Why having everything on a belt is a bad idea

[-root]

Zisteau just released his latest factorio episode and I thought it warranted a bit of a comparison and discussion :

Zisteau's Factory

In the episode, Z plans to put 80% of items onto a belt. :

Here is a link to one of my very early fixit episodes with me giving a whole heap of reasons why it is bad. It features a factory where it takes the "ultimate central bus" design to the n-th degree and at the same time shows why it is a really bad idea.

Factorio Fixit 8

Basically, this kind of idea is going to be very hard to upgrade and very slow heading into the late game. The lack of throughput is going to be extremely hard to overcome the further down the line you get. In fact, in my playthrough with Oni, we're starting to hit some throughput issues .

On the topic of aesthetics I think that a compact factory with shit running everywhere that actually works is a lot more attractive than a bloated one that creeps along at (to borrow Oni's expression) the speed of smell .

TL:DR? Don't make things harder for yourself than they need to be.

[sillyfly]

While I agree that Zisteau's approach is a bit too much, I must admit I find it oddly aesthetic. Beauty comes in all shapes and sizes, not just lean and efficient
But yes, if you have high throughput in mind, the central bus is not a very good idea.

[ssilk]

I also shaked the head about putting copper cables on the belt.
The lack in throughput for green circuits could be already seen in the last video...
But well. I think he tries to think in categories of newbies: They like this "all on a belt"-idea. I see that a lot, they really like it.

And it makes really sense to have such a central bus for copper, iron, circuits and maybe some more...

It's eventually our "duty" in the forum (and wiki and more) to always put a finger on this and say: "Well, nice and clear, but useless..." ... like in this thread....

[User_Name]

I also shaked the head about putting copper cables on the belt.
The lack in throughput for green circuits could be already seen in the last video...
But well. I think he tries to think in categories of newbies: They like this "all on a belt"-idea. I see that a lot, they really like it.

And it makes really sense to have such a central bus for copper, iron, circuits and maybe some more...

Not sure about circuits.
They are 1 iron 1.5 copper. Not that much benefit over just transporting iron and copper.
Making them on the spot in significant numbers don't take much space, and you'll get the benefit of inserter stack bonus, which is almost always useful, given that the demand for circuits for the most recipes is tremendous.

Let's compare that to steel.
1 steel plate per second requires 9 electric furnaces, 9*4*4 = 144 tiles without infrastructure (inserters, belts, poles, spacing).
It will take 5 items per sec from your metal bus.

1 electronic circuit per second requires 2 assemblers = 2*3*3 = 18 tiles (in fact you will get 1.7 circuits per second from this setup)
It will take 2.5 items per sec from your metal bus.


I'd rather added belt for steel.

[ssilk]

And it makes really sense to have such a central bus for copper, iron, circuits and maybe some more...

Not sure about circuits.
They are 1 iron 1.5 copper. Not that much benefit over just transporting iron and copper.
Making them on the spot in significant numbers don't take much space, and you'll get the benefit of inserter stack bonus, which is almost always useful, given that the demand for circuits for the most recipes is tremendous.

Well, can be discussed. It is, cause you need so much of it. Simple calculation: Let's assume your INPUT is 2 belts of copper and 2 or iron. 4 in total. And as output you need ... red, green and blue science packs. Then between you need - all in all - about 50% green circuits. All in all, measured on the input and about 10% plus/minus.

So in total you spare two belts inbetween (one with copper, one with iron) and add about one with circuits.

1 steel plate per second requires 9 electric furnaces, 9*4*4 = 144 tiles without infrastructure (inserters, belts, poles, spacing).
It will take 5 items per sec from your metal bus.
1 electronic circuit per second requires 2 assemblers = 2*3*3 = 18 tiles (in fact you will get 1.7 circuits per second from this setup)
It will take 2.5 items per sec from your metal bus.

No question, that steel makes sense on belts, but you need much less of it.

[Xterminator]

I gotta agree with root here. While it may look nice (or horrible) pretty fine line there, I usually find it very inefficient to have much else besides the basic materials on belts. Usually the only things that I have going through the whole factory on belts are iron/copper, sometimes steel and that is about it. The rest of the stuff, I just make in place when I need it.

