Fix the transport belts already

[Holy-Fire]

This is a thing that I (and possibly others) raised in the past, and I hope it will be solved already.

Each type of belt has a maximum theoretic throughput. According to sslik's measurements (some of which I've reproduced independently) these are roughly 12 items/s for basic belt, 20 for fast and 30 for express.

These numbers themselves are a bit of dark magic because it would be more logical if the ratios were 1:2:3 which is their speed ratio. But let's say we take these throughputs as given.

The problem is that it's very hard to actually reach these throughputs. If your basic belt is curved throughput is reduced by about half. You'll have to reinforce the turns with fast belt to be close to max throughput. Fast belts require express turns, and still take an even bigger hit over theoretical throughput.

If you want an express belt with turns, the only way to maintain throughput (which I haven't tested) is to split with an express splitter, make turns with the two lines, and then merge with an express splitter.

Likewise, if you merge two basic belts with a T-junction, you need to reinforce the junction to end up with a full throughput line.


In addition to these turns business, if you have a line of smelters and you want to load a belt, you won't reach full throughput since you'll have gaps that take space but are still too small for the smelters at the end of the line to insert into them.


Why not just make it so that... Belts have a given throughput, and you enjoy it without fussing too much, no matter which turns and inserters combinations you have?
(I'm sure it's a very challenging problem from the viewpoint of developing the game engine, but from a player perspective it just makes little sense).

[Garm]

I actually like them this way. I know the set of rules they operate at, and I know it is possible to solve or avoid the problems using in game mechanics. And at the end of the day instead of boring lines I end up with fancy networks riddles with different flow normalizers.

[ssilk]

Hm. The solution is sometimes not simple and especially for the last problem you describe with the furnaces in line, there is no real software solution possible, because fixing it would mean, to change the deepest algorithms of a physic simulation.
For example make the items somehow "magnetic" to close the small gaps between them and leave space for bigger gap, in which can be inserted. But the rules are the same, if they are on the belts or the ground. How would that look, if items move by themselves?
Or we make the items on the belt more elastic. That would need some more CPU power.

You see: This would bring in new effects, new physics, which have also side effects. The only way to make belts "just working" is, that they are "simulated", that not every item on it is a real piece of information in your computer, instead it is a formula and the graphics is a imagination of a full belt.

I don't see it. In my eyes that is the heart of factorio, that it creates a somehow physical model and that has its own rules which are valid for every object in that world and that it doesn't show any similar looking estimations. The current state is a good compromise between a dozen different things. I see it as task of the player to see the belts not only as transport, they are part of the game.

I learn how to deal with the belts just by watching.

Sometimes I turned the game speed to only 10%, just to see step by step, what's going on. And the funny thing is: in real factories, there are specialists, which look mechanical processes with high speed cameras.

Just by watching I found out how to use faster belts to compress the items. Often enough it's enough to place one fast belt in front of an inserter, so that he has space to insert.

I find a solution for the most problems, just by watching, but in general I learn to avoid situations. Line of furnaces too long? Remove some and make a second. T-junction? Make a splitter to join them instead. Too many curves? Make two belts or avoid curves. And so on.

I tried to document those insights into the wiki, but how should I tell, when to use which trick?

That's part of the game.

[JackGruff]

The way belts work gives the world some physical quality, it feels like the items are being transported by the belt and not just following a line.

I had much the same grievances when I started. In retrospect, I didn't know what I was doing most of the time.

My suggestion to you is that you look at your setup and try to move things around for more efficiency. I use to make my offloading train stations use splitters but now I've made a really simple one that offloads just as fast, takes advantages of all lanes and doesn't use any splitters. I also hope you are not using belts for copper wire for circuit boards.

As for turns, as ssilk says, use more belts, side by side. You have to accept that a belt's throughput is affected by its course, that's just one of the rules of the game. Don't take their throughput in a straight line as dogma.

I wouldn't mind faster belts. If you want to be able to pack more throughput in a smaller space and are prepared to pay for it, then ok. Btw, you can vote yes here https://forums.factorio.com/forum/vie ... f=5&t=1929

[LoSboccacc]

why do you have so many turns in your belts?

that's the question here. for the belt where throughput is essential I managed to have them straight trough all the factory (it is not a huge factory, but..)

[Holy-Fire]

Of course I can work around these quirks... I just don't see how this makes sense from either a realism or gameplay viewpoint, and it leaves me spending mental energy on things I find neither interesting nor entertaining.

[TGS]

Of course I can work around these quirks... I just don't see how this makes sense from either a realism or gameplay viewpoint, and it leaves me spending mental energy on things I find neither interesting nor entertaining.

They aren't broken. They work a certain way. You want them to work a different way. That in itself does not make them broken. Broken would be if they were not working the way they were intended. Just consider it variance. Something the real world has plenty of.

[ssilk]

If you don't want to spend energy - which is your right as player - then I recommend, what I do, when I won't spend energy in them: I make a splitter and create a second belt. All you need to care about then is to leave enough space and make from time to time a new splitter. or in other words: parallelization rulez!

I've made a factory like that, the needed space is about a third bigger. And at some point I couldn't resist to optimize the flows, using both sides, etc. after that the thruput was phenomenal. So included the planning with two belts per item (mostly iron and copper) into my planning, when starting a factory and till now this is really a good idea.

[ray4ever]

I'm not sure if they have changed something on the belt speed with 0.9, but now 8 electric furnaces (without modules, speed = 1.0) on each side (16 in total) are perfectly able to completely saturate a basic belt without turns. Any additional furnace put in line wasn't able to put a piece of iron plate on the belt so it seems that a straight belt is able to transport 16 items per second.

Testing method: I had 10 electric furnaces (which were always on stock with iron) on each side of the belt, blue inserters to put the plates on the belt and in total 6 blue inserters to get the iron plates from the belt and putting them into chests (only 4 were always working, the other were just for some overflow as 4 were not completely able to keep up with the belt.

[sparr]

Some of us like dealing with the emergent efficiency concerns of simulation rules.

https://wiki.openttd.org/Corners

If the slow-down in a turn bothers you, use fewer turns!

[immibis]

I agree with TGS and sparr that this is not broken. You can move items by using belts. If you want to move more items, use more or faster belts. It only gets complicated when you want to squeeze every last drop of throughput out of your belts.