Powder Game: the website that turns pixels into physics

Powder Game is a free browser toy from the Japanese studio DAN-BALL that drops a pinch of sand, water, or fire onto a blank canvas and lets physics take it from there. It has been running at dan-ball.jp since 2007, and it is worth a look because almost nothing else from that era of the web still works, let alone keeps shipping updates.

In brief

• Powder Game is a falling-sand physics sandbox: pick an element, paint it on, and watch it react.
• It was made by the developer ha55ii under the DAN-BALL label and has been live since 2007.
• The current build offers more than forty elements and a built-in gallery where players upload and vote on creations.
• It is the game Sandspiel's creator names as his primary inspiration, which makes it a quiet ancestor of the modern falling-sand revival.

What the site actually is

The page loads a single square canvas and a dense grid of buttons. There is no tutorial, no account wall, and no splash screen asking for a newsletter signup. A visitor picks an element, holds down the mouse, and draws. Sand falls and piles. Water flows and finds its level. Fire spreads to anything flammable and gutters out when it runs out of fuel. That is the whole premise, and it has not needed changing in close to two decades.

The studio page lists the current release as Powder Game ver9.8, carrying a "Copyright(C) 2006 DAN-BALL" line and crediting ha55ii alongside collaborators named tylor and otascube. The version number is the tell. This is not a frozen relic but a toy that has been revised steadily, element by element, for years. Each click paints a patch of the chosen material, and the right mouse button holds a second element so a visitor can, for instance, carve a channel of water with one hand and set a fire with the other.

The interface looks intimidating at first, a wall of tiny labelled squares, but it resolves quickly. There is a speed control, a pause, and a set of number keys that switch between tool modes. Nothing explains itself in a paragraph. Everything explains itself the moment it is tried, which is the oldest and most reliable tutorial a toy can offer.

The aesthetic is period-correct and proud of it. The canvas is small, the palette is bright and basic, and the pixels are large enough to see individually. None of that has been smoothed over in the name of looking modern. A visitor arriving from a web of gradients and rounded corners meets a flat grid of primary colours that could have loaded in 2007, because in every way that matters it did.

A simulation built one cell at a time

Everything happens on a pixel grid. Each cell holds one material, and every frame the simulation asks a small question of each cell: what are you, what is next to you, and what should happen next. A grain of sand with empty space below falls into it. A grain with water below swaps places. Fire next to oil becomes more fire. Repeat that a few thousand times a second and sand behaves like sand without anyone ever writing a rule that says "this is a dune."

This is a falling-sand game, a close cousin of the cellular automata that computer scientists have poked at since Conway's Game of Life. The difference is that Powder Game never asks a visitor to care about the theory. The emergent behaviour, dunes, floods, firestorms, is just what happens, and the pleasure is in watching simple rules produce complicated results.

The elements, and why there are so many

Powder Game ships with more than forty placeable materials, which is a lot for a toy that fits in one browser tab. The basics are intuitive: powder, water, oil, salt, and ice. Then it gets livelier with fire, gunpowder, C-4, nitroglycerin, and magma, which exist mostly so that the careful structure a visitor just built can be removed in a satisfying hurry.

Other elements are quieter and stranger. Seed grows into vines and trees. Cloud rains. Thunder arcs between conductors. There is metal, glass, stone, and fuse for building and wiring, and there is a clone tool that spawns an endless supply of whatever touches it. The roster reads like a chemistry set assembled by someone who was never told to stop adding things.

The interactions matter more than the count. Salt melts ice. Acid eats through most solids but not glass. A laser passes through some materials and reflects off others. Because every element knows how to react to every other element, the combinations multiply far past the forty-odd buttons on screen. A visitor who learns that gunpowder carries fire along a drawn line has, without noticing, learned to build a fuse, a timer, and a trap.

The parts that move on their own

A second category is less about materials and more about agents. Powder Game includes a small stickman "player" the visitor can steer, plus ants, birds, and fighters that wander the canvas and interact with whatever they bump into. The ants follow walls. The birds flock. The fighters pick fights. Add a fan and the air itself becomes a force, pushing powder and feeding flames.

The result is that a scene is rarely static. Build an enclosure, drop in a few ants and a food source, and the thing starts to look like a terrarium that runs itself. Leave it alone and it keeps evolving, which is a large part of the appeal. Powder Game rewards setting something in motion and then simply watching, an increasingly rare invitation on a web that mostly wants a click.

Wind, pressure, and the illusion of a world

What separated Powder Game from the earliest sand toys was air. The simulation tracks wind and air pressure across the grid, so a fan does not just nudge nearby pixels, it builds currents that ripple through the whole scene. Smoke rises and drifts. Explosions produce a shockwave rather than a tidy puff. Fireworks need lift to climb.

This is the detail that made the toy feel like a world instead of a screensaver. A player could build a windmill, point a fan at it, and watch the blades turn under pressure they could not see but could clearly feel in how everything else behaved. The physics stayed cheap enough to run on a mid-2000s computer, yet rich enough to reward hours of tinkering.

Pressure also made the destruction legible. An explosion in a sealed box behaves differently from the same explosion in open air, because the simulation has to route the shockwave somewhere. Players learned to channel blasts, to vent them, to aim them. That turned demolition from a reset button into a design problem, which is a surprisingly deep thing to find in a free browser toy with no instructions.

The gallery that turned a toy into a community

Powder Game was never only a solo sandbox. The game includes an upload system: finish a scene, save it, and it joins a public gallery that other players can open, run, vote on, and then fork for their own tinkering. The studio page describes weekly, monthly, and per-user rankings, with rules that stop people from voting for their own work and that retire older entries from the charts.