Of course it can be situational. I will use belts to transport like circuit s or something a short distance to 1 thing that requires it. But never through the whole factory. :p

[The Lone Wolfling]

The problem is that no-one ever seems to leave enough space.

People forget that almost every recipe in the game consumes more items than it produces. (Exception: copper cable). Which means that your line of belts will have to be at its largest at the start, and shrink from there. At least if you want the throughput numbers to work out.

Doubly so if you're putting terminal items into logistics chests.

[Drury]

I love belts, and I love cramped spaces, and figuring out where belts go in cramped spaces is like orgasm.



Here's a screenshot of my game, edited a bit by the devs. They removed a bunch of stuff, somehow managed to screw up somewhere and clog up absolutely EVERYTHING and placed some roboports. In my game there were pretty much no empty tiles. Every little nook was filled by a bendy belt, inserter, electricity pole or an assembly machine. Eventually I managed to upgrade to express belts with no issues.

Is it the most efficient thing to do? Nope. Is it nice to look at? Your mileage may vary. But it's fun as hell. I love bootleg designs quickly thrown together on the go, cookie cutter setups are boring.

Not sure what's the point of telling people how to play their singleplayer videogames. If I didn't like playing this way, I wouldn't.

[-root]

But yes, if you have high throughput in mind, the central bus is not a very good idea.

But thats the thing, it is a good idea, you just don't put every single item onto a belt and run it through the factory.

Central Bus Design Concept - Youtube

Here is a video I recorded last night of building a central bus factory without the bloat. It took me roughly 30mins once I had everything sorted.

[HerrJoebob]

TL:DR? Even though this game is prerelease I've decided there is One True Way to play

FTFY

[-root]

TL:DR? there is no perfect way to play yet. however there are good ideas and then there are ideas that need work.

FTFY

[ShneekeyTheLost]

But yes, if you have high throughput in mind, the central bus is not a very good idea.

But thats the thing, it is a good idea, you just don't put every single item onto a belt and run it through the factory.

Central Bus Design Concept - Youtube

Here is a video I recorded last night of building a central bus factory without the bloat. It took me roughly 30mins once I had everything sorted.

The problem here is that, for example, gear wheels are being produced in several substations, which is rather redundant. I'd rather have all the things requiring gear wheels together, then gear wheels line ENDS.

Now, there's a fair bit you want to produce en masse that require gear wheels. Belts, inserters, splitters advanced versions of the aforementioned, red science packs, and if you are playing on a difficult map, turrets. This is all able to be done in a fairly straightforward manner because advanced versions require the basic versions, so those lines can be concurrent with the proceeds from the lower tier feeding into the higher tier. However, of note is that other than upgrading to better versions, the only item that is really required elsewhere are going to be your belts and inserters for green research packs, and the red packs going to your actual science facility.

So you've got a few lines of production splitting off from the main bus that require gears. For that space, gears get their own line. But once you are done with those items, it terminates, probably into a passive supplier chest so you've got some gears on hand for other things that may not be needed en masse but will certainly be needed to build from time to time (rockets, cars, and assembly machines would be examples, although if you're really wanting to do a 'gray ooze' scenario, assembly machines might be made enough to warrant being included in the former category).

As materials come and go, they are added or removed from the main bus. Under no circumstances is everything on the main bus at any given time, but you have centralized component building in one location rather than have it scattered all over your base.

I can certainly agree that once you've built everything you are going to build, running lines for that item is pointless. However, having, for example, gears built at six different locations is just inefficient when you can get all your gear needs taken care of with a third that many.

[Peppe]

Looking at where resources tend to flow seems like instead of 2 copper and 2 iron lines running down the length of a base it could be served with 2 green circuit lines and 1 each of copper and iron.

I was planning to basically take two lines each of copper/iron smelting at the front of the base to form typical quad line, but funnel most of it off to run green circuits at the start of the base and replace a line each of copper and iron with green circuits. Every other copper sync in the game should be covered by a single yellow for a long time. Iron still has some heavy consumers but a single red line should carry it to late game.

