I’m an analyst by training, with a somewhat disturbing fascination with numbers. Not so much in the discrete numbers but the story that numbers can tell. There are 3 main phases – the pure numbers, the sanitation of the numbers, and then the trend of the numbers. Reddit has a neat DataIsBeautiful section that presents the mid-point of analytics. This is the flex point where data goes from being objective to subjective. You can make numbers tell any kind of story… CNN and FoxNews use the same data sets to spew completely different opinions after all. That aside, making the numbers do something is the foundation of factory games.

Factory games are a sort of inverted view, where the goal is set up and you are given some basic numbers to sort out. Like if I said, collect 100 lumber. You would go out and cut some trees to get there. Which trees, and where is up to you. Maybe you want big trees with a lot of lumber, or tons of smaller ones that are faster. Maybe you skip doing it and instead build a sawmill to speed up the process and save time later. The only thing that matters though, is getting 100 lumber.

At the start, factory games are simple affairs. Click a few buttons, collect things, make things, next. They quickly spiral to more complex things, were you need multiple steps working together, in order to meet an order. If I moved from 100 lumber and said make me 20 tables, well you’d need a saw, nails, and lacquer. Then I’d say, make me 50 houses. Then a town. Then a country. And so on.

The gains at the lumber space are linear. You will collect 100 lumber in 5 minutes of a game. Building a table though, each one may take a minute, so you’re 20mins to complete that task. And as you move farther along, more and more time to get there. Time to create is a limiting factor, but only in the linear sense. If you make more factories to make tables (say 20 of them), then it takes only a minute to make all of them. Of course, you now need 20x the speed of collecting the base material to feed the factories. And then the logistics of moving the material from the collectors to the factories.

But then you need to worry about space. You can’t just pop up a factory anywhere. Building them too far apart means you need to ship material all over the place. You will end up with “hubs” of factories that are closely linked and can share the supply chain. They need to be close enough to the base material and the transport method needs to support large volumes so that you always have material on hand to make things. Nothing worse than a factory sitting idle because it has no material to craft. And yet, you need real estate to build this stuff. There’s only so much land available to build upon, so things need to be tightly built next to each other to optimal.

Finally, you need to power all of this stuff. Sawmills might work with water, if they are near it. But if they are out in a field, then you need to get power to the mill. Same with the factory. Making power means finding enough sources of energy, and the mechanisms to process them, in order to over-supply your needs. You want more, so that you have flexibility to add more factories quickly. Not enough power means that your things run slower (if at all).

Frostpunk

Frostpunk is a really interesting game because it’s about failure avoidance. Most builders, if you step away, things just don’t progress. Frostpunk instead escalates bad events, either sickness, hunger, power issues, or ever increasing levels of cold. You rarely have enough resources to do anything, and need to send out expeditions to collect more materials. They take time to get there, time to get back, and there’s a chance they will die. You need to always be pushing the “safety blanket” level to best absorb the ever shifting disasters.

You’re limited in base resources, but also in power and location. Buildings fit into slots of sort, and you have a maximum in any area. Then you need to get power there, and importantly, heat the buildings where people are located (night buildings without worked don’t need heat, robots never need it).

It isn’t a factory game though, because it has multiple levels of RNG. Maybe you’ll get a giant 3 day storm, or maybe you find a huge hoard of coal. Your best plans can be completely trashed because of it. Factory games are about predictability. Frostpunk instead does a superb job explaining the limitations of a factory game.

Math Exercise

Factory games often benefit greatly from spreadsheets. Understanding the rates of items is key to optimization, so that you are as close as possible to the place where you are using 100% of capacity, without causing backlogs. If the table factory consumes 10 wood per minute, and you have 20 factories, then you need to be collecting at least 200 wood per minute. 150 per minute and then factories will stall. 250 per minute and you create an overstock of lumber that needs to be stored somewhere.

That’s a simple example. Multiple tiers of complexity later, you’re going to be consuming multiple products that are created at different rates with different material. Maybe factory 1 needs 100 wood per minute, and factory 2 needs 20… and factory 15 needs 75. Figuring out the base demands means you need to find more material and harvest more things. I’ll have a good example of this in a future post.

The spreadsheet helps to keep track of all the requirements, or the quick math needed to see how things work with each other. In particular if you’re looking to optimize, as even the smallest of numbers early on can have cascade effects at the tail end. The number of times I’ve had bottlenecks and tried to figure out why was often resolved through a look through the numbers.

It’s really hard to make a factory game, hence why there are so few. Factorio is the platinum standard. It spent 4 years in Early Access. Cripes, its hard to make any game, but the people who plays these games are looking at every detail. They are nitpicky by nature! So when something like this does come around, and it works, holy cow does it feel good.