The desert arrived after the moves

The Windswept Wastes in Ori and the Will of the Wisps, 2020, reads as a desert level, and the desert is the last decision, not the first. The level started as mechanical problems: a burrow, a gust, a threat that lives in the ground, a space that wants a dash and then punishes it. Sand, ruins, and colour arrived to give those problems a single place to stand.

The other order is how levels get pretty and empty. A theme is chosen, art is made, and the verbs are asked to take a tour of the art. Players remember the tour for a minute. They remember a movement problem that could only have happened in that space for years. If you cannot state the problem without the word desert, you do not have a level yet. You have a backdrop.

The Windswept Wastes in Ori and the Will of the Wisps did not start as a desert. It started as a set of mechanical problems, and the sand arrived later to hold them together.

The finished area holds a treacherous climb, platforms that fall apart underfoot, digging through sand, enormous laser beams, and a closing chase away from a monstrous worm. None of it came from asking what a desert should contain.

One of the first jobs was narrower than a whole area: find level-design situations worth building around platforms that crumble the moment the player lands on them.

At that stage the element was almost nothing: a rough prototype, and one question repeated over and over. How many distinct situations can a falling platform produce?

Some of the sketches reused a single platform twice. Jump on it, stay off the ground, cross to the far side, then use it again once it reformed to reach higher. Another sketch had the player fall deliberately, recover mid-air, and double jump back out.

The habit behind all of this is workshopping. Take one piece you already have, learn its properties, squeeze it against constraints, and hunt for arrangements that show you something you had not seen.

The strongest idea to come out of that workshopping asked the player to destroy the platforms deliberately, so a projectile could travel the length of the room.

Normally a collapsing platform is a threat, because standing on it costs you the ground you are standing on. In this room the vanishing is the point. Every platform has to be cleared away without the player ever touching the floor, so the shot has an open lane.

That inversion is what the process is for. One object can serve as floor, timer, hazard, switch, route, and finally as an absence you create on purpose. Nobody reasons their way to that list; they build small arrangements until it appears.

Burrow, the ability that defines the area, was not invented to fill a desert. The desert was chosen partly because the ability needed somewhere to live.

The question being asked was much wider: what fresh kind of movement could sit alongside a moveset that was already deep? Digging looked promising for a Metroidvania, and part of the inspiration came from a traversal idea in Sonic Colors.

Early sketches used burrow to open up spaces Ori could not otherwise reach. Spikes blocking a wall jump stopped being a wall, because you could tunnel through sand and launch out past them. A ledge above your jump height stopped being a ceiling, because you could dive into sand, aim, and fire yourself over it.

Suspending balls of sand in the air stretched the move further. Enter from whichever side you like, launch out along a straight line, and every pocket of sand becomes both an aiming device and a source of momentum.

The sketches that mattered most were the ones that fell apart without the new move. That is a fair test for any ability. It should not behave like a key cut for one door. It should change how a player reads space and how they cross it.

Sketches only take a move so far. The next step was rebuilding those situations as blockouts and asking the programmers and animators for something playable.

That handoff is where an imagined ability turns into a control problem you can actually feel. The first working burrow was slower than what shipped, and it turned less crisply.

That sluggishness was intended. Sand was meant to be slightly awkward to move through, with some drag on it and a little of Ori's usual nimbleness taken away. Playtest after playtest argued the other way, and the move kept tightening.

The deciding factor was the rest of the game. Ori is precise everywhere else, and a loose move in the middle of it reads as a fault rather than as texture. Burrow still feels unlike anything else in the moveset, but it now speaks the same language as the character around it.

Crumbling platforms and burrowing began as unrelated ideas. What joined them was a problem that only shows up once you try to lay out a Metroidvania.

An early challenge hung collapsing platforms over lasers. Land, wait, glide, and drop through the beam on the right beat. As a sequence it worked, and it worked exactly once.

A linear platformer can live with that; the floor falls away behind you and the game keeps moving forward. A Metroidvania cannot, because players are expected to come back and cross the same room later with more in their pockets.

The fix was to make the collapsing surfaces out of sand. The platforms could still break away and reform, but on the return trip the player could simply dig straight through them.

That one change bought a great deal. A collapsing platform stopped being only a temporary floor and became a sand surface, a burrow route, and a piece of traversal that reads in both directions.

Once the player has burrow, the area deliberately stops pointing. That creates a problem of trust: how do you leave people free to experiment without losing them?

The answer is convergence. The Windswept Wastes offer several ways back to the main route, so a player can wander, work out what burrow is really for, and still end up where the level needs them.

The result reads as openness while quietly staying a route. Poking around is genuinely rewarded, and every direction still drains toward the same place.

