Game Review: Factorio: Space Age (pt 1)

Factorio is my favorite game of all time. I played it very early on, then periodically after that. When I started, the graphics were much uglier, there was no nuclear power, biters dropped little purple orbs you needed to use in science, ninety percent of the current QoL was missing, and it was still one of my favorite games.

Before the Space Age expansion, I had ~1200 hours in the game, partly because it was my go-to game when I was a stay-at-home dad and my son was napping beside me on the couch. I’ve played not only vanilla Factorio, but a lot of overhaul and other mods. These are the overhaul mods that I’ve finished:

• Bobs
• Bobs + Angels
• 128k
• Krastorio 2
• Space Exploration
• Exotic Industries
• Freight Forwarding

Additionally, I made it to the terraforming stage of Nullius and py science 2 of Pyanodon’s, but didn’t finish either of them. This is all for context, where I’m coming from in this review. I have no idea what it’s like for a new player, but my guess is that it feels complex as all hell.

The Space Age expansion expands the game by adding in 4.5 new planets (Vulcanus, Fulgora, Gleba, Aquilo, and space itself) as well as a major-but-optional mechanic, quality. I’m dividing up this review along those lines, which is the natural way to do it, but in theory all these things are meant to work in harmony with each other, so I’ll be trying to take that into consideration. Spoilers will follow in each section, but the Factoriopedia has everything right from the start, and the devs consider it a game that does not actually have spoilers, so take that as you will.

In my opinion, the real spoilers are the designs for things you build along the way, but there will also be some screenshots of those.

The Same Old Early Game

You start on Nauvis with a crashed ship, a pickaxe, and abundant mineral deposits. If you’re new to the game, red science and green science can easily take 20 hours to figure out, particularly if you’re playing with biters on. For me, it was about two hours to build designs that I have built maybe dozens of times before. The basic furnace stack that handles incoming iron, copper, and stone has not changed, and will not change.

If I consider the basic gameplay of Factorio to be the design and decision process, then there’s no gameplay here. Each entity needs to be placed by hand, and you can make rows of things by running up and down, but still … it felt like a slog to me, and this is the first ~4 hours of Space Age, assuming you’re going moderately fast and making a beeline to bots.

Once you have bots, it gets much less tedious, and you can start slapping down blueprints, expanding the base as rapidly as the machines can turn raw materials into finished buildings. There are a few differences from the base game, including terrain generation, some stuff with trains, science checkpointing … but it’ll all be well familiar to veterans, and in my opinion, is pretty skippable. I set up walls to keep the biters out, trains to supply the variety of turrets on the wall, solar and nuclear, and outposts for as much resources as I would need for the next few dozen hours, then made my first space platform and began the actual expansion stuff.

Space!

Space platforms are created by launching a starter pack up, which you can then send materials to. Bots aren’t allowed in space, and your character isn’t either, and it seems to me that a lot of the game design was built around wanting the player to grab resources from out of space and do some complicated belting to keep everything organized and prevent it from locking up. There are no chests allowed in space, and the only thing that acts as a container is the central hub of the platform, of which you can have only one. This means that if you want storage, you have to route everything through this big warehouse, and it gets complicated the more you have items going in and out.

I would say that generally I think this works from a gameplay perspective, but there are a few things that are needlessly obtuse or unfriendly, getting in the way of the platform design stuff that’s supposed to be the star of the show. One of them is definitely “automatically request materials for construction”, which will send up an entire stack of something you only need one of. This is an issue in the early game, assuming you didn’t overprepare on Nauvis to have a base with ~20 rockets per minute. Frontloading this difficulty, which becomes less serious later, is bad design, and you end up having to manually go through rocket loading to not waste enormous amounts of resources.

(The easiest way I’ve found to do this is to make a blueprint of the ship, click “add section” on logistics to make it a logistics group, set a requester chest to that logistics group, then unselect that logistics group once everything is there, then use an inserter to feed that stuff into a rocket and manually launch it every time it’s full, and even that sort of sucks, because the blueprint makes a logistics group that will have the hub and extra platform in it, and holy hell is none of this intuitive or friendly, why could they not just have coded it so that rockets would auto-combine things into groups?)

