Lessons learned. Please tell me more.

[Laie]

(full-res png also available)

My first factory. The general idea was that I'd pipe in resources from the left and bottom, consume them on the right, and add new production at the top.

Lesson learned: I should have disposed of my neighbors much sooner. Constant defense was quite a distraction. Instead I switched to AP ammo quite early... now *that* took some resources, increased pollution considerably, and only made the problem worse. The whole military science sub-assembly is only there because I expected to need lots and lots of wall and ammo. In the end I mostly needed the bricks.

Another lesson: It requires much less input than expected, hence all the empty space on left side. So much for drawing up a floor plan before I have the faintest idea how things will play out. Well, speaking positively, I certainly didn't paint myself into a corner there

Now for a question: I'm doing 45-50 science/minute and new research comes in faster than I can implement it in production. The science subassemblies are built such that I could just extend them to the right and increase production by a factor of 2-4x; However, I'm getting the idea that I won't scale up production of individual items any time soon, while the number of different Items I want to produce will increase quickly. Is that so? Will it go on like that? Then I could perhaps risk adding a second tier of item production to the right, so that it all doesn't stretch out only in one direction.

For this and all following questions: I'm on default settings and my only goal is to "finish the game": build and launch a rocket.

I'm totally clueless as to how I should integrate the refinery. As you can guess, it's currently far away at the bottom of the map; I wonder if I should move it up closer to where it's products will be needed. Mostly that comes down to the question of how I will deal with liquids: is there any point in trying to cluster the production of liquid-dependent things around the refinery, or will I end up stringing long pipes no matter what?

Oh, and... For the time being, I stocked up on intermediaries (Steel, Cogs, Circuits, ...) and crafted the finished products as I went along. Even if I could -strictly speaking- go on like this, I guess I will soon need some real assembly lines to keep myself stocked with all that I need. Could you perhaps give me a hint as to how much space I should reserve for consumer goods production? If not in absolute terms, then perhaps as a fraction of factory size?

Lastly... I may be like the guy who wants to drive in a nail and wonders whether it's better to use an empty bottle or an old shoe. In that case, please don't spend much time explaining the pros and cons of either, but just tell me about hammers instead.

[Mskvaer]

Hi,

I've been there, too, even recently. Now I like the experimentation, I have avoided looking at "solutions". I only got the rocket done on the 3rd game - the previous two got stuck on impossible dense layouts (something you have avoided) and getting exhusted on maintaining wall/turret/ammo. 3rd game I just kept hammering (with a shoe or a hammer? ) away at it. When faced with Rail layout I started a seperate Sandbox game in Peacfull mode and just did Rail experiments. (build rails, transport something to and fro, tear down, repeat) Armed with better knowledge I then rearranged base in the game. This way I spent about 70 game hours in Game3 getting the first rocket and 50? 100? hours in sandbox learning about features. By sheer ignorance I avoided logistic wiring/circuit and logistic robots - I just beelined for the rocket ignoring tech that wasnt needed (except for rockets in miltary and armour+ suits) Bonus two achievments.

So - IMHO - you're doing great. Hope you enjoy doing it!

My experience is that some "consumer items" you need lots of when they are "newly discovered", so I build assembly lines and use them as fast as they are made, and a little later in the game they are just stockpiling. As you can not undo an item, stockpiling is a very mixed blessing. I never seem to produce steel fast enough, though.

I have never measured my Bottle/minute or suchlike. This is one item where I happiily stock pile; two insertes and a big box as a buffer in the belt line. My labs (typically 8) only run 60% of the time, as I do not keep up with using all the tech I keep discovering. It takes time to layout the refinery, it needs resources or I am out purging nests, so the labs can idle for a while, stockpiling bottles.

Just remember : The Factory Must Grow !

[Serenity]

Oh you sweet summer child. It's purple and yellow science that will really strain your production. So it's good that you left space. You need a lot of steel for the rails in purple science. And red and blue circuits need a lot more than what came before.

Look at the costs. Blue science needs a lot of machines, but only because they craft slowly. The costs are still relatively low:
https://wiki.factorio.com/Science_pack

But how you much you need really depends on your goals. You can launch a rocket with very little if you just wait long enough.


