It seems almost inappropriate to call Zachtronics’ Opus Magnum a puzzle game. I tend to think of it as a problem-solving game, or an engineering game. Its ethos—gentle, forgiving, gradualist, concerned with incremental improvement rather than strokes of genius—is at odds with most of the puzzle games I’ve played.
It came to my attention first through the stunning looping gifs it produces: flat little worlds of constantly moving machinery, reminiscent of the atomic-level manipulation of SpaceChem. I don’t normally play a lot of puzzle games. I’m fond of tiny PuzzleScript experiences, and I’ll play puzzle platformers like Fez and first-person games like The Witness, but I’m not an enthusiast of the genre. I suspect I’m not alone in loving Opus Magnum in spite of my cooler reaction to puzzles in general, though.
It achieves something beyond what a game like The Witness tries to achieve. Something, I’d argue, more fertile and appealing than the maze-filling in that game, or the convolutions of Fez. Opus Magnum is less about breaking through frustration and more about steadily building up to something—less about getting into the designer’s head and more about expressing your own way of thinking.
How to turn lead into gold
Every level in Opus Magnum is a simple problem of alchemy. You have inputs ("atoms" or "molecules" of fundamental elements such as earth, salt, or gold), and expected outputs. Your goal is to assemble mechanisms and program their actions so that they transmute inputs into outputs.
As the game’s story progresses, the player is tasked with manufacturing everything from fuel for alchemical airships to poisonous lipstick. It’s clear at every point that, while there’s a difficulty curve, the puzzles are very much designed around the story—they stem from narrative ideas, rather than purely mechanical exploration of the design space.
Finishing a level isn’t a matter of finding a solution but triangulating your way towards one.
Most puzzle games can’t do that, because they need to very carefully and deliberately ramp up their difficulty. They have to make use of every nook and cranny of possibility in a mechanic before moving on to the next one. Opus Magnum is broad and open-ended in a way that lends itself to this kind of structure, though. It doesn’t have to wring every bit of value out of its mechanics.
Like a lot of puzzles, they break down into smaller problems. The difference is that those parts are completely independent—once you’ve solved a subset of the problem, you know that solution works and it just has to be joined to the rest. Finishing a level isn’t a matter of finding a solution but triangulating your way towards one. It’s climbing a mountain, not leaping across a gorge.
A lot of puzzle games let the player make mistakes that only become apparent later, sometimes requiring the entire solution to be scrapped. Even the solitaire puzzle that accompanies Opus Magnum, Sigmar’s Garden, is like this, as are countless other puzzle games. Different parts of the solution interfere with one another, gradually constraining how the puzzle can be solved so that only one solution is valid at the end—think of how every square in Sudoku has to exist in agreement with every other square.
A tiny but incredibly effective example of this interdependence is PrograMaze, a mindbending PuzzleScript game. PuzzleScript is designed for tile-based puzzles and Sokoban variants, and PrograMaze presents the simplest possible problem in this space: move a blue tile to the orange goal. The difficulty comes from how you control the tile by writing a simple program, and the bits of memory that your program lives in are also the space that the tile moves through.
Essentially, you have to build a maze that also expresses a path to solving itself. It’s a game entirely about the player getting in their own way, which to me distills the quality puzzle games sometimes have that gets them called fiendish, twisted, or cruel. There’s a meanness to making the player their own villain—it's hard to execute well, and not for everyone.
Opus Magnum avoids this completely. It doesn’t want you to get in your own way, and it doesn’t want its problems to feel like mean-spirited tricks. It never hides something in its levels meant to elude your attention until you find it halfway through, invalidating your tentative solution.
Opus Magnum never demands an efficient solution
Because solutions are assembled from parts that can be built independently and compartmentalized from one another, you can start from either end of the problem, or you can complete different sections independently and integrate them at the end. The more complex a level, the more angles of attack you have to start solving it. The most daunting levels in Opus Magnum act as invitations—you might not be able to see the whole solution at once, but you can see how you could do a part of it. And once that subset is done, what remains is less complex.
A friendlier brainteaser
Most of all, this is a forgiving game. You always have unlimited space and resources to assemble your solution, limited only by your inputs. It ensures that mechanisms stay in sync as they cycle through the solution you’ve assembled. Letting an atom linger in place after being transmuted won’t cause it to turn into something else entirely. Opus Magnum never demands an efficient solution, never demands you prove you’re good enough before letting you advance.
The designers clearly realize most of the joy in Opus Magnum comes from optimization, and so they’ve built a game in which finding an initial, crude solution is easy. Taking a crude solution and making it faster, more efficient, more compact, and more elegant, is a pleasure absent from most puzzle games. There’s a comfort in knowing at that point that you can’t fail. You’ve already found your answer, you’re just seeing if you can make it better.
Opus Magnum explicitly encourages you to optimize size, cost, and speed with its histograms. These histograms are so much smarter than a leaderboard. Instead of asking "Can you be number one?" a fundamentally boring question to which the answer is almost always no, they ask, "Can you make it to the 70th percentile? The 90th?"
They all but ensure players will have something to feel good about with every solution. Invariably, you’ll be operating either faster, cheaper, or smaller than most other players, simply by virtue of how those three properties are at odds with each other.
Making your machine faster entails adding more components to it, which increases its size and cost. Making it smaller and cheaper means making it slower. And often the very cheapest machines need extra room to work, particularly as the levels get more complex. Every solution has to fall somewhere within this triangle. Perfectly optimal solutions that hit all three targets don’t exist for any but the most trivial problems, making optimization a matter of personal style. Some players will tend towards "tall" solutions with numerous devices working in unison at fast speeds, other might go for "deep" solutions in which small numbers of devices perform long, complex programs.
If a good puzzle feels like a duel between designer and player, Opus Magnum is much more like a conversation.
And there’s a hidden, third optimization that can’t be quantified in a histogram but is also very much encouraged by the game: aesthetics. All Opus Magnum solutions are visually satisfying, as illustrated by the mesmerizing looping gifs they produce. But some solutions are more elegant, more beautiful than others. Opus Magnum revels in its machine dance. Building a good-looking construct is its own reward.
Zachtronics games have always stood slightly apart in this way, but Opus Magnum takes this philosophy of design and runs with it. There are no tight constraints like the limited board and code space of Shenzhen I/O, no room for messing yourself up like the finicky integration of SpaceChem. Perfection is unattainable but you have the room to build whatever you want. Opus Magnum is a puzzle game that you play rather than stare at in frustration.
Many puzzle games have built-in puzzle or level editors, but Opus Magnum is the only one I’ve played where you can use that functionality to entertain yourself. It’s problems aren’t created by working backwards from a solution—you can pose a question you don’t already know the answer to. Rather than the tightly prescriptive jumps of Super Meat Boy, it’s the open-ended parkour of Mirror’s Edge, where vast spaces of play open up from almost any starting condition and true surprise is possible.
While a good puzzle feels like a duel between designer and player, Opus Magnum is much more like a conversation. It’s posing a question not because it wants to see whether you’ll get the answer right, but because it’s genuinely interested in what you have to say.