That loop, make a thing, share it, watch strangers react, improve it, is ordinary now. In 2007 it was not. Powder Game quietly built one of the web's early user-generated-content communities around a physics toy, long before "UGC" was a term anyone used in a meeting. The uploads ranged from mechanical contraptions to pixel art to crude animated jokes, and the ranking boards gave the whole thing a pulse.

The gallery also doubled as a school. A newcomer who could not work out how to make a working cannon could open a top-ranked one, run it frame by frame, and read the design out of the pixels. Skills spread through the community by imitation rather than instruction, the way they do in any workshop where the finished work is left out on the bench for everyone to study.

The studio behind the canvas

DAN-BALL is small, and Powder Game is only one of its browser toys. The same studio kept iterating on the idea for years, releasing sequels that reviewers covered well into the following decade. The games portal Jay Is Games, a long-running chronicle of browser games, has published walkthroughs for later entries including Powder Game 4, which places the series firmly in the category of a project that outlasted most of its contemporaries rather than a single viral afternoon.

The restraint is part of the charm. Nothing about the site has been redesigned to chase a trend. The layout is the same cramped, functional grid it has always been, which is exactly why a visitor who last played it in school finds it instantly familiar. The studio treated the toy as something to maintain, not something to relaunch, and the web is quietly better for it.

Where it sits in the falling-sand family tree

Powder Game did not invent its genre, but it did a great deal to popularize it. According to Wikipedia's history of the falling-sand game, the form traces back to a Java applet posted on a Japanese blog around 2005, which gave the category its name and set off a wave of imitators. Powder Game arrived soon after and expanded the formula with its pressure model, its roving characters, and its sharing gallery.

The clearest proof of its reach is Sandspiel, the glossy modern falling-sand toy. In his essay Making Sandspiel, published in 2019, the developer Max Bittker thanks ha55ii's Powder Game as the primary inspiration for his own project, and describes growing up inside the Powder Game community of roughly 2007 to 2013. Bittker reports that Sandspiel went on to collect more than 70,000 shared creations, a direct echo of the gallery model Powder Game pioneered. One toy's childhood habit became another toy's founding idea.

The lineage runs wider than one tribute. The falling-sand idea seeded a whole shelf of projects, from the open-source desktop favourite The Powder Toy to the acclaimed roguelike Noita, which built an entire game around every pixel being simulated. Powder Game is not the root of that tree, but it is the trunk through which a generation of players, Bittker among them, first discovered that a pile of pixels could behave like a world. Its influence is out of all proportion to its tiny, unchanged footprint on the web.

Why it still runs when its peers are gone

Most of the web that a 2007 teenager loved is unreachable today. Flash is dead, Java applets are blocked by every modern browser, and the hosts that served them have long since lapsed. Powder Game survives because DAN-BALL kept maintaining it and migrated the game to run on a plain HTML canvas, so it still opens in a current browser with no plugin and no fuss.

That survival is the thing commenters keep remarking on. When Powder Game resurfaced on Hacker News in June 2024, the person who posted it described sinking hundreds of childhood hours into it, and a reply simply said they were amazed it was still available. An earlier 2021 thread carried the same note, one parent mentioning fond hours spent playing it with their kids. The game outlived the nostalgia cycle and came out the other side still functional.

There is a lesson in that for anyone who builds things on the web. Powder Game made no attempt to be a platform, a brand, or a funnel. It had no app, no login, and no plan to be acquired. It was a good toy that its maker kept alive, and keeping it alive turned out to be the entire strategy. Two decades of quiet maintenance beat a dozen flashier launches that are now 404s, which is a result worth sitting with.

How to approach it today

A first visit rewards low ambition. Draw a hill of powder. Pour water on it and watch it slump. Set a corner on fire. Drop a block of C-4 into the middle and see what the pressure wave does to the rest. The point is not to win, because there is nothing to win. The point is the same one that keeps people watching waves or feeding coins into a pachinko machine: a system that responds, consistently and a little unpredictably, to a small input.

Patience changes what the toy becomes. Spend five minutes and it is a fidget. Spend an afternoon and it is a workshop, with half-built machines saved to the gallery and a growing sense of which elements can be trusted around which. The ceiling is high precisely because the floor is so low, and the climb is entirely self-directed. No level gates the next one, and no score tells a visitor they are doing it wrong.

For anyone who enjoys that register of calm, aimless web, Powder Game sits comfortably alongside the rest of the quietly single-minded internet. It pairs naturally with a toy like Sandpile, which chases the same meditative "pour a material and watch it settle" instinct from the Boring Websites network. Both ask for nothing and give back exactly as much attention as a visitor cares to spend.

Powder Game is a reminder that a website does not have to be big, loud, or new to be worth keeping. It has to work, and it has to do one strange thing well. Nearly twenty years on, a blank canvas at dan-ball.jp still does.

FAQ

Is Powder Game free to play?

Yes. Powder Game runs in a browser at dan-ball.jp with no account, no download, and no sign-up. A visitor opens the page and starts drawing.

Who made Powder Game?

It was created by the developer known as ha55ii, working under the DAN-BALL studio, with collaborators credited as tylor and otascube on the current version. The game has been online since 2007.

What can you actually do in it?

Place more than forty elements, from sand, water, and fire to gunpowder, metal, seeds, and a steerable stickman, then watch them interact under a simulation that tracks wind and air pressure. Finished scenes can be uploaded to a public gallery for others to run and rate.

How is Powder Game related to Sandspiel?

Sandspiel is a later falling-sand toy whose creator, Max Bittker, names Powder Game as his primary inspiration in his 2019 essay Making Sandspiel. Powder Game is effectively a direct ancestor of the modern falling-sand revival.