Here is what you need for 45 science per minute:
https://factoriolab.github.io/list#p=fo ... el-furnace

Around 9 yellow belts of iron and copper each. 3-4 belts of green circuits, of which 37% go to red circuits and 50% to blue circuits


Btw, belting copper wire for an extremely fast craft like green circuits is not such a great idea. You need 3 wire assemblers for 2 circuit assemblers. That can be neatly done with direct insertions. Belting cable is fine for slow crafts like red circuits.

Could you perhaps give me a hint as to how much space I should reserve for consumer goods production? If not in absolute terms, then perhaps as a fraction of factory size?

You mean things you need for yourself to build the factory? Just look at how many different items there are:

Mall

Mall.jpg (558.22 KiB) Viewed 4220 times

And that doesn't include stuff like power armor equipment. Laser turrets, solar panels and accumulators also aren't there as I do those seperately.

Of course you can leave some things out. You don't necessarily need to mass produce artillery turrets for example. And things like combinators can be handcrafted if you don't use many of them

is there any point in trying to cluster the production of liquid-dependent things around the refinery, or will I end up stringing long pipes no matter what?

Plastics can be worthwhile near the refinery as it consumes most of the petroleum and just needs coal. Sulfur too. Acid is needed in several locations, so that can easily be piped. Lubricant just needs heavy oil, so it makes more sense to make it next to the refinery and pipe the lubricant. Solid fuel/rocket fuel needs a lot of light oil, so that's better near the refinery too.

But there are plenty of options. You can have a small liquid bus with water, petroleum, acid and lubricant

[astroshak]

As far as the refineries go ... I tend to train Crude Oil in, and make stuff with it at a central location.

I will make Plastic, Batteries, Sulfur, Explosives, and belt those to the rest of the base where needed. I will also make Lubricant and Sulfuric Acid and pipe those where needed.

Lubricant is used in making Electric Engines, as well as the “Blue” belts/undergrounds/splitters. Sulfuric Acid is used in mining Uranium, and in making “blue circuits”. None of these get manufactured near the oil processing plant, so they need their liquids transported from the refinery area.

[Zavian]

I'm totally clueless as to how I should integrate the refinery.

My general preference is to build it near the factory, as it is easier to pipe/rail crude oil to the refinery, than to pipe/rail/belt refinery outputs long distances to where they are consumed. Pick a clear space where it isn't going to be in the way, with access to water and room to expand.

Could you perhaps give me a hint as to how much space I should reserve for consumer goods production?

Many of the basics and most the intermediates for common consumer items will be produced as part of your science builds. I generally add a chest (limited to a couple of stacks) to make it easy to pickup some of these. (Later on you can replace these chests with logistics chests). For other stuff that I want in bulk I will add dedicated assemblers, somewhere convenient, often adjacent to a science build that uses the same (or similar) intermediates. (eg laser turrets use the same resources as robot frames, so I often build laser turrets beside bot frames).

Building a dedicated mall area is also a common strategy. (Yellow + red belts + splitters + undergounds, red + yellow + blue inserters, pipes, mining drills and lvl 1 assemblers all only need iron + gears + green circuits. Add steel and you can also automate lvl 2 assemblers. Add red circuits and you can also automate stack inserters + roboports etc). But I never bother to automate everything. Provided I have the intermediates, it is generally more convenient to just hand craft things I only use occasionally.

Here is what you need for 45 science per minute:
https://factoriolab.github.io/list#p=fo ... el-furnace

Around 9 yellow belts of iron and copper each. 3-4 belts of green circuits, of which 37% go to red circuits and 50% to blue circuits

That is a little misleading, as it includes the resources for space science packs, that won't be needed before launching your first rocket. It also includes military science which will often be idle.

Try https://kirkmcdonald.github.io/calc.htm ... BBnL/yAQ== instead.

So a total of 5.3 yellow belts of iron ore (3 of which are for steel) and 4 belts of copper ore (2.25 for green circuits), and 1.5 belts of green circuits.

Once you have researched the rocket you will also need resources to build it, but the goal is simply to build a rocket then you can stop research and redirect all the resources to building the rocket.

Once you do launch your first rocket, then you will want those extra resources for space science. Plus you will probably want to scale up production to 75/90/150 or more per minute, as techs will start getting more expensive.

[foamy]

My own refinery preference is to pick somewhere well out of the way (refineries are big, right up there with chip farms and smelting arrays) but with direct access to water. If there's a large enough lake I'll probably build it out with landfill, even. Bring in oil, by pipe at first but by train as soon as practicable (it's often usually the very first train stop I build) and haul out the petroleum gas and, before going to single-product stuff is worthwhile, the lubricant as well. Since nothing else uses it except for flamethrower turrets -- which work just fine with crude anyway -- I don't bother putting light oil on the railnet and just turn it straight into rocket fuel at the refinery.

