Factorio Thread

Shig

Also the tracks are all one way so multiple trains can use the same tracks at the same time.

That just sounds like you're trying to create a traffic jam. The advantage of two parallel tracks is that incoming/outgoing are separate, so you don't ever run into issues of two trains meeting each other face to face and getting stuck.

That's what I did though. The trains even drive on the left like they should.

Wackity Smackity doo

also I pity anyone who doesn't use
TRAINS
to solve all their problems

Fucking wewlad. Enjoy your tracks getting nom'd by biters occasionally.
Nice train stations though.

...

nah nigga that ain't no problem, as long as you don't build your tracks 5 feet away from their base they won't touch em, and even then if you're producing any polution they skip over everything else for the source of that. They haven't touched my rails once all game for those reasons

also what are those bluish blotches in the south-east, solar panels? I haven't put any down since the devs updated the map output so I haven't seen em yet

there's a mod for that, Bob's Mods I think.

Biters don't touch my tracks but sometimes one gets stuck on a power pole and cuts the power for an entire outpost. Should really look into powering outposts with fluid-wagon shipped steam or something.

Never heard of anyone doing that before, could be an interesting experiment. If there's a lot of forest, building power lines through tree clumps can help eliminate random biter intrusion also.

Interesting idea, you could also ship water depending on which one is more space efficient.

I got curious and tested it out, and it seems water is converted to steam at 1:1 ratio. So shipping steam is definitely better.

nope

fagtorio.dynu.net

Switching to a new map, deathworld with lower resource frequency and increased richness.
Per chance there's demand, I can upload the old map or host it on a different address.

IP: fagtorio.dynu.net
pass: vidya
Mods:
puu.sh/wawgx/645ae2a4a1.rar
For balancing belts
CONTROL + R: rotate the pickup direction clockwise, skipping over the drop direction (SHIFT + R before 0.15)
CONTROL + SHIFT + R: rotate pickup direction anti-clockwise
CONTROL + F: toggle inserting items on near/far side of a transport belt
CONTROL + SHIFT + F: cycle inserting items on left/middle/right of destination - allows choice of lane on belt corners, or belts moving directly away from inserter
Allows long underground belts, they use equivalent amount of belts from your inventory. Same for pipes.

It's not my idea

This shit is infuriating.

oh god, they're coordinating their attacks

Fun fact, biter nests cannot spawn in at a certain distance from man-made objects, so players have often taken to placing "victory poles" down whenever they purge a nest. You can keep them from spawning in that way, or you can just not put turrets all over the fucking place because they will aggro over to them from farther distances otherwise. The only thing with high priority than pollution to a biter is a turret.

I thought that got taken out.

It's a good idea from what I've seen steam doesn't turn into water if stored for a long time so it should be possible to use trains to transport it where it's needed.

How large of a bevel do you guys make in the corners of your walls? Is 20-30 blocks large enough?

I'd probably go for about the size where a laser turret on one side of the bevel can cover a pace beyond the walls of the turret on the other side

Hmm, I guess I'll use the measurements for tier 1 turrets. Bob's adds 4 additional tiers, each expanding the range by 2 blocks. On that note, is there any advantage of using gun turrets, other than not needing power poles (but needing conveyor belts and inserters instead)?

Higher DPS particularily against spitters - their damage with uranium ammo is absurd, but a decent chunk better even on piercing. Easier to research and produce (particularily if you have petroleum shortages since battery mass production kind of eats through that). No drain and no need to build extra power storage to compensate for the spikes. Logistic chest fed is probably better than belt fed in the long term, belts need you to keep a huge buffer while keeping the turret at lower effective ammo levels.

Oh yes, I completely forgot that 0.15 adds depleted uranium rounds. This will be useful, as bob's enemies mod adds biters and spitters with various resists, including up to 95% laser resist. I might mix sniper turrets into the laser lines (base shooting speed 0.5/s, base damage bonus +2400%).

