Science, Fiction, Life

What is Zerg creep, really?

Artwork from StarCraft, showing a creep-infested platform near the planet Char.

“What the hell is that? Looks like the ground there is alive.” – Jim Raynor

Creep: that purple, fibrous, living mat that extends from zerg “buildings”. Its ominous presence always tells you you are entering unsafe territory, unless of course you’re playing as zerg, in which case it says “welcome home”! But what exactly is it?

Well, let’s think about it scientifically. It is produced by zerg buildings and spreads across any available surface. According to the Starcraft Wiki (which I will use as an authoritative source for all the minutiae of StarCraft trivia that I don’t know) Creep has a cellular structure: it’s not just mucus. The Wiki also says that creep can absorb sustenance from the underlying terrain, that it can be spread by “spores” or excreted by several units, and it provides nutrients to zerg buildings. It is averse to high temperatures but can grow in space and over water.

So, does any real-world living thing match this description? Surely not, right? Wrong. The closest analog that I know of are slime molds. You’ve probably seen slime mold before without knowing it. They are a type of fungus, often brightly colored and found growing on damp logs in the forest. They are also incredibly weird.

Some types of slime mold are made of multiple cells joined together to form a super-cell with a shared cytoplasm. Ok that’s weird enough, but the other kind of slime mold starts off as a bunch of separate single-celled organisms, which can then coalesce into a multi-celled organism.

In terms of similarities with creep, slime molds are spread via spores but can also grow and multiply when they encounter nutrients. Slime molds can become quite large, and form branching networks of cytoplasm, allowing the leading edge of the slime mold to stream nutrients back to the rest of the “organism”.

A slime mold branching out and looking for food.

When the going gets tough for a slime mold and nutrients run out, it can transform and form structures called sporangia, which distribute spores. In some cases, separate cells will coalesce into a single “creature” in order to do this.

There are some definite similarities between slime molds and creep. They both come from spores but can also grow across nutrient-bearing ground. They both can transport nutrients to locations within them that need them, and are averse to hot, dry conditions.

There are some aspects of slime molds that would have been very interesting if they applied to creep. Most notably, slime molds have been reported to show some rudimentary intelligence. No, they don’t sit there and ponder the meaning of life, but they have been able to choose the most nutritious food and can “solve” mazes to get to food sources. These aren’t true intelligence, they are actually an example of something called ant-colony optimization, often used in computer programming.

The slime mold starts out evenly spread through the maze, but when food sources are placed at two ends, the slime mold retracts from everywhere but the shortest path.

The idea is that you don’t know what the best set of steps to reach a certain goal is, so you test things out randomly. Some sets of steps don’t give you the goal, but others do. The ones that do give good results are reinforced, while the ones that don’t, are not. The analogy is that ants start off randomly searching for food, but when they find food, they emit pheromones encouraging other ants to follow the same path, so eventually you end up with the familiar narrow stream of ants going from the nest to the food and back. The exact same principle applies to slime molds.

That’s nice, but weren’t we talking about creep? Yes! My point is that creep could behave like this too. There’s not necessarily a need for it to be spread evenly across the ground. It would make more sense for it to have thick branches connecting zerg structures (so that large amounts of nutrients could be provided) while narrower branches near the leading edge of the creep could do the work of absorbing nutrients.

Just to play devil’s advocate though, I can see why it might spread across the ground evenly (other than because it makes the game more intuitive to be able to see a clear boundary to the creep). If it is able to suck nutrients out of any surface, then it wouldn’t have to concentrate on certain areas. And by not coalescing into thick “veins”, the creep is more robust: there’s less chance of a building being cut off if there are many smaller veins feeding it.

Finally, all of this brings me to an interesting point: if creep can extract it’s own nutrients, and if it is the way that zerg structures are fed, why do the zerg have to mine for minerals? They should just be able to engulf a mine in creep and let it do the work! That would certainly make for a different zerg strategy, especially if a “creeped” mine could not be used by other players!