Plastic is so very simple to make -- and expands on belts so much -- that I prefer to build it as close as possible to the site of consumption and ship/pipe in coal and petroleum gas. Ditto acid: petroleum gas and a tiny trickle of iron are more than enough to satisfy the only things in the sciences that need it, batteries and blue chips, so I prefer to simplify things and just make that on-site as well. Uranium mining does need it on the net eventually, but it's a tiny amount you can make basically anywhere you have petroleum gas, water, and iron, all of which are very common ingredients. You can literally build a plant right next to a train stop for that.


An actual lesson learned though:

Use filter inserters whenever you're feeding something other than an assembler, and set filters on your train wagnons (middle mouse button). Very few things are more annoying than a part of your factory stalling out because the wrong kind of ore wound up in your smelters or something.

[Laie]

I've been there, too, even recently. [...] So - IMHO - you're doing great. Hope you enjoy doing it!

I have never measured my Bottle/minute or suchlike. This is one item where I happiily stock pile; two insertes and a big box as a buffer in the belt line. My labs (typically 8) only run 60% of the time, as I do not keep up with using all the tech I keep discovering. It takes time to layout the refinery, it needs resources or I am out purging nests, so the labs can idle for a while, stockpiling bottles.

Thanks for the heads up!

I started with a 3x3 lab setup, later expanded to 4x4 -- that's faster than I can create the science packs, but if I don't keep my eyes on the ball and forget to start new research ASAP, it piles up so quickly and I want to pick up the slack. I've buffered military science in a box, but now that you mention it, i should probably do that for all science packs.

Btw, belting copper wire for an extremely fast craft like green circuits is not such a great idea. You need 3 wire assemblers for 2 circuit assemblers. That can be neatly done with direct insertions. Belting cable is fine for slow crafts like red circuits.

So a setup like this would balance out?

Many of the basics and most the intermediates for common consumer items will be produced as part of your science builds. I generally add a chest (limited to a couple of stacks) to make it easy to pickup some of these.

I've done that too, my assembly line for Red&Green Science creates a sufficient excess of belts, circuits and inserters to keep me well-stocked. That was here, though, while cogs and pipes where over there and I thought it would be more convenient to have it all in one place. That probably doesn't matter all that much anymore once you have logistics robots?

Try https://kirkmcdonald.github.io/calc.htm ... BBnL/yAQ== instead.

I knew about kirkmcdonald's calculator, but didn't notice that I could check for more than one output -- that made it pretty hard to get a big picture. You have just given me a strong hint, thanks!

[Zavian]

So a setup like this would balance out?
http://factorio.schnobs.de/spool_balance.jpg

Yes although it needs faster inserters. (A lvl 2 assemblers making copper wire makes 3 copper wire per second. At starting tech a yellow inserter moves 0.83 items per second. A blue inserter move 2.31 items per second. So typically use at least one blue inserter per copper wire assembler. (2 for full speed production). You can also use a one to one layout, with lvl 2 assemblers (craft speed 0.75) making wire, and lvl 1 assemblers making circuits (craft speed 0.5).

I've done that too, my assembly line for Red&Green Science creates a sufficient excess of belts, circuits and inserters to keep me well-stocked. That was here, though, while cogs and pipes where over there and I thought it would be more convenient to have it all in one place. That probably doesn't matter all that much anymore once you have logistics robots?

I typically add most of my early game chests near the belt/inserter setup for green science. It already has the ingredients for red + blue inserters, splitters, undergrounds, gears, pipes, miners, and lvl 1 assemblers. Depending on how fast you are using things, the belts might not supply enough resources for everything to keep up, but they are all fast to handcraft anyway, provided you already have gears + circuits. Do make sure you have a box of pipes somewhere. Makes handcrafting underground pipes much faster).

Gears and green circuits will be boxed near green science. Red circuits will be boxed at red circuit production. Engines at the blue science setup. Military stuff in the mil science build. So for me most of the stuff I box early game is fairly central. Late game I mostly let bots keep me supplied. (Although I do build an outpost train, if I'm going to keep playing on the map).

But ultimately it about player preference, so do what works for you.

[Serenity]

I've buffered military science in a box, but now that you mention it, i should probably do that for all science packs.

