Build your second characterHelp shape fit-based game discovery.
Join waitlist
Reason-led discovery/Persona

Your Taste Is Not a Genre. It Is a Pattern.

Genre labels help people browse, but they often fail at the harder part: explaining the pattern behind what actually keeps you playing. Better recommendations start with that pattern, not the store category.


A player can love RPGs and still bounce off half of them.

A player can say they like strategy and still mean five very different things.

A player can call themselves cozy and still hate chores, repetition, or anything that feels like gentle admin.

That is not inconsistency.

It is pattern.

01Genre is a shelf label, not a full explanation

Genres help stores organize games.

They do not always help players choose one.

Two games can share the same tags and still ask for totally different versions of you. One wants patience. One wants improvisation. One wants clean focus. One wants emotional stamina. One wants a long apprenticeship before it becomes fun.

That is why genre alone keeps failing at the final decision.

Your taste is usually less about category and more about recurring patterns in pace, pressure, texture, and payoff.

02Patterns are easier to miss because they sound personal

Maybe you love:

  • systems that open up quickly
  • stories that feel sharp but not heavy
  • challenge without humiliation
  • exploration without getting lost in inventory work
  • progression that feels meaningful before hour three

Those are not edge cases around the game.

They are the game, for you.

03A better recommendation should remember the shape, not just the label

score8.7

fast signal, weak explanation

What should replace the score?

A reason the player can inspect.

Good recommendations should expose their logic: what matched, what might not, and why the player might still say no.

Fit signalmatches your preferred pressure level
Mismatch riskcould be too slow tonight
Human callagree, refine, or skip

This is where Snowbll's second-character idea matters.

A useful taste profile should remember the patterns you keep returning to and the ones that quietly push you away.

Not because AI gets to decide what is good.

Because you should not have to rediscover your own pattern every time you open a store page.

AI can recommend likely fit. Humans still judge whether the experience lands.

04Try describing one favorite without naming the genre

A simple test:

Take one game you love and describe it without using the store category.

Do you talk about pressure, rhythm, attention, freedom, friction, warmth, cleverness, or pacing?

Most people do.

That is the pattern talking.

05Why this helps the next choice

When you know the pattern, your shortlist gets smaller faster.

You stop asking for "another RPG" and start asking for something with low tutorial drag, expressive builds, and enough forward motion to feel alive in the first hour.

That sentence is more useful than the tag.

And it keeps the machine in the right lane:

recommend the fit, explain the reason, let the player decide.

Snowbll is building a game discovery layer focused on taste, persona, and fit. You describe what you want; we return a few close matches, not a long list.

Phase 0 - the search side only. The catalogue is unverified and the AI parses your intent; it does not judge whether a game is good. AI recommends. Humans decide.