A critique of Factorio’s power generation systems

[Jakiepoo]

Factorio is a game about building increasingly advanced industry. Progression is at the heart of the game – is the game’s ultimate aim. To reach technology advanced enough to build a rocket and fly to space. As such, throughout the game, there are several technologies that help you to advance the technology you use so that you can make better use of the resources you have – time, power, and literal physical resources.

A fantastic example of this comes in the form of the various oil processing technologies. The starting technology (basic oil processing) is enough to start your production of plastics and other oil products. However, the higher yields of products as well as the more advanced materials (such as lubricant) are locked behind tackling a more advanced system. The advanced oil process is literally more complex – requiring the player to deal with three products rather than one, potentially requiring additional processes like cracking to balance production with need.

Now consider the difference in setup cost between the basic and advanced oil processes. The difference boils down to a pump, possibly a few chemical plants and a bunch of additional pipes (which are essentially free in the grand scheme of things). The “price” of the advanced process is not primarily or solely in material resources, but rather in the player being clever enough to implement it. A cracking operation is not difficult to set up, but it requires some critical thought and implementation of the circuit network (a system which, although fundamentally simple, is not massively intuitive for a new player). And whilst there is a literal difference in price, possibly even a large one in the case of some technologies, that difference is necessitated to make the progression feel balanced and natural.

By thinking about the problem and implementing the more complex but more efficient system, the player can increase their factory’s productivity (which is perhaps what truly separates advanced and basic technologies). This is a big part of what makes Factorio gratifying to play. The hook of the game is in inviting the player to think about their problems and solve them in creative ways to get the best result. This idea, demonstrated with oil processing, is true of the rest of the game too.
Implementing a smelting setup or a crafting line can be as simple as one might desire. But there are optimisations that can be done in order to improve efficiency in terms of resource usage and time efficiency. You can make your factory objectively better if one is willing to put in the time and energy (and maths) required and doing so is innately satisfying.



And this brings me to my criticism of Factorio’s power generation system. In general, Factorio’s power generation isn’t as complex as the rest of the game. Going into this, I want the reader to keep a question in mind. The question I put to you is this: what is a good price for a more advanced technology? What should the player have to pay in order to get to the more productive methods? Bearing in mind that a “good price” is one that makes the game “good” (satisfying, engaging, balanced, fair, enjoyable, etc.) to play. As designers, surely, this is our aim.

For the most part, the power generation systems (all three of them) are very basic. Take the steam boiler to start. The boiler is the most basic technology, and there isn’t really much of an optimisation to be done on a basic boiler setup. You could maybe change the fuel over to solid fuel and add a burner inserter rather than an electric one to ensure the system never goes down, but neither of these is very difficult. There is also perhaps some fun to be had designing a circuit network system to automatically route fuel to where it is needed, but ultimately, most people never feel the need to do so because saturation (as in, providing the power unit with so much fuel that the belts are constantly full) is such an effective method that more advanced systems are pointless. Unless you plan on fuelling your boiler with rocket fuel (a prospect so ridiculous that I am sure to try it next time I play) there isn’t much need for anything more complicate than a belt and an inserter.

And this is fine – the boiler is meant to be basic. but consider similarly tiered systems; basic belts and assembly machines are used throughout the game for low priority or niche purposes, and setting them up requires an understanding of mechanics like splitters, inserters, and (depending on the purpose) potentially the circuit network. Similarly, the boiler will probably last you a good portion of the game, at least until you have access to solar panels, and quite likely all the way up until nuclear power, since solar is so relatively low in power output. Despite this longevity, there are no real later-game optimisations that can be made for the boiler.

Solar panels are no better – indeed, they might even be worse. Solar panels themselves need little more than to be placed down with an appropriate number of accumulators. Working out this exact amount ultimately comes down to understanding how your factory uses power – which can be found literally by looking at a graph of your factory’s power usage. Again, however, this is fine. Solar panels exist for the purpose of generating power either without producing any pollution or in remote situations where connecting to the main plant is impractical or not desired.
This brings us then to the most advanced power generator in the game. Nuclear reactors are without any doubt extremely advanced technology. The author is too lazy to look up exactly how much more energy a reactor system produces compared to a boiler system but knows from experience that even a huge boiler system can be compressed down into only two or three reactors. Reactors are indeed, “big bois”.

