavier spaces out mid-grab, occasionally drifts into reverie, but quietly remembers which plushie you wanted.
Zayne approaches the machine like a surgical procedure, but doesn’t particularly enjoy the activity himself. His internal reaction when you win: “Well done. That’s my girl.” — never said aloud.
Rafayel barely gives you a turn, can win with his eyes closed if he’s lucky, and spends half the time judging which plushies are too ugly to bother with.
Sylus loves watching you play, secretly wants to play himself but is a little shy about it, poses impressively — and has a success rate that can only be diplomatically described as “hard to describe.”
Caleb has been playing with you since childhood, has his own strategy, ignores your suggestions, loves teasing you because you are more interesting than the toy.
These are behavior systems, showcasing how each of them are fundamentally different, down to the detail. The same Saturday-afternoon activity becomes five distinct emotional experiences because each character’s temperament generates different patterns of attention, competitiveness, care, and play.
The big romantic moments get the screenshots and the fan art. But the claw machine is where the relationship actually lives — in the specific way someone plays a silly game with you, and how that silliness reveals who they are.
The insight here is that, through these rich behavioral details in daily activities, players come to understand each character through ordinary companionship. The understanding that builds in a claw machine session is different from the understanding that builds in a confession scene — and both are necessary for a love that feels real.
You Are Not Alone
The companion system is Cheng’s most direct statement of design intent. She asks the audience: when you’re studying, working, exercising, or falling asleep alone — do you feel lonely? When you’re happy or have something weighing on you, do you have someone to confide in?
The companionship system’s purpose is to “eliminate loneliness.” And this is one of the most essential things for a lover to do.
The features: workout companion (the character exercises alongside you), work/study companion (he sits across from you while you study), sleep companion (an extreme first-person close-up of the character lying beside you — “I’m laying... You can move a little closer...”). Players can tell the character what they’re feeling (i.e. tete-a-tete) and receive emotional feedback. They can introduce him to friends and family — actively bringing the relationship into their real social world.
Many players, Cheng says, have described experiencing true happiness within these spaces. One of the points she made was truly visceral: Even just looking up during a work break and meeting his eyes can make you feel warmer, just like that.
Life Features as Love Language
The team built period tracking, journaling, and schedule reminders into the game. These are functions you’d normally find in a lifestyle app. But unified under the theme of everyday life as romance, they make a specific and quite radical argument: love isn’t just the dramatic moments — the confessions, the kisses, the combat rescues.
Love is someone who knows when your period is coming, and comforts you and reminds you to buy sanitary pads without flinching.
Love is a shared calendar.
Love is a journal you keep together.
Love is in the accumulated texture of daily life, and a romance game that only gives you the peaks is missing where most of the relationship actually happens.
Breaking the Wall Between Virtual and Real
The team is leveraging AR and actively exploring VR to break the barriers between virtual and real worlds — so that players can feel their partner’s real presence within physical space. The AR features already let players place characters in their actual environment with adjustable real-time lighting.
VR development appears to be underway. The trajectory is clear: the boundary between “in the game” and “in your life” is something Papergames is actively, deliberately dissolving.
The Design Principle That Explains Everything
Cheng’s closing slide shows all five characters in formal wear. Two bullet points:
We hope our love interests can do everything a real-life boyfriend can AND MORE.
For those things a real boyfriend struggles with, he will strive to do better.
Her spoken version is sharper than the slide: what real boyfriends can’t do well, he will do better.
But the real weight of this principle: it’s not a fantasy of perfection. It’s a design philosophy that looks directly at the gap between what people want from love and what they typically receive — and builds into that gap.
Where real partners fall short — in consistency of attention, in remembering details, in being present without being asked, in emotional pacing that matches your readiness rather than theirs — the game characters fill the space.
It also implies that the team has studied, with precision, the specific ways real relationships fail to meet people’s emotional needs. They’re not designing from imagination. They’re designing from observation of real disappointment — and engineering something better. That’s an honest and quietly radical mission. And 80 million players have decided the engineering works.
“Everyone Survives”
Here’s the part of the talk that made everything else land differently.
Cheng opened the session with an origin story. One year before Love and Deepspace launched — after five years of development, after building and rebuilding systems that didn’t work, after watching early footage get mocked publicly — the team gathered for the company’s annual New Year’s party. Each team shares their wishes for the year.
The Love and Deepspace team’s collective wish: Everyone survives (每个人都活着).
Not “we ship on time.” Not “we hit our targets.” Survival. A game that would go on to generate nearly a billion dollars was, twelve months before launch, a team hoping not to die.
