Variance Beta May Update

What we've built, what changed, and why it's going to be worth the wait.

Dwight D. Eisenhower famously said: “Plans are worthless, but planning is everything.”

In February we outlined a plan for Variance Beta - Closed Beta at the end of March, Open Beta a month later. We worked hard to hit it. And somewhere along the way, I realized that hitting it wouldn’t give us a game that was actually fun - and we’d set ourselves a high bar with Variance Alpha.

I always imagined Beta as a major step up that keeps everything that made Alpha good. But Beta is a new game, with big new pillars: mobile gameplay, character progression, and large procedural levels. Each one is its own challenge. We could have shipped in April with all of them in place - it just wouldn’t have been fun yet.

We’re building a game where you fight familiar enemies in unfamiliar ways. Remember our big spider enemies, the Wardens? Imposing, but never that dangerous - dash behind one, slash it to pieces, done. In Beta, with a whole party of characters in play, an enemy that simple isn’t enough.

This time the Warden has been reworked so you can’t just slip past it anymore, and imagine right now it has Brok pinned. He does everything he can, burns the last charge of his shield, and he’s about to go down... and then Raya comes in hot, blazing past Claws and Orbions, carving into the Warden’s back with a flurry of lightning-fast strikes. The Warden registers the new threat just as it’s about to break Brok’s shield - and turns to face her. That’s the opening Brok needed: he can finally charge his ground-shattering Hammer Blast.

Now imagine you’re the one playing Raya, and all of this is happening on mobile. Intuitive controls, just a few buttons - with optional auto-melee to back you up. An interface clean enough that you knew, at a glance, that Brok was in trouble: where he was, and how close to the edge.

That is what we want the Beta client to deliver.

Our February plan called for desktop-first gameplay, procedural maps, and 3D - but it never put fun first, and fun is the whole point. Measured against what actually matters, the plan turned out, as Eisenhower says, worthless. And yet without it we’d never have reached the point where the real game can even take shape.

We keep working, building, iterating, improving.

It’s taking more time than we first planned - and it will be worth the wait.

Why we’re not at open beta yet

Two decisions matter more than the rest.

Mobile-first

Variance Beta was always meant to have mobile. The engine was built from day one to handle it, and we’ve tested that it can. What changed isn’t whether - it’s when. We flipped the priority: mobile leads, and the beta ships there first.

Why? Because if both clients take roughly the same time to build, mobile reaches millions more players. And a beta with more players means better feedback, a tighter loop, and a stronger game.

Thanks to that new direction - we also spent time iterating more on our internal infrastructure to prepare it for cross-platform gameplay.

PC is firmly in the roadmap. The PC client you’ll get will be the higher-fidelity version - same gameplay, more visual headroom - built on everything the mobile beta teaches us. It just isn’t part of this first beta window.

Procedural generation

We knew procedural level generation was a must-have for Beta. We accounted for the technical challenges of building the generator and authoring its content. What we didn’t fully account for was how much iteration it takes to make procgen feel great - not just work correctly.

Beta levels are large. But a large level is only interesting if there’s something in it to explore. Our answer is the roguelike layer we’re building on top of those levels: chests to chase, trinkets that reshape a run, and between-level beats that give a long level rhythm. The question wasn’t whether we could make it work - it was figuring out the right questions to ask. We found them, and the answers make the game significantly richer.

This is the work that took the most time, and we’re proudest of it. The chunk system doesn’t just solve today’s level design - it means we can add new features in future updates without redesigning entire levels. It cost us time, and it bought a foundation we’ll build on for years.

And some legitimate tangents

Mobile and procedural generation are the big lifts, but they aren’t the only things that moved over the last three months. Three more are worth a mention - and two of them will get a full post of their own later.

• Art pipeline
  We’d been working with external studios on art, and over time the operational overhead grew heavier than the art itself. So we made a call: we’re better off building the pipeline ourselves, with Kodo. It’s faster, it produces results that fit Variance better, and everything we build along the way feeds straight into a product we’ll ship later. That story deserves its own post, and it’ll get one.