Given this, you may think that the cost of a reactor, both material and in design, would be high. And whilst it is true that reactors and centrifuges are expensive, reactors are not particularly difficult to design, besides perhaps fuel reprocessing, steam management and shutoff to prevent energy waste. I do not mean to say that there are no challenges in designing an efficient and good reactor – the problems above, especially the last one, present a real challenge for the player to solve. For me, the reason the reactor design seems so egregious is because the setup cost seems to be less than the crafting cost. I have spent far more time crafting reactors and centrifuges than I have spent in setting them up in an intelligent manner. This is, of course, simply the author's personal opinion, and in general, the nuclear reactor is more complex than alternative solutions, and therefore, a more satisfying solution to implement. If I had to choose only one solution to redesign with more complexity, it would probably be the boiler, with the solar panels being a fair second. The point I make is that reactors are a very expensive thing to build for something that requires so little setup.
I also believe that there isn’t necessarily a huge change required to make boilers more interesting. Simply allowing for beacon interactions would immediately make boiler setups much more interesting to design, as they more or less fulfil all the specific requirements that make a beacon useful (densely packed machines, constant operation). It would also be interesting to see how one might precariously balance the beacons’ own power usage with the net gain granted by boosting the generators. Indeed, beacons are a very complex (read- “good”) mechanic that allows for a huge amount of variation for a player interested in making impactful marginal gains.
There are obviously also other ways boilers could be made more interesting – perhaps allow the player to manage the amount of pollution produced with additional logistics. And the same is probably true of solar panels, although I have difficulty thinking of any definitive solutions myself.
As far as reactors go, one possible solution to making them far too complicated (in a good way) would be no longer requiring all three heat pipe connectors for the adjacency bonus. For instance, perhaps if there is only one connection, then the bonus is 1/3 that of what it would be with all three, etc. in this case, the formerly simple question of “what is the most efficient reactor layout?” becomes much more complex. As it is now, a double row of reactors is naturally more efficient than any other since others are either clearly less efficient (allowing fewer connections per reactor) or impractical due to inaccessibility.

To wrap up, I want to bring us back to the question I asked at the start of the criticism. What is a good price for a more advanced technology? Hopefully, I have demonstrated that requiring the player to think about how to implement an advanced solution makes it a more satisfying system to play with. And I also hope that I have demonstrated that to an extent, there is still room for improvement in this department within Factorio’s power generation mechanics. This concludes my essay.


TL;DR – boilers are very simple, especially when compared to other technologies around similar tiers and use cases, and some additional functionality could make them more interesting to play with, especially considering that they are the power generation system used by people for probably most of the game.

[foamy]

Actually getting the full throughput out of a large scale reactor is significantly more challenging than given credit for. There's some nasty hidden throughput issues with heat pipes and water/steam flow that require careful engineering to avoid.

(Also: I *do* use rocket fuel in my boilers. It's a handy way to vent petgas when I need lots of lubricant.)

[Krazykrl]

Factorio is a game about logistics and automation; therefore whatever item you are attempting to automate isn't about the end, it's about the complex logistics that goes into producing the endpoint of said item.

Yes, the systems (when built); are simplistic. But this is by game design. Each individual entity only does one thing, convert stuff to other stuff.

If you take a look at anyone's factory, the endpoints for products are just a few in quantity. Since the vast majority of the factory is just the logistics and production entities to support the vast quantities of intermediate items required.

---

Power production is intended to be simplistic as a game design decision.

Tedium increases as you increase complexity; and this would compound into an exponential growing ratio of Tedium/Complexity. Past the first few minor bumps in complexity, you're adding tedium for no benefit.

---

Factorio is the mess of the history of the story of the journey of that chunk of iron ore, not its ultimate destination.

[netmand]

That's not a critique. It's a brain dump!

[Ranakastrasz]

I remember being disappointed with the reactor system when it was added. Mainly because it had hints of the Industrialcraft reactors, which were very complex and detailed, with a number of tradeoffs. The reactors we got were much less interesting. But, they worked, and could replace massive solar fields, so there is that.

[5thHorseman]

I have the opposite thought amusingly.