Does anyone else have the same issue? On the one hand, even 16 labs are gonna be fully provided by 4 assemblers of each science (at least the early kinds), especially if you speed module them up, so it's not like you need to expand very much. On the one hand, maybe I would want more labs in the future, in which case I am fucked.

I tend to build the first short term solution that comes to mind. Which means that I tear down my entire base upwards of two times per playthrough.

With how complicated and annoying producing shit like circuits is in my modpack, I am not sure if I have the patience to redesign is more than once-twice per section, individually. Pic related is tier 1 and 1.5 circuits (basic electronic boards require basic circuit boards + loads of components and solder).

I guess it's time to move walls… Those chokepoints are gonna be easy as fuck to defend.

That is a criminal lack of resources.

Playing with Resource Spawner Overhaul. Resources are spread out, and the further away from the spawn, the richer they are.

Either the power poles inside the solar array (that are clearer on the sand and not so much on the grass) or the robports and power poles connecting parts of my base.

It did. If you're not careful, biter nests will (attempt to) spawn right the fuck on your rail lines, and when the biters find this isn't possible because there is something in the way, they will destroy a large section of the rail, and any trains that happen to come along.

Wow! What a lucky spawn. You'll never lack for space in this map.

Where the fuck are all the biter spawns?

This is very early in the game. I cleared out a few spawners when I built the walls.

Even still they're really far apart and small. Even early game for me they're basically a sea of red beyond the borders.

I did increase the spawn area to very large, but that's about all I did. And that doesn't increase the safe zone by THAT much. Otherwise, the settings are default.

If you take out that one nest you can secure those valuable resources to the south.

I have a diagonal train and I want to make a line going south, with each of the lines being one way. I'm just kind of dumb, and I don't know I can effectively attach the vertical two lines to the diagonal lines, and where to put signals/what type of signals to put. Will I have to just make it so the angled section becomes pure horizontal, then return to the angled run? What's an effective way to set this up so that the lines all interact properly, without having any direct X type crossings?

Give Bob's Mods a try sometime. Making a single green circuit will make you cry tears of blood.

Train junctions are already obtuse in this game, so if you're going in blind I'd recommend short pure horizontal stretches for junction simplicity. Railway traffic junctions are the one thing in this game that I don't understand the rules for, so if you want examples of working railway junctions look here imgur.com/a/CxXxd. Honestly if this game gets a Thread General pasta this belongs in it.

I tried to do some doodling with my train thing. Alot of those things show where a rail straight cross, perpendicular to the line and I was trying to avoid that for, reasons? I think this is what I'm going to try, but I really don't know. Trains are hard.

It's called a roundabout, they're pretty simple to do and can easily support 4 directions.

The bottom pair of rails are the wrong way around assuming the colours are the direction. It goes from driving on the right to driving on the left.
Rails aren't that hard. Just remember this simple rule when making intersections: chain signal going into the intersection, rail signal coming out of the intersection.
Chain signals read the next rail signal and show it to the train incoming. If there are multiple possible tracks to go from the intersection the chain signal leads into, it will be able to tell which one is which and stop or go appropriately. Having one before your stations split off and one after where the train stops at the station is a good idea because it means your trains will stop before entering if a train is waiting at the station. Also putting a few rail signals along your tracks will allow trains to queue up along the straits rather than only at the intersections.

Also remember to put the signals on the correct side of the tracks so the trains are going in the directions you want them to (or on both sides if you want it to be a two-way track.

Fixed your shit, fag.

no bully, i was gonna split them

Mind that these green circuits are not vanilla green circuits. They are tier 3 (or 4) circuits in bob's mods. The tier 1 circuits are in this picture:

Do any of you guys have a unified railroad construction philosophy? I usually have 10 tiles between opposing tracks, and the trains go on the right side. The 10 tile distance lets me easily create 3- and 4-way intersections. However, it does take quite a bit of forest cleaning when going through forested areas.

Openttd autism also applies to factorio, without the bridges or tunnels though

Into the trash it goes.

Gorillion of intermediate components galore. It's a shit.

ya that's sounds, fun

i don't see any 90's there. That's why there's 7 layers of inner track spaghetti

I've