Going slightly out of sequence here, but I’ll talk about the space stuff all at once here. Over the course of normal play, I think the intent is that you design approximately five ships:

• A space science ship that sits in orbit, collecting materials from asteroids and doing bare minimum processing on them to turn them into space science, which gets sent back down to the labs. I made one very early on and then didn’t ever have much cause to touch it again, except to send up some better assemblers and slightly expand it with no major changes.
• An inner planets ship with chemical plants, engines, furnaces, and an ammo assembler that feeds turrets to shoot down asteroids, which the grabber arms then take chunks of for the materials to run the chemical plants and be made into ammo. (I dubbed this the Dart-class, pictured below is the SS Christopher Wren.)

• An Aquilo ship with rocket turrets to shoot down the larger asteroids that the normal turrets have problems with. This requires advanced asteroid processing to get sulfur and coal synthesis to make coal, which gets made into explosives to make rockets. Probably at the same time you’re switching over to advanced fuel processing with calcite. (I dubbed this the Jacknape-class, pictured below is the SS John Napier.)

• An outside the system ship with rail guns to shoot down the largest asteroids. This requires making rail gun ammo, which needs steel and copper wire, and to power all that you’re probably not going to use solar, which gets much worse out at the edge, so likely you’ll be doing nuclear or fusion. (I used a lightly modified Jacknape-class for this, though it would have been better to do a full redesign.)
• A shattered planet ship that is capable of harvesting promethium, which I have not actually made yet, but requires scaling up even more.

Overall, I found the increasing complexity of designs to be very pleasing, even if it sometimes felt a little bit forced. Not having bots I can maybe understand, but not having chests felt like a very blatant design decision rather than something that came about naturally from considering space and what it means, especially since the belts still work. Designing the SS John Napier was one of my favorite parts of the entirety of Space Age, partly because it was so constrained, and I knew that my individual decisions were creating individual problems of my own making.

I will say that space is where Factorio shows its limitations far more than elsewhere. In programming terms, Factorio uses something called a “surface”, and each planet is its own surface, as is each ship. Surfaces cannot interact with each other, and in the mods I’ve tried where they do (more than just hooking up inputs and outputs) it’s always been a bit jank. Still, this means that there are a lot of things that cannot be done:

• Docking one ship to another
• Having a ship land on a planet
• Having a ship have any verticality to it

The ships also look a little … well, bad. They look like a bunch of things have been placed on a flat slab, especially when they get larger. This can be helped a little bit by adding walls around the ship, but it doesn’t help much, and there’s no aerodynamic consideration, so the ideal design is probably a big box of some kind, and the space platform that everything is built on looks even less ship-like than everything else. The exception is the engines, which look awesome, but I don’t think having one element look really cool makes up for the rest looking a bit weird.

Funny enough, the Space Exploration mod actually does do some of the things that these ships don’t do, like docking, landing on a planet, etc. It was a bit jank there too, but it did kind of sort of work. And those ships needed to take aerodynamics into consideration, though I can’t remember what the formula was like, and it was pretty opaque.

I do not need to have the entirety of Kerbal Space Program inside of Factorio, but I do think there are a lot of things that are neat about space that they just decided not to touch. The planets are in static positions, always the same distance from each other, and there’s no need to worry about launch windows or delta-V or gravity slingshots or light-speed communication delays any of the other cool rocketry things. Some of that would be a nightmare to implement, other things would probably not be very fun, but it feels like there was a lot left on the floor.

It’s interesting that spaceships in this game are self-sufficient by nature, gathering materials from asteroids and never needing resupply. It’s also interesting that there are two basic modes for ships, in-flight and in-orbit, with different considerations for defense and production, though I don’t think they ended up doing all that much with this distinction. If spaceships could land on planets, you could have three distinctions, and if they could be flying through interstellar asteroid-less space you could have four, and I think that would be cool, but the focus of Space Age is mostly on the new planets, not on the spaceships.

It’s at this point that I’ve realized that this review is going to be very long, so I’m splitting it into parts. The four planets will be the next part, but before I wrap this up, I can talk about one of the other things that came with the expansion: quality.

What Quality is Quality?

I would say that of the 140 hours that Space Age took me, about 30 hours were spent messing around with the “quality” mechanic, and of those, most were “wasted” in the sense that they did not meaningfully make a better factory, even if I enjoyed the process.