I think there should be one more tier of power generation, fusion reactors. Super expensive to build, runs on water, and spits out constant power like solar panels with no down side or ups hit outside that single pump.

[JimBarracus]

In hindsight its not very complicated.
Nuclear reactors should explode when they reach 1000°C.

[azesmbog]

With large electricity consumption, there is simply no alternative to nuclear power.
For me, the biggest design problem is the water supply. No more problems. But without this it would not be interesting to play.

[astroshak]

The reason nuclear seems so much easier to get up and running is how easy it is to find good, working blueprints online to just plop down.

If you try to design one yourself, its not that hard (its not really hard to design anything in Factorio until you start delving deeper into the circuit network than “turn off if X < Y” type of logic) but it is one of the more interesting things, because of the changing and odd numbers. 40/80/120/160/200(!!) MW power generation per reactor depending on adjacency bonus? 5.82 MW per Turbine? 103 water/sec Heat Exchangers? 1200 water/second Offshore Pump? Kovarex produces how many Uranium per second, for how many fuel cells, so one machine can fuel how many reactors?

Unless you are going for nuclear weaponry, such as the green ammo for some high power Turrets, the only reason to even bother with Uranium is nuclear power. As such, refining the Uranium, Kovarex, and the nuclear power plant are all considered, while somewhat piecemeal, also somewhat as a whole. This makes nuclear power potentially a bigger puzzle than anything else in the game; you’ve already solved making radar, lower density structures, etc. for the rocket by the time you need to start feeding them into the silo, so the puzzle there is simply one of routing resources already generated rather than starting on a whole new resource entirely.

I’d say that nuclear power itself is fine. They just need to do something with fluid mechanics to reduce the UPS hit and to make them a bit more intuitive (since most people do not bother to consider gravity based flows when we have pumps to pressurize lines and push fluid through).

[wosthisdo]

There are lots of mods that play with power generation (most overpowered), but I'm hoping that Factorio 1.1 might update the power distribution system.
It seems silly to me to have GW of power going down a single thin copper string on a wood pole. If the power lines had a range of max throughputs it would give the incentive to upgrade to medium & big poles, put in step-down transformers, buffer with capacitors or pump storage etc. as power generators were expanded or uprated.
I expect it would be a challenge to code efficiently enough to keep ups up, but it would add a new dimension to planning and upgrading. We've all had brown-outs, but having to go around and replace all the copper wire that had vapourised under the load from a laser defence would concentrate the mind wonderfully

[Krazykrl]

There are lots of mods that play with power generation (most overpowered), but I'm hoping that Factorio 1.1 might update the power distribution system.
It seems silly to me to have GW of power going down a single thin copper string on a wood pole. If the power lines had a range of max throughputs it would give the incentive to upgrade to medium & big poles, put in step-down transformers, buffer with capacitors or pump storage etc. as power generators were expanded or uprated.
I expect it would be a challenge to code efficiently enough to keep ups up, but it would add a new dimension to planning and upgrading. We've all had brown-outs, but having to go around and replace all the copper wire that had vapourised under the load from a laser defence would concentrate the mind wonderfully

I dunno how fun power distribution would be, it sounds like a mess of pipe-hell; except these are 5+ connection power poles with single-pixel connectors. Along with the fact that any destruction of the power network would be automatically repaired via conbots, since blueprinting is a major game feature that makes endgame and post-endgame bearable once you get past any scale.

And there is the other issue with capacity:

The other fact is that the real-life capacity of high-voltage power lines is utterly ridiculous. "For example, a 100 mi (160 km) span at 765 kV carrying 1000 MW of power can have losses of 1.1% to 0.5%. A 345 kV line carrying the same load across the same distance has losses of 4.2%." Currently, single runs of high-voltage lines can exceed a couple gigawatts of rated power (and Factorio (ostensibly) takes place in the future, where room-temperature superconductors are probably viable.)

---

Factorio does really well with brownouts/slowing factory production rates. Something I HATE in games, is if you get a momentary single-tick spike; stuff explodes/needs manual intervention. In a game about automation, unless something runs out; or runs into you... your factory should hum along in the backround.