I think I heard someone call this compressing your resources before putting them on a belt? So instead of sending say 12 iron/s and 12 copper down two yellow lines I can send 12 green circuits a second down one line which is the equivalent of 12 iron and 18 copper.

Looking at my current game stats green circuits are the biggest resource consumer by far, so to me it makes sense to make them part of that core bus.

I'm thinking 6 base wide belts:
2 Iron (Reduce to one after green circuits) = ~shoot for 30-40/s
2 Copper (Reduce to one after green circuits) = ~shoot for 30-40/s
2 Green Circuit (12 total = 4x inserter area, 3x for red circuit, 1x for Modules, 4x Blue circuit) = 24 green/s
.5 Red Circuit (8x for science, 16x for rest for modules/misc) = 3 red/s
.5 Blue Circuit (6x blue circuit) = .4/s
.5 Steel (~10x) = 1.4/s
.5 Battery (5x) = 1/s

Steel+battery share a belt to blue science and periodically I find I need one or the other in different spots, so it is nice to have available on a main line. They aren't high volume though so can get by with just half a yellow belt.
Red and Blue circuits are low volume as well and tend to be consumed together in most spots, so good to run along with the main trunk to the different areas.

I plan a science area of 1 pack of each per second the 5/6/12 ratio.

I think it is a descent plan... Looking forward to restarting

[ccik]

I really like the idea of having a central bus system. But it is obvious, that not all items are suitable to be transported all throught the base. Mainly copper wire and everything that is only needed once.

I've found for myselft, that to have a effective central bus system, there are 3 important points:
1. The central bus has to be wide, with enough room to put a second belt for every item you want to transport. Especially true for iron and copper. For these two, i normally split them up right at the beginning, and then have two belts each going downwards. In later game, when i need more iron and copper, i can just easily cut the connection, and have two belts coming in from my train station, so i can double the available materials without changing much of the belts. This also applies for some other items, like green circuits.

2. One of the most crucial factors in a central bus system is the priority of items. Everything down the line only receives a fraction of the total materials available. Meaning, my gear production at the very beginning receives 50% of all iron, my blue belt factory down at the end receives basically only the items not used before. Therefor, you just cant put factories down, they have to be in the right order. Intermediate products at the beginning, important stuff next, and items you only need once a while at the end. This system is kind of self regulating, if my factories at the end don't receive any items, they shut down, meaning the intermediate products will back up. This shuts down the intermediate factories, providing more raw materials for the factories at the end. But this system of regulation is not effective. This brings me to the next point.

3. I always overproduce intermediate items. If a belt is not backing up, there are not enough items.The bad thing here is, that i have a lot of items waiting on a belt and probably wasting a lot of efficiency. On the plus side, i can keep the bottleneck always down the line. Every factory has the bottleneck, and by overproducing intermediate items, i can assure that the bottleneck is always in the final product. That beeing said, if i wonder why my blue science production is too slow, i already know i need more factories producing them. And that is the only thing i have to put down, 2 more factories, maybe later upgrading intermediate stuff.


I really like the central bus design. It may not be the most efficient or the fastest solution, but i think i looks nice, get's the job done, and it is easy to analyze where your bottleneck is.


Edit: here is a screenshot of my current factory:




Note the space in between my iron and copper lines for expansion, altough this is a little bit too much. Also, i'm currently lacking green and red circuits, so that is my next planned update for the factory.

[Gamati]

As someone who found this game through Zisteau and having watched him play other games for over a year now, I think he's really trying to play this game in a way that's accessible to most of his channel. His general style toward games is to go way overboard, usually sacrificing efficiency for something that is excessive and interesting. I'm sure he realizes that having a dedicated belt for nearly every product isn't actually the best strategy- but I think the IDEA of it is what he's attracted to. Until Factorio has been a part of his channel for several months, I don't think any of his videos will be geared toward playing the game at the level of Negative Root or similarly amazing players simply because he's playing the game both for non-Factorio-playing subscribers and for people who are just getting into it.

