SpaceChem

This is the seventh engineering puzzle game I've reviewed today, and the one all the others were descended from. It's also, by Steam's numbers, the least liked of them.

Being first is expensive.

What it invented

You control two "waldos" — pincers that follow looping instruction paths you draw on a grid. They pick up atoms, bond them, rotate them, and drop the results into outputs. The two waldos run simultaneously, which means the real problem isn't chemistry at all: it's concurrency. Two things moving on independent loops that must never occupy the same square.

That's a genuinely brilliant central idea, and it's why originality scores a 9.5. Every game in the table above is a variation on the question SpaceChem asked first: given a constrained machine and a required output, what's the most elegant thing you can build — and then, how much smaller can you make it?

The histograms were here too, plotting your cycles and symbols against everyone else's.

Why it rates lowest

Not because it's bad. Reading all hundred of its most-helpful negatives, two things stand out.

The on-ramp is the weakest in the set. Tutorial appears in 16 of a hundred negatives — the highest rate of any of the seven — with explanation and learning adding another 13. The studio hadn't yet worked out how to teach its own ideas, and later games handled it better; Opus Magnum in particular teaches almost wordlessly.

The criticism is mostly self-selection, not complaint. The most-upvoted negative, at 134 votes, is gracious: "Due to its extreme complexity, SpaceChem will never be a title for everyone. Too deep and articulate for casual players… Clearly, it was not made for me." The second, at 84 votes, opens "It's not a bad game, but you should think carefully about whether or not you are the target audience."

That's the same shape I found in SHENZHEN I/O, whose 905-vote top negative also recommended the game. In this genre, a low score frequently means "wrong audience" rather than "poor game" — and across seven titles today, that pattern has held remarkably well.

A fair criticism that is specific: it looks its age. Visuals score a 6.5, the lowest of the set. And Linux or macOS problems appear in 7 negatives spread from 2013 to 2021, so cross-platform support has been imperfect for most of its life.

Reading the dates

Negatives cluster in 2014–2015 — 37 of the hundred — but that group carries a median of 3 votes, and the 2022-onward group just 2. Neither cluster is weighted enough to represent an event; this is ordinary spread across fifteen years rather than a moment where something went wrong.

Recent English reviews sit at 94 of 100 positive, with a median playtime of 15.6 hours. The critics' median is 2.7 — the familiar fast-bounce pattern of the whole genre.

Achievements are punishing but not extreme: 20 in total, the most common held by 35.9% of players and the median by 8%.

The verdict

SpaceChem earns an 8, level with TIS-100 and SHENZHEN I/O.

It's a foundational game — the concurrency puzzle that an entire genre grew out of — and at NZ$14.75 it remains one of the cheapest ways to find out whether this kind of thinking suits you. Fifteen years later the core still works.

It doesn't score higher because the studio's later work simply teaches better, looks better and respects your time more, and because a 2011 interface is a 2011 interface. That's not a criticism so much as an observation about progress: SpaceChem's biggest problem is the six games it made possible.

Play Opus Magnum first if you just want the best one. Play this if you want to see where it came from — and it's worth seeing.