Conveyors OP?

[thetoolcrafter]

I think trains are underused for a lot of people, simply because conveyors are so useful. They need no power and the only main disadvantage is that they go in one direction. To change this, maybe power could be needed for conveyors? Or, imagine having a "motor conveyor" and a "conveyor". My idea is maybe every 10 conveyors or so, you need a motor conveyor, powered by electricity. It would make conveyors more complex and also more people would use the rails and trains

[Mylon]

Conveyers don't need power but they could use a nerf. Yellows are too good for most cases. Even reds are really good. Blues are just fine.

[Deadly-Bagel]

Ugh, that would just be tedious beyond belief. Perhaps if you only needed one motor per belt but even then it destroys the possibility of having complex little bits of belt running around and merging with other belts. You would also need to introduce a Powered Belt for the little bits where a motor isn't possible, either that or place the motor on the belt itself but again it's just tedious and doesn't really help the gameplay.

Also, if we're talking iron-per-tile rails are cheaper than belts (rails are 5.5 iron + 1 stone = 4 tiles / belts are 3 iron = 2 tiles). Of course making it two way and adding signals and the train make rails more expensive BUT for long-term rails still end up being cheaper.

When you lay a belt to an outpost it can ONLY service that outpost. If you want to add another outpost you need to lay a whole new belt from start to finish, you can't reuse belt paths.

I posted a detailed analysis here.

[HurkWurk]

a better solution would be for belts beyond yellow to use power.
yes, i know this isnt ideal, but it makes more sense and gives people a reason to think about trains for speed rather than long red/blue runs. yellow would still be viable for long haul / low use.

the entire game is based around water pumping itself and conveyors being free.

[Deadly-Bagel]

Why would you use red or especially blue belts over rails? There is zero reason to do this other than complexity, and if that's a problem for you, you're playing the wrong game!

[HurkWurk]

i never use trains. no point. i belt content directly to where it needs to be.

the only value of trains is diagonal movement. but squaring off belts allow me to make city squares anyway, which lays out better for factory/smelting rows.

[Deadly-Bagel]

You've provided no evidence of why there is no point in using trains. Maybe because there isn't any.

Throughput is obvious, belts just can't compete. It is easy for a single unloading station to have eight belt lanes of throughput. If you need more you need only add more carriages.

But surely belts are cheaper? Well, in terms of cost effectiveness, two way rails are only comparable to two yellow belt lanes. Two red belts is far more expensive even when you include the signals and loops, and become ludicrously more expensive the moment you add a second outpost. The train itself is some 280 iron, less than the cost of adding a third belt lane. So trains are actually cheaper than belts.

So as I said, the only reason you would want to use belts for long distance transport is because you're lazy but even that has its limits, to expand a train network you just link to an existing rail but belts you have to run aaaaallllll the way back to base. Have fun with that

[HurkWurk]

you are completely discounting the train station costs in terms of chests, robotics, etc, to store and forward what you are moving.

take a basic case, a distant iron field. you have several choices on what to do:
1. mine the iron and ship it back on belts to be smelted at your "city".
2. mine and smelt the iron, ship back the places on belts to be directly input at your city.
3. do either of the above, but setup a train depot to collect the goods and another to drop them off.

a yellow transport belt will move 13 items per second.
a red will do 26, a blue will do 40.

to equal this early game you need 6 fast inserters. but since you need to buffer for the train (unless you want production stoppages) you need at least 12 plus chests.
thats just to equal it... you actually need to exceed it in order to have the same performance (otherwise the constant belt rate will out perform the train travel time issues)


trains are basically nice for density and large movement of items that are used at an irregular rate.
belts are better for items used at a constant rate. especially if you create an outpost that outputs finished goods rather than raw materials.
I can also see that trains would likely be better on certain map settings. (sparse resources)

[Deadly-Bagel]

Am I?

Let's be thorough. Rails are 5.5 iron per 4 tiles, but we'll go two-way so 5.5 iron per 2 tiles. Rail signals are 6 iron and 1.5 copper each which is needed twice every 30 tiles. So 5.5 * 15 + 7.5 * 2 = 97.5 resource per 30 tiles (assume 1 copper = 1 iron).

The cost per blue belt is off the chart for long distance so we'll compare it to a single red belt lane. Red belts are 11.5 iron per tile, so 345 iron per 30 tiles. That's over 3.5 times the cost of the rails and signals, just that 30 tiles has a difference of 247.5 ore.

