Exactly why is it bad to transport things with belts over long distances

[Ultra]

most people who play this game seems to agree that its a bad idea to transport things over long distances with belts. I wanna know why. Well the obvious answer I will probably get is that its alot slower than trains. But why is that a problem.

Lets say I have a huge circuit factory that is producing a compressed belt which is traveling for 5 minutes until it reaches its destination. The destination will recieve a compressed belt. which is the same amount that the factory is producing. I dont really see the drawback. you could argue that as soon as the factory starts I will have to wait 5 minutes until the circuits are used. but that belt storage will be used up as soon as the factory stops. Can someone please explain in detals why its a problem. I dont understant why having long belts will slow down my production.

[KatherineOfSky]

The answer is: you will get FAR HIGHER throughput with a train than a belt, esepcially at those distances.

Also: you use up too many resources creating belts when train infrastructure is much cheaper. You have to defend the entire length of the belt, (especially in 0.13.x) since biters are much more agressive, whereas train track is generally safe from biter incursions.

Another benefit: you can use trains. Trains are fun!

There are probably CPU use reasons as well, but I don't know the details enough to put that information forward.

As far as slowing down your production -- you just answered your own question -- it's constantly delayed. It takes you 5 minutes to see if there is a problem on the line... you have to out there... then takes 5 minutes for fixed line to get back to the factory.

I would highly recommend trains, especially for end-game... one "compressed belt" is insufficient. I need 4 full-compressed (blue) espress belts for decent rocket production. The thought of producing miles of blue belts is wastefulto say the least!

[brian_va]

^ exactly. 1 compressed blue belt just cant bring in enough ore for high production. only real solutions aside from trains would be to run multiple belts from the deposit to the factory, or bring in belts from multiple deposits. early game it could be fine while production is still fairly low, but at some point only having that 1 belt will lead to plate shortages.
but, at the end of the day, its your game. if you want to belt in ore from the other side of the map, knock yourself out.

[Ultra]

Originally posted by KatherineOfSky:

The answer is: you will get FAR HIGHER throughput with a train than a belt, esepcially at those distances.

Also: you use up too many resources creating belts when train infrastructure is much cheaper. You have to defend the entire length of the belt, (especially in 0.13.x) since biters are much more agressive, whereas train track is generally safe from biter incursions.

Another benefit: you can use trains. Trains are fun!

There are probably CPU use reasons as well, but I don't know the details enough to put that information forward.

As far as slowing down your production -- you just answered your own question -- it's constantly delayed. It takes you 5 minutes to see if there is a problem on the line... you have to out there... then takes 5 minutes for fixed line to get back to the factory.

I would highly recommend trains, especially for end-game... one "compressed belt" is insufficient. I need 4 full-compressed (blue) espress belts for decent rocket production. The thought of producing miles of blue belts is wastefulto say the least!

I agree that Trains are fun. I have built a rail network myself and it worked very well. But I dont quite understand the math which explains why you get more throughput with a faster transportation line. Once evrything is up and running the distance should in theory not affect the throughput.
if a problem occurs on a 5 minute belt line it will take 5 minutes until that problem affects the rest of the production since there is alot of belt storage. but it will also take 5 minutes after the problem is fixed until it affects the rest of the production. Why is that system worse than having a 10 second long belt line where it will take 10 seconds until a problem affects the rest of the production and 10 seconds until a fix affects the rest of the production.

Im not saying that having long production lines are a good thing but rather that in theory it should not make a huge difference, If it makes a huge difference I want to know why.

One problem with having long lines is that its more expensive to build.

[AlexMBrennan]

As far as slowing down your production -- you just answered your own question -- it's constantly delayed. It takes you 5 minutes to see if there is a problem on the line... you have to out there... then takes 5 minutes for fixed line to get back to the factory

It is silly to argue that this doesn't affect trains or indeed every other kind of transportation: Trains need a sizeable buffer to work well (at least one train's worth of cargo in chests ready to be loaded onto the train), and that buffer is going to obscure production problems just as badly as a buffer in the form of a 5min long conveyor belt.

[Eddab]

The biggest advantage to use train is the throughput as mentioned. One train car can hold 2000 ore, and is about 5 blocks long. The same length transport belt can hold about 7*5=35 ore.

So, if the traincar is moving, on average, as fast as the transport belt, one traincar is equivilent to 57 belts in parallel.

Now there is the cost in materials to consider, and even a tiny train with one engine and one traincar, and a looped traintrack costs next to nothing compared to the same throughput as belts.
Trains and train tracks are a lot more flexible.


That is the math behind it, but as said before, do what feels best. Cost is only a matter of collecting more resources, ie time.


*edit*
Forgot to add more math :P
Lets pretend the train has a 5 minute round trip, from empty at the factory, to full at the mine/mines/ back to empty at the factory.

One seriously long blue belt withh have a throughput of 40 units of ore per second. 5 minutes throughput is 40*60*5=12000 units.
This is equal to one train engine and 6 train cars. If the train can do 120 km/h at top speed, with 2 minutes of travel and 30 seconds load and 30 second unload time, this is a track of close to 4 kilometer in length, and if every piece of belt is 2 m squared, that is 2000 pieces of belt, if it is all on one straight line. Add more belts if it has to go diagonal.

[ExavierMacbeth]

https://wiki.factorio.com/index.php?title=Transport/What_Transport_for_which_case%3F
Math toward the bottom...

Recommended Belt Limit: 500 tiles (used as distance for all maths)
Yellow Belt: 700/min
Blue Belt: 1750/min
Trains: 1500-3000 (Ore vs Plates) per wagon (Speed varies depending on load/unload layout/speed)
Bots: 3000/min requires ~2400 Logistics Bots & 4-5 Roboports

What sorta throughput do you need for your setup :)
The longer the belt the more trains become king.

EDIT: Wow I just realized how inefficent my current base really is. Its only about 4-5 ports long but I pretty much am moving everything except copper/iron plates & ores with the bots :P

[Ultra]

Okay I understand that train is more efficient than belts. due to the fact that the train can hold more items than a belt. But Im not sure I understand why shorter belts are more efficient than longer belts. they both can hold the same amount of throughput.

[ExavierMacbeth]

Mostly it comes down to costs.

At those same 500 tile range (not including Assemblers to make it or inserters to handle loading/unloading on either Belt or train)

Yellow Belt: 1500 Iron Plates
Blue Belt: 10,750 Iron Plates
Train: 2195 Iron Plates, 30 Copper Plate, 250 Rock (Straight Track + 2 engines + 1 Cargo, Dual Engine so it can reverse)

If 700/min throughput of the yellow belt is all you need then your golden. If you need more then the belt price starts to skyrocket those kinds of distances compared to using a train. If you have the Iron fields for it then your choice :)

Edit: And yes I broke the Steel & Engines parts down into their Iron Plates counts for the math.

[brian_va]

at some point the return on investment makes longer belts less attractive, meaning you will have to spend a disproportionate amount of the potential resources to just bring them in. while they are not purely wasted as they can be reused, you will have to use even more belt the farther out you go, meaning even more of a loss from the resources that are being brought in. while that belt can also be repurposed, meaning you can transition from iron to copper once the iron runs out, there will be a period of downtime while the belt storage is allowed to clear before it can be fully switched over.

you will also have to run multiple belts as the game progresses to provide enough ore, so you are going to have to double or triple (or more) the cost of transportation (belt costs).

[Ultra]

Okay thanks for all the answers. I just got one more question which is kind of related to this topic. What about long pipelines. If Im not mistaken liquid travel through pipes instantaneous. Is there a drawback with extremly long pipelines or Is is more efficient to use trains even though there is not liquid tanker int the game yet. Is there a limit of how far liquid can travel through a pipe? I got no idea how pipemechanics works.

[ExavierMacbeth]

Also what Brian said. If you extend that range by another 500 tiles the additional rails for the train only will cost you an extra 1375 iron & 250 Rock... Less if the rail production makes more than 1 per resorce cycle.

So if you get to the point where you start doubling your yellow belt lines or extending the distance the cost gap just keeps growing.

[brian_va]

i dont think it is instant, but i could be wrong on that. hover over a pipe and it will show you when it is pumping liquid through.
i use train for this as well. all my resources go into a station and then allocated where they are needed. the resouces ususally come in from several deposits, so it keeps management a bit cleaner and easier.

[ExavierMacbeth]

For Pipes, there is a little travel time involved.

Over extreme distances (and I easily run 1500 tile pipelines to fuel my "firebases" flamethrowers. I recommed using the Underground Pipes. The entire underground length is treated like 1 pipe segment so they are less prone to Preasure falloff issues (which is why there are small pumps you can build) over long distances.

But I think most pipes handle 30 fluid/sec at full preseure even if they only show storing 10.

[Ultra]

Originally posted by ExavierMacbeth:

Also what Brian said. If you extend that range by another 500 tiles the additional rails for the train only will cost you an extra 1375 iron & 250 Rock... Less if the rail production makes more than 1 per resorce cycle.

So if you get to the point where you start doubling your yellow belt lines or extending the distance the cost gap just keeps growing.

So if I Understand corectly trains get exponentially more efficient than belts the further the distance. due to the fact that cargo wagons have higher throuput than belts as well as a train network is cheaper to build. But you dont want trains for shorter distances because of the onloading and loading aspect

[brian_va]

Originally posted by Ultra:

So if I Understand corectly trains get exponentially more efficient than belts the further the distance. due to the fact that cargo wagons have higher throuput than belts as well as a train network is cheaper to build.

yep

But you dont want trains for shorter distances because of the onloading and loading aspect

debatable, this one isn't so black and white
if you already have an unloading station setup and track is running fairly close by a deposit, the resource cost to drop a loading station is pretty small. while running belt might be just as easy, it is cleaner and usually easier to just have it come by train and get mixed into the existing distro system then integrating more belts into the system.