Obviously the creep is still pretty science fictional. I mean, it can grow in space! There are actually some real-world spores that can survive in space, but I think the whole “zerg don’t need spacesuits” issue needs to be tackled in a future post. But from now on, when you see creep, think “slime mold” and when you’re out in the woods and you see a slime mold, be glad you don’t have to watch out for zerglings!


  1. Kevin

    I’d like to link to Ed Yong’s post on using slime molds to model rail systems, but my internet is being quite poor. Maybe later… As to why creep isn’t vein-like: It could be due to the fact that zerg units move faster on creep and the colony as a whole benefits from this more than it would from economical creep spreading.

    Possible future post: the issues that mech-type armored vehicles would have a la the viking, thor, strider, colossus, immortal, and formerly the goliath.

  2. Pon

    Zerg units only “morph” they don’t “grow”. That means the only time their body really undergo structural growth is when they’re inside a cocoon (albeit from larvae into units, from drones into buildings, or from zerglings into banelings!). Every units you see must already be “adults” and therefore the “food” they eat would only be required for energy and not structural growth.

    This could explain why they only need minerals when they’re building or morphing (to use the blue glowing crystals to build those hardened carapace and armored piercing spines) but otherwise when buildings are feeding from creeps they only requires minimal amount of energy to sustain what’s already there.

    Also, creeps are generated by building which cost minerals (in starcraft2 queens can force creep tumor out of their bowels for free, but it only extends the range of already extending creeps), so it’s highly likely that creeps also function as means of how minerals are distributed throughout the colony. Don’t you require creep to build most buildings anyway?

  3. Kevin

    The only building that you don’t need to place on creep besides a hatchery is the output of a Nydus worm, which is a change from starcraft 1. Here’s the link about slime molds acting as a good model for finding efficient paths around relatively natural obstacles.

  4. Ryan

    Nydus canal outputs don’t have to be placed on Creep? Well that’s awesome!

    • Ryan

      Testing threaded comments.

      • Ryan


  5. Steve

    It reminds me of the red weed in War of the Worlds.

  6. J

    Interesting. could we then think of the connection to the zerg’s “living” buildings as something similar to mycorrhizae symbiotic relationships with plants?

  7. Josh

    Great blog! Hope that this could be expanded more. Its really interesting, particularly that link you provided in your most recent post. Anyway, Zerg creep may be a matrix, something like a living support. It probably contains the channels for nutrients and other essential substances to be delivered to Zerg buildings, which are essentially living organisms as well.

  8. JtiksPies

    A great post. Learning about the zerg biology, I never really paid much attention to creep. I’ve always thought about how a larvae can turn into any 8 subspecies on demand, and some of thsoe into others. The best reason I’ve found would be a “quadruple helix” where instead of our double helixes, their’s can unattach, move up or down and reattach to create different sequences. Or they could also unattach and switch to a adjoining strand and attach there. A pair of double helixes side by side constantly switching whic side they are no for more complex structure. Which brings up the problem where zerg DNA (maybe RNA?) is supposably very simple (ie. the zerglings are so simple they are always born twins) The best counter would be that this pair of helixes also happens to be very short, therefore simpleier than a normal double helix of a human.

  9. Freddie

    I considered the entire zerg colony as a single living organism, a living Gaia. By that, we could assume that the zerg is controlled by her basic homeostatic processes. The DNA of the zerg could be quite long, possibly coding for every possible organism (from sling to ultralisks). Larve spawned by Queens produce basic cells which are pluripotent, similar to stem cells, that will differentiate into specialised cells with specific functions. These differentiations may be induced by homeostatic requirement of the zerg colony.
    The zerg buildings can be seen as parts of the brain or secretory proteins, while the units involved are the muscular motors.
    For example, in the human body, during a fight, the body induces the hypothalamus to secrete adrenaline. Similarly, when the zerg organism encounters mass colossi, evolutionary feedback will develop the organ called ‘Spire’ which produces factors that allow differentiation of Larvae StemCells into Corruptors/Mutas

Leave a Reply

Your email address will not be published. Required fields are marked *

© 2020 Ryan Anderson

Theme by Anders NorenUp ↑