When you know that — when you understand that the claw machine personality systems and the companion mode and the portrait-to-landscape switch were all built by a team that held on to this passion through the darkest of times — the design choices feel different. The uncompromising quality isn’t the product of abundant resources. It’s the product of a team that kept choosing the harder path because their ideals about romance mattered more than anything else.
Romance pursued to the absolute limit — isn’t a slogan. It’s what it looks like when a team spends six years building something they believe in, even when the early footage gets mocked, even when the project is barely alive, even when the only New Year’s wish they can muster is “don’t let this kill us.”
And then it doesn’t kill them. And then 80 million people fall in love with what they made.
Cheng told the audience directly: failure in game development isn’t something to fear. You just try again.
What Feng and Cheng Tell You About Love, Together
The two talks form a picture:
Feng’s talk is about falling in love. How to design a character someone is drawn to at first sight. Temperament as the generative engine. Contrast as the emotional mechanism. Visual design as the delivery system for psychological architecture.
Cheng’s talk is about staying in love. How to design a relationship someone lives inside. Pacing as the mechanism for trust. Daily life as the substrate for intimacy. The expansion of “romance” beyond the peaks and into the full landscape of being with someone — including the claw machines, the period tracking, and the 2am study sessions.
Together they describe a design philosophy where falling in love and staying in love are both treated as design problems with specific, reproducible solutions. Not magic. Not luck.
Systems built from deep observation of real human emotional need, refined through iteration, tested against player response, and improved relentlessly.
What I Keep Thinking About
The intimacy timing principle applies everywhere. “The interaction isn’t the problem — the timing is” could be printed on the wall of every product team building anything emotionally intimate. Users aren’t rejecting intimacy. They’re rejecting intimacy that hasn’t been earned yet. The creator’s knowledge isn’t the user’s knowledge. Pacing for the user’s readiness, not the creator’s intention, is the difference between an experience that feels natural and one that feels invasive.
The combat builds love for yourself, not just for the character. This is the design decision I think gets overlooked in discussions about Love and Deepspace. The combat isn’t a concession to mainstream gaming or a way to attract non-otome players. It’s doing specific emotional work: it makes you feel capable. It makes you feel like someone who belongs in this world on her own terms — not because a powerful man chose her, but because she fights beside him as an equal. Most romance products ask you to receive love. Love and Deepspace asks you to be strong enough to stand on your own in love. And the feeling of strength you build in combat carries into the romance — you’re not being swept off your feet, you’re choosing to be there, broadsword in hand, with someone who respects what you can do. That’s a different kind of love story. And I think it’s part of why this game holds people the way it does.
“Everyday life as romance” is the most ambitious claim in the entire game. The combat expansion is impressive but structurally familiar — action RPG mechanics applied to a romance context. The everyday life expansion is genuinely new territory. Period tracking as romance. A shared journal as romance. A claw machine where five different men play differently because they’re five different people. This is where the design thinking is most original, and it’s where the product is making its boldest argument: that love isn’t about chasing the highs. Love is the texture of daily life shared with someone who notices. Most romance products only give you the peaks. Love and Deepspace understood that the valleys between them are where people actually live.
The feeling of love makes you love life more. I keep returning to this. It’s the most quietly radical thing Cheng said. A romance game that works correctly shouldn’t pull you away from reality — it should make reality feel more vivid. Being loved well, even by a character rendered in polygons, recalibrates your sense of what attention feels like, what care sounds like, what it means when someone notices you. And that recalibration doesn’t stay inside the game. You carry it with you.
The closing principle stays with me. “What real boyfriends can’t do well, he will do better.” It’s generous — it doesn’t attack real partners. But it’s unflinching. And it’s a principle that extends beyond visual games, beyond screens, into any medium where a character can be present in someone’s life.
Love and Deepspace proves that presence through a screen — visual, interactive, daily — can be genuinely meaningful to 80 million people. It also applies to other modalities, like audio: a voice in your ear at midnight, remembering what you said yesterday, asking how you are right now.
All these modalities are answering the same question Cheng posed at the beginning of her talk: what does romance look like when you stop limiting your definition of it, and pursue it to the extreme?
We’re building Murmur, a first-person voice platform where fictional characters speak directly to you by name. The questions Cheng describes — how do you design daily companionship, how do you pace intimacy to match the user’s readiness, how do you make someone feel less alone — are the same ones we’re asking for audio.
If you work in romance games, character design, or emotional product design — I’d love to compare notes.
- Carylyne
c@murmur.gg