[wosthisdo]

I guess I just like the idea of having to plan and upgrade a proper power grid in the transition from startup - mid - end - megabase. The trade-off between robust/redundant grids and their logic vs brittle but cheap appeals to me. Maybe as a compromise the poles could take steady damage if their wires are overpowered so that bots could handle the maintenance most of the time, maybe with power switches wired as circuit breakers to protect from transients. I still think a major overpower event should vanish some of the wires (maybe the >50% damage ones).

[wobbycarly]

I don't think any further complication of the power network (as opposed to power generation options) adds any FUN to the game, and first and foremost games should be FUN.

[DarkShadow44]

There are lots of mods that play with power generation (most overpowered), but I'm hoping that Factorio 1.1 might update the power distribution system.
It seems silly to me to have GW of power going down a single thin copper string on a wood pole. If the power lines had a range of max throughputs it would give the incentive to upgrade to medium & big poles, put in step-down transformers, buffer with capacitors or pump storage etc. as power generators were expanded or uprated.

I don't think that will become part of the base game. It can be modded though... Do you mean something like https://mods.factorio.com/mods/JeroenDStout/highvoltage ? I planned to bring that up to 0.18 at some point...

[eradicator]

There are lots of mods that play with power generation (most overpowered), but I'm hoping that Factorio 1.1 might update the power distribution system.

Never gonna happen. Electrical system is heavily ups-optimized. And such a change would turn every factory currently in existance into garbage, forcing players to do major reworks for no fun gain at all. Minecraft on the other hand has a bunch of mods with different tiers, and a bunch of more mods with completely unique power systems. If you haven't tried i very much recommend you do.

[wosthisdo]

There are lots of mods that play with power generation (most overpowered), but I'm hoping that Factorio 1.1 might update the power distribution system.
It seems silly to me to have GW of power going down a single thin copper string on a wood pole. If the power lines had a range of max throughputs it would give the incentive to upgrade to medium & big poles, put in step-down transformers, buffer with capacitors or pump storage etc. as power generators were expanded or uprated.

I don't think that will become part of the base game. It can be modded though... Do you mean something like https://mods.factorio.com/mods/JeroenDStout/highvoltage ? I planned to bring that up to 0.18 at some point...

Certainly something like that. Obviously 'fun' is different for different folks! I just think it's very overlooked - apart from poles I'm not sure it's been updated since the dark ages. I have no LUA experience but would be happy to devote some time to help develop a mod.

[wosthisdo]

There are lots of mods that play with power generation (most overpowered), but I'm hoping that Factorio 1.1 might update the power distribution system.

Never gonna happen. Electrical system is heavily ups-optimized. And such a change would turn every factory currently in existance into garbage, forcing players to do major reworks for no fun gain at all. Minecraft on the other hand has a bunch of mods with different tiers, and a bunch of more mods with completely unique power systems. If you haven't tried i very much recommend you do.

What better time to 'turn every factory currently in existence into garbage' than post 1.0 I'm really looking forward to all my blueprints being wiped and starting over (not playing again 'till 1.0 'cos getting a bit of burnout). I've played MC since the alpha, don't really see the connection.

[DarkShadow44]

Certainly something like that. Obviously 'fun' is different for different folks! I just think it's very overlooked - apart from poles I'm not sure it's been updated since the dark ages. I have no LUA experience but would be happy to devote some time to help develop a mod.

Well, Ideas are always welcome. Not sure if there is already a discussion about that. Just need to keep the engine limits in mind, the rest can be modded...

[eradicator]

I'm really looking forward to all my blueprints being wiped and starting over (not playing again 'till 1.0 'cos getting a bit of burnout).

There's not going to be any wipes. Devs know better than to treat their players like trash. I'm not sure if you understand how important for factorio this is. Too many other indy devs ignore this and kill their own player base.

I've played MC since the alpha, don't really see the connection.

I'm talking about modded. Industrial Craft (+Greg), Rotarycraft, Electricraft etcpp. There's lots of mods with complex power distribution systems.

For a mod the only realistic way i see would be enforcing very complex restrictions on pole2pole connections, as there is no moddability of the actual power system.

[wosthisdo]

Obviously not a popular suggestion here, and in danger of getting too far from OP. Will shut up now and look at suggestions thread.