It depends on your playstyle and goals. If you stick with a relatively low amount of science it helps with costly research projects that take several thousand packs. So it gets you to launching that rocket and "winning" faster.

But if you want to continue beyond launching one rocket and take your time it's not needed. You will probably have a larger amount of science production so even end game research is quite fast. And belts also buffer quite a bit of science if they are long enough.

I knew about kirkmcdonald's calculator

It's the best known one because for a long time it was easily the best one. But with Factorio Lab there is a contender now. It does some things much better like allowing you to specify the output in number of belts. Mixing belts and assembler levels. Or adding all the sciences with a single click

And in both of them you just need to click on a product in the list to temporarily remove it from the calculations. So you don't need to to remove it entirely open a new window to see how things change

I thought it would be more convenient to have it all in one place. That probably doesn't matter all that much anymore once you have logistics robots?

Even when you get bots it's good to keep most things central so you don't have to wait for them to fly to you from all over the place.

[starlinvf]

To the original question.... Liquids tend to be easier to handle if you move crude in bulk to a central refinery area, and pipe output products to locations you need them. Barring mega factories that needs an entirely different style of logistics, a basic factory for rocket launch can reasonably work (for most of the game) with 2 Refineries with linked outputs to a common storage area, and set up and Oil Cracking chain (Heavy -> Light -> Petroleum) between storage tanks for each type to act as a product balancer. You do this because if any of the outputs get backed up, it stalls processing entirely.

The typical problem with oil refinement is that you need a lot of petroleum for plastics at the onset, but have no use for Heavy or Light oil until much later (and why it stalls easily). You can convert Excess light and heavy oil into Solid Fuel to prevent stoppage. However....if you care about the math, this is inefficient. Hence Oil Cracking chain to break it down everything into petroleum, until you're ready to use Light and Heavy for other purposes.

For a simple setup, you can use Power switches to manually turn oil cracking on or off as needed. For an intermediate setup, you can get a simple circuit network of SR Latches to monitor the levels for you.

If your production allows for higher SPMs, you can easily expand the number of refineries on the main network, or set up a second one in another convenient location, and find a way to move excess fluid products between networks. Depending on distance, barrels and trains might actually be useful to avoid complex pipe set ups in the middle of the factory. But in most cases, a decently planned factory, and some reserved floor space around your belt ways should have enough room to allow pipes to cross from one side of the factory to the other.

You don't "need" Coal liquefaction; but does prove useful in late game for solid fuel/rocket fuel production, since coal is way more abundant then oil wells. You can basically set up the entire production line around the coal patch, and transport the finished rocket fuel to the silo.

[mergele]

[..] For an intermediate setup, you can get a simple circuit network of SR Latches to monitor the levels for you. [..]

Way excessive already. Just do a pump with a wire to a tank that activates when the tank is almost full.

[BicycleEater]

I tend to find the early game the most fun - the rush to blue science, and nuclear power, and tend to stop playing at that point, or sometimes launch a rocket, then call it a day.
I usually play games with the goal of reaching nuclear power as soon as possible - it's not efficient, but its a good goal - and then relax from there.
I also find the game a lot more fun when playing without any pre-planning whatsoever, it produces more organic factories which are more fun to work with, and in the post-victory stage, I tend to ignore my old base anyway, building with lots of rail based systems.
My opinion on this is generally: main buses are amazing, but are almost always better small, having just a copper line and an iron line next to each other works fine. If it needs more capacity, run a belt to a late area in the bus, when it starts to run dry and go from there, it means you make more progress earlier on.

I find that (at least up to winning the game), there is no reason to go beyond 45-50 science/minute, as not may researches require that much, so I tend to not bother making early sciences very scaleable, you probably won't need it anyway (if you do then assembly machines MK3 are an option).

I tend to play on default, (or rail-world if I want it easy), and find that pollution doesn't matter that much, it is too abstract to measure easily, and massed gun turrets with a short local ammo belt are cheap and devastating.

For building the consumer good area, I tend to find it doesn't need to scale that much, even into the relatively late game. All in all, they tend to end up about the size of the blue potion facility, if you include the engine factory - maybe 20 machines? I also tend to build a second area, when I get onto needing more stuff in large supply (e.g. red belt, chemical plants, etc.)

The advise I am giving is how I play the game - not very scalable or efficient, but generally good enough to win the game pretty quickly (I reckon if I aimed for it I could get the within 15 hours achievement with this approach).