Quality divides almost everything in the game into tiers, with higher tiers having better features, which depend on the specific building or product. Resource extractors do less resource drain. Production buildings get better crafting speed. Weapons get better range. Some things get faster and require more power for that speed, while others get speed without needing more power.

There are a few sticking points with quality.

One of them is that machines cannot use a quality product if they’re not set for a recipe that requires it, meaning that an “uncommon” gear cannot take the place of a common gear. I assume that this was either an engine limitation or a deliberate challenge for the players, but either way, I don’t like it. Quality does kind of make sense, since it’s something that exists within real world manufacturing, where parts need to be within certain tolerances, but it wouldn’t be the case that a gear that’s inside a narrower band couldn’t be used for purpose that’s in a wider band. Factorio is the wrong game to be making real world comparisons for, but the argument is that an uncommon gear shouldn’t be enough to gum up the works.

One of my plans for quality was to “skim” quality parts. The last machines in a stack of assemblers would be given quality modules, and of the thousands that they made, a few would be high enough quality that they could go into a chest, and that chest would be used for making personal equipment and spaceship parts, where they potentially make the most difference. At a certain point, I misconfigured one of these setups, and some quality gears got on the belt, which gummed up the entire factory and required me to clean several lines and restart a bunch of processing. This is a skill issue, yes, but it’s an unpleasant complication of quality generally.

Quality comes from quality modules, and in general, the modules are a matter of trade-offs, whether you want more speed, more efficiency, or to make the most of materials. Quality … well, quality is an enormous complication. You can’t simply put in machines. You need entirely new setups for it, and even skimming feels like kind of a weird and gross way of doing things.

Here’s how I wish it worked: You put quality modules into machines, and they can make quality things at a set chance. Those products can go down the line and be used in any recipe that requires lower or equal quality. Uncommon gears and chips would get consumed by machines that make normal quality engines or whatever. This would instantly solve at least half of my frustrations, but it would also be simpler, and not so much of a challenge.

How it works now is that you either silo away all qualities from each other, or you engage some kind of recyclotron that attempts a craft and instantly junks it if it’s not quality. This is one of my tileabale parameterized recyclotrons:

Blue chests request normal quality materials, machines make the base product, anything not at the desired quality gets recycled, materials go on the belt to be made into more of the desired thing. There are some circuit conditions set up, one to shut down the machines if the desired number of quality machines have been made, and another to set the inserters to only pull from the chest if there are no materials on the sushi belt.

I think this is interesting, but if this is all quality is, then the juice isn’t worth the squeeze.

Before building my final ship, I set up full quality on Fulgora, at a place isolated from the main base. It separated out every item at every tier, then used roboports to put things together. It was more interesting than the recyclotron, with better/faster/cheaper results, but still kind of meh, and I kept wondering why I was spending all this time trying to make a chemical plant that was twice as good when I could have built a second chemical plant for half the cost.

My other major gripe with quality is that it makes blueprinting a pain in the butt. First, because the speeds of machines are different, which throws off ratios, but second, because if I want my machines to be of the best quality available, there’s no way to easily do that. What I want is to have a tool where I drag across a bunch of machines and say “upgrade these in accordance with the highest quality in the logistics network”, but what I have to do instead is count the number of each type of machine, then manually go through and replace them, and if I do this, then I have to manually go upgrade machines as more become available, and this means that I can’t just copy sections of the factory to duplicate them, because they’ll be at a mishmash of quality on buildings. I spent a lot of time fiddling with the upgrade planner, which I didn’t enjoy.

The Fulgora setup, at endgame, is currently making the legendary quality modules necessary to make the legendary quality modules necessary to make legendary quality buildings of all kinds. I think pouring enormous resources into that makes for a megabase, but mixed quality faces lots of usability concerns, and I think of all the approaches (skimming, recyclotron, mass sorting) the recyclotron is the one that I’m most likely to end up actually using in future playthroughs.

Which is to say that I think quality as a mechanic is one or two steps away from being good, as much as the rewards do often feel worthwhile. The puzzle of quality has not, for me, been a highlight.

In the next part of the review: the four planets.