[Tenandra]

As someone who found this game through Zisteau and having watched him play other games for over a year now, I think he's really trying to play this game in a way that's accessible to most of his channel. His general style toward games is to go way overboard, usually sacrificing efficiency for something that is excessive and interesting. I'm sure he realizes that having a dedicated belt for nearly every product isn't actually the best strategy- but I think the IDEA of it is what he's attracted to. Until Factorio has been a part of his channel for several months, I don't think any of his videos will be geared toward playing the game at the level of Negative Root or similarly amazing players simply because he's playing the game both for non-Factorio-playing subscribers and for people who are just getting into it.

I agree with Gamati, but also i would like to mention that efficiency is not the only fun to find in the game, a lot can be said for aesthetics, and elegant design. I am also not sure how Negative root and others see the end game in Zisteau's world playing out, is it that you see the system being designed as grinding to a halt in the end? What real issues do you see with the design in terms of game completion? If its is just a matter of well it will take 1.5 seconds extra to produce this and that component that can be said by many a player as an acceptable way to achieve the result. But, one thing I guess us new players paticulary those of us who follow Z should keep in mind is that those players who have been with this game for a long while toiling and displaying its virtues that we don't stamp all over those loyal players and keep in mind this is clearly a small friendly community that we need tomake an effort to adjust too.

[katyal]

Unless you plan ahead and know how much of each item you will need then the biggest problem with everything on a belt is throughput. As the line gets longer and more processes use up materials on the belt you will find the belts getting more and more empty yet adding more miners or furnaces won't help because in this unfortunate situation the belt itself is your choke point.

That being said I'm actually quite facinated by Z's attempt and I'm following it closely. To me one of the big selling points of this game is its sandbox nature. I'm always on the lookout for people doing Let's plays with different play styles or self imposed restrictions because they're good places to learn and find setups and layouts to suit special needs.

[DerivePi]

Things that should not be on a belt in the base game:
- small, medium and large electric poles
- iron chests (they should never be made)
- alien goo (you have to handle it so you might as well put it where you want it)
- turrets
- shotgun ammo
- poison caps
- cars, locomotives, tanks
- destroyers
- Express belts, X-splitters and X-UG belts
- Steam generators, boilers, water pumps
- Assembler 3s
- electric miners
- long inserters
- wood chests
- log bots and con bots

The main advantage of the central bus is, with concentrated production of items in one spot, shortages can quickly be identified and addressed. An example would be a shortage in gears - identified by no or light amount of gears on belt. Addressed by adding/upgrading a gear assembler next to the existing gear assemblers. OR if not enough iron plate making it to the gear assemblers, upgrade the iron plate furnaces. If you have green circuit production dispersed, it is not as obvious.

[-root]

I am also not sure how Negative root and others see the end game in Zisteau's world playing out, is it that you see the system being designed as grinding to a halt in the end? What real issues do you see with the design in terms of game completion?

You've hit the nail on the head . That is exactly what is going to happen. As more and more things get added, the throughput will become more and more of a problem and it'll eventually stop.

I think there's something important to be said. The way Z is going about his build is genius. I think that's important to recognize. The design and the execution is on a whole new level and I am starting to see what he is getting at. My point is, I just want to make sure that all the new players don't take what Z says as gospel and realize that there are some better ways to do things. He is, as you said, trying to make the game accessible by creating a factory that is really visually appealing and I'm starting to see his vision.

[Tenandra]

I am also not sure how Negative root and others see the end game in Zisteau's world playing out, is it that you see the system being designed as grinding to a halt in the end? What real issues do you see with the design in terms of game completion?

You've hit the nail on the head . That is exactly what is going to happen. As more and more things get added, the throughput will become more and more of a problem and it'll eventually stop.

I think there's something important to be said. The way Z is going about his build is genius. I think that's important to recognize. The design and the execution is on a whole new level and I am starting to see what he is getting at. My point is, I just want to make sure that all the new players don't take what Z says as gospel and realize that there are some better ways to do things. He is, as you said, trying to make the game accessible by creating a factory that is really visually appealing and I'm starting to see his vision.

I have watched Zisteau for some time, i pretty much figured early on that there is little point trying to copy a build, i simply dont have the level of patients required or indeed as would be required in this instance the painstaking head banging required to get everything to the right dimensions. I have been diping in and out of your fixit series with some success in helping, me with my own build, which i will say is far more simpler. So allow me to thank you for that help..