• How the team works
  Life happens - and over this stretch it happened to nearly everyone on the team to a varying degree, leaving less time to go around. Rather than stretch a leaner team thin, we focused: we trimmed our product portfolio down to what truly matters and put our weight behind tooling we own and control, so we can do more with less. That refocusing took real time, and it’s part of why the update moved - but the team on the other side of it is steadier and more focused than before. The fuller story is its own post coming soon.

• Variance Alpha
  It used to open only on weekends. Since March it’s permanently available - jump in any day, whether you’re new and curious or just want something to play while Beta cooks. We wrote up the details when it went live.

None of these are the headline - but they round out the picture of where the last three months went.

Where we are

Variance Beta is what Variance Alpha was meant to be. It builds on Alpha’s foundations while adding a whole new layer of content and systems.

• Levels, for example. Alpha had a set of hand-authored levels that were strung randomly together to create a run. That approach worked for Alpha - but we knew the next step was real scale. We made the call to invest in procedural generation because hand-authoring has a natural ceiling, and run variety with a fixed level pool eventually plateaus.

  Beta delivers on that investment. Beta levels are very large and composed of “chunks” - small pieces of a level that can slot anywhere. We author the chunks, the generator assembles them. Two runs through the same biome now play differently - different routes, different fights, different things to find - and it scales the way hand-authoring never could.

  We’ve also expanded the toolkit available for level design. Where Alpha had, for example, glowing beams that could be turned on and off, Beta has proper moving platforms, elevators, one-way platforms, damaging hazards, and more.

• Combat is maturing too. We’ve kept the same customizable weapon and projectile system and made it far more robust - projectiles can now interact with each other, so beams clash and a well-timed shot can answer an incoming one. Arcing, physics-driven projectiles land where you aim them, and the whole system supports far more interesting weapons and abilities with a much cleaner feel. Characters now share an advanced health model - health, armor, and shields, each with its own function - which becomes another lever that sets one character apart from another.

• Enemies have leveled up. Alpha enemies could follow you across even ground; Beta enemies jump across gaps to reach you and pick their attack based on where you are - so kiting from one safe distance no longer works. Familiar foes like the Wardens have been reworked to match. Alpha tactics will need an upgrade.

• Visual style keeps getting refined. We’re merging 2D and 3D techniques - preserving the anime aesthetic of Alpha while leaning on 3D for better visual clarity and modern features like dynamic lighting and post-processing.

If you took the codebase from February and the codebase from this week and put them side by side, you’d see a different game. The extra time isn’t a sign of slow progress. It’s two specific scope decisions we made to make the beta worth your time.

The journey continues

Here’s what we’re planning to do.

Closed beta first, with a window of mid-to-late Q3 2026. Closed beta is for the core community - the players whose feedback shapes what open beta becomes. Leading with a smaller, focused test gives us a faster iteration loop and a better open beta. Closed-beta-first is how we make open beta better.

Who’s in: Closed beta is open to Genesis Portal holders and players with 10,000+ Aura - the core community we committed it to from the start.

What closed beta will have:

• A mobile-first build (PC client comes later, not in this beta)

• Procedurally generated levels - large, varied, and full of things to discover

• Two playable characters - Brok and Raya - with a third arriving at open beta

• Co-op multiplayer with friend-invite joining

• A proper roguelike loop with chests, trinkets, and scoring

• Between-run character progression - simple to start, expanding over time

What follows closed beta:

• Open beta, adding a third playable character, matchmaking and the broader feature set

• We’ll name an open-beta date once closed beta gives us the confidence to set one. We’d rather hit the next milestone we name than guess at the one after it.

To borrow a line from our February update: consistency over specific schedules. We set an ambitious April target and learned from it. We’re setting a mid-to-late Q3 window now, with real room built in.

---

Follow Variance on Variance to get updates directly in your inbox.

Join the conversation on our Discord