As you're only using one red belt I'm only going to unload from one side of the train. That's 12 Stack Inserters, 12 Steel Chests, 12 Fast Inserters and a Train Stop, and two are required, plus the train itself. Again I'm going to assume 1 Iron = 1 Copper. That's ((87 * 12) + (40 * 12) + (12.5 * 12) + 37.5) * 2 + 575 = 3998 ore and 24 Plastic.

The plastic is difficult to factor in but the break-even distance on just the ore is 485 tiles. Once you factor in the loop it's about 500 tiles. Keep in mind this is compared to a single red belt, and this setup will probably happily provide a constant supply of two red belts at least. My last game I was feeding four blue belts from two carriages (unloading both sides) and didn't have any problems, not sure if I was consuming four solid lanes at any time but I was at least getting pretty close. Would easily have been going through four red belts anyway.

Additionally since you can't share belts between resources or outposts, trains become drastically cheaper when you start expanding. Rails themselves can carry a very large number of trains without discrimination of contents, and that unloading setup can supply two blue belts easily with sufficient trains coming in. Double up on inserters and chests to supply four blue belts. To expand to that with belts you'd either need four lanes of blue belts going all the way to your outpost (lmao) or six lanes of red belts, then the break-even point becomes less than 200 tiles, and that's assuming you're not even reusing the rails for other outposts.

[HurkWurk]

i was talking yellow belts actually.
6 fast inserters can fill 1 yellow belt.
3 iron per tile, 90 per 30 tiles. slightly cheaper than rail, but obviously slower.


also i can put mixed goods on belts. it just meens i need filter inserters at the other end, and a lot more calculations on storage required to make sure my production and consumption are similar, but if i only want 2 goods (copper and iron for instance) i can just use one side of the belt for each, or run a second belt.

running another belt is what i normally do because by doing so, i can lay them in such a way that they naturally flow where needed for assemblers.

[Deadly-Bagel]

Oof, I couldn't imagine trying to run a base on a single yellow belt of iron, let alone half that! Well as long as you're having fun I suppose =s

[hoho]

also i can put mixed goods on belts

Same with trains - you can set filters to vagons to only accept certain types of items. Those filters are free as they're just an UI setting.

running another belt is what i normally do

The moment you put down another belt, you *double* the required resources for the infrastructure. A half-decent rail system can handle a LOT of trains sharing same tracks.


Sharing of tracks is what will make trains be overall cheaper than belts. If you only run a single line between mine<->base and never share tracks, of course you will not see much use in trains.


Also, the cost of infrastructure is nearly meaningless in factorio anyway. It's a tiny fraction of the cost of building base, science and rocket(s)

[Deadly-Bagel]

Yellow belts are 375 iron per 500 tiles per lane. It's not a great deal, if you want 2 blue belts worth you need 6 lanes = 2,250 iron, still cheaper than train. The investment is instead a massive consumption of time setting all that up. Even if you include bots it's no better because moving to different outposts you have to uproot most of it to reclaim the resources, not to mention the sheer number of belts you need to carry with you.

Realistically you could spend that time progressing in the game but there's no convincing some people =P

[InsaneFox]

i was talking yellow belts actually.
6 fast inserters can fill 1 yellow belt.
3 iron per tile, 90 per 30 tiles. slightly cheaper than rail, but obviously slower.


also i can put mixed goods on belts. it just meens i need filter inserters at the other end, and a lot more calculations on storage required to make sure my production and consumption are similar, but if i only want 2 goods (copper and iron for instance) i can just use one side of the belt for each, or run a second belt.

running another belt is what i normally do because by doing so, i can lay them in such a way that they naturally flow where needed for assemblers.

Fast inserters? Why not use stack inserters? Yes, they're more expensive, but they can offload a cargo wagon's storage into adjacent chests in a matter of seconds, freeing up the train to keep moving while another row of stack inserters deposits the materials onto a belt for processing and production.

Also, I'm curious as to what scale you've managed to use this strategy with, or if you're even comparing trains to conveyors in a situation where trains actually make sense. Usually in my playthroughs, trains only start to shine in the endgame where I'm looking at launching my second or third rocket, it is entirely possible on normal difficulty settings to launch a rocket or two with the resources gathered from the land you expand your main production facility onto.