A difficulty curve is not a slider. It comes from knowing which layers of complexity live inside a single mechanic, and introducing them in an order the player can absorb.

Burrow has a very easy floor. Stand on sand, hold down, press the button. Aiming a launch out of a hanging ball of sand sits far above that, because it asks for trajectory, timing, and clean execution all at once.

Placing those layers is the designer's judgment call. What can somebody handle the moment they get the move, what should wait until it is familiar, and which combination is asking for mastery rather than comprehension?

One rule worth keeping is not to burn through a mechanic's whole ladder in a single stretch. Teach every escalation back-to-back and then walk away from the element, and the level starts to feel like coursework. Interleaving mechanics and revisiting them later produces both more variety and a curve that feels less deliberate.

The area is dense with secrets, and almost none of them are arbitrary. A secret is doing its job when finding it feels like reading the room correctly.

A hidden passage might be given away by a small lip that pulls the eye. A breakable wall might be drawn slightly differently without announcing itself. The clue is there to be caught, and catching it is left to the player.

The aim is never to make somebody leap off the screen and hope. It is to let them notice something out of place, test the hunch, and be right.

In a Metroidvania that matters more than usual, because hunting for things is most of the pleasure. A hidden reward should pay off attention, not paranoia.

One room asks the player to tunnel through sand while a giant laser beam sweeps the space. It was among the earliest prototypes built for the area, and it lasted because it dramatized what burrow is for.

Tuning it was a fight. It went through repeated iteration and internal argument, because the first versions asked a lot. It stayed because it ramps the player toward the harder burrow challenges ahead, and because people remember it.

There is a general test buried in that. A hard room is far easier to defend when it is expressive as well as hard. It should make the case for why the ability exists.

The second act of Ori and the Will of the Wisps lets the player take several areas in a flexible order, and every bit of that flexibility lands on the level designers as a constraint.

Players show up carrying different loadouts, which means solutions nobody planned. A challenge built around the sand worms, for instance, can be sidestepped entirely by somebody who already has the right tool from elsewhere.

A fair amount of that was simply accepted. If a player solves a room with something they earned in another region, and the solution is still interesting to watch, that is what freedom costs and part of what it is worth.

Balance was the harder half. Nobody wanted an area that told players they had come too early by beating them up, through enemies or challenges pitched beyond them. If the game let you walk in, you should be able to finish what you found.

Windswept Wastes breaks that rule slightly, because it eventually gates story progress. You can arrive early, explore the desert, and leave with burrow, but at a certain point the game asks you to go elsewhere before it continues.

A hard stop like that risks feeling flat, so the level cushions it. A fast-travel point sits near the sealed temple, and the player still leaves with a new ability, upgrades, rewards, and a stretch of play behind them.

The Windtorn Ruins near the end of the area were meant to be a dungeon, built around a second new ability called drill, which would have bored through thick rock blockers.

Drill was cut for lack of versatility. It kept looking like a key for one type of door rather than a way of moving that changed how the game was played.

In its place came a more linear, story-driven stretch, and the pacing improved for it. The player has just come off a run of demanding burrow challenges, and one of the game's hardest escapes is waiting on the other side. A breath belongs between them.

Not every level has to climb the whole way. Sometimes the correct move is to drop the mechanical intensity so the next spike has something to rise from.

The closing escape has to be hard enough to be thrilling, and no harder, because past a certain point excitement turns into trial and error.

A reasonable target is that most of the sequence flows on sight: the player reads the screen, understands most of what is coming, and performs under pressure. A smaller slice can hold surprises.

Even the surprises have to be reactable. If something drops on the player's head with no readable warning, the death registers as cheating rather than as a lesson. The feeling you want is: I saw that too late, and I will not next time.

Dying once or twice is part of learning a sequence like that. Dying five or ten times is the pacing coming apart.

Far more was designed for this area than could ship. A heavy block that would drop once you tunneled the ground out from under it. A bigger worm trailing acid behind it. Thin sand platforms that flung the player off like trampolines.

Some were expensive to build. Others were novel without being better. The work was finding what players could grasp and enjoy inside the constraints the level actually had.

That narrowing is why burrow carries the Windswept Wastes so well. It is not desert set dressing. It shapes traversal, secrets, backtracking, the difficulty curve, the set pieces, the unplanned solutions, and the escape at the end.

A game's best level usually comes out of that kind of subtraction. Many sketches, many prototypes, many good ideas left on the floor, and one mechanic strong enough to hold the whole space together.

In the Flockbay app, list three movement problems before you name a biome. Ask for a chain of rooms that pose them. Play the chain in greybox, or as close as the build will get. Only then ask for the sand. If the sand version is not clearer, the theme is in the way.

An AI metroidvania maker and a metroidvania template are the genre. Three rooms, not a map.