Though, the main advantage I've found with trains in my playthroughs is the ability to use locomotives as a means to supply outposts on logistics networks separate from my main factory. While conveyor belts are great at moving raw materials around a single facility or even for moving raw materials from a mining outpost to your main factory, they aren't a viable means of keeping long distance outposts supplied with produced construction materials (walls, turrets, power poles, belts, etc.)

Granted, long distance outposts and supply lines are definitely not necessary to launch rockets, but do provide a quicker means to expand your material intake than you would have by simply expanding your walls to enclose whatever materials you need.

[MalcolmCooks]

I suspect people who find "no use" for trains have their map generation set to make resource patches very close together

[Hannu]

There are so many ways to play and in some ways trains are not needed. You can say the same almost everything in Factorio. There are solutions without belts, without bots, without modules etc. It shows versatility of Factorio and makes it a great game. If you want to play with ways which do not need trains then just do not use trains or make train because you like it even it is not necessary. But do not try to force your way for everybody.

I do not care if belts need electricity. If they needed, bots would put electric poles as easily as long belt lines now. I consider infrastructure stuff as practically free in Factorio. They do not eat many percents of my production in any significant phase of the game and one blue assembler makes so much of any product that I have to never wait. But I do not want to feed 12 blue belts of iron and 8 blue belts of copper from station to my base with nerfed belts. I also do not see wide belt lines from ore deposits kilometers away as a reasonable way to transport hundreds of thousands of item in hour compared to ore trains which can share tracks with each other and all other trains and which are easy to reprogram after changes in ore production.

[Hannu]

you are completely discounting the train station costs in terms of chests, robotics, etc, to store and forward what you are moving.

These costs are completely negligible. If I use 6 filter stack inserters, 6 steel chests and 6 fast inserters per wagon, building of them and wagon and part of the locomotive needs less raw materials than one wagon can hold. Train station has paid itself after the first unload. Costs of infrastructure is nothing in Factorio. I have thought that would it be more interesting or annoying if infrastructure costed significantly (for example 100-1000 times current materials and time) and I had to think what I can afford to expand or in what sequence should I build things. Now I can make a small beginning base (typically capable of handle one red belt of iron) and when I get trains and bot logistics I can begin to build tens of times larger megabase and I have to never think can I do something or wait that single blue assembler makes everything needed.

a yellow transport belt will move 13 items per second.
a red will do 26, a blue will do 40.

to equal this early game you need 6 fast inserters. but since you need to buffer for the train (unless you want production stoppages) you need at least 12 plus chests.
thats just to equal it... you actually need to exceed it in order to have the same performance (otherwise the constant belt rate will out perform the train travel time issues)

There are no real needs for trains in early game. At least with normal resource settings. You can also make the base able to launch rockets with very low amounts of stuff, and if you have good sized deposits near your base you can send first rockets without outposts. My first factory used one red belt of iron when I launched my first rocket. I had two assemblers for light structures and control units. But if you want to have larger base, for example 4 blue lines of iron is quite common, and have settings (or mods) which gives ore patches with large distances, it become soon impractical to make large distance large volume transports with belts.

trains are basically nice for density and large movement of items that are used at an irregular rate.

Or so high rate that belts would be impractical. It is so much easier and more versatile to make rail lines and signal systems than kilometers long 8 belt wide transport buses.

belts are better for items used at a constant rate. especially if you create an outpost that outputs finished goods rather than raw materials.
I can also see that trains would likely be better on certain map settings. (sparse resources)

Have you ever really made a train world like base with belts instead of rails? It is easy to allocate slots for 40 different products in a train and run a loop through most stations in which just needed items are taken or loaded. But you have to make belts from every factory to every other factory (in both directions of course) to get same functionality. Or have you ever tried to transport million items per hour from your iron smeltery to your base? Or have you tried to collect million iron ores from 10 patches around 16 square kilometers. And of course hundreds of thousands of coppers, crude oil etc. A railroad can handle easily 4 8 wagon train per minute (even more is possible, but I want to do "realistic like" rail network and trains instead of Factorio optimal)

[Deadly-Bagel]

I've already given stats and numbers on all this Hannu so you're just repeating me at this point. And actually the train station was surprisingly expensive at almost 2,000 ore each for just single-side loading (still less than a cargo wagon's contents but surprising nonetheless) which makes it difficult to compete with belts in terms of cost.

Cost is irrelevant in most games I agree but ike I said there's just no convincing some people ^^

[Hannu]

I've already given stats and numbers on all this Hannu so you're just repeating me at this point.

You are right. I read this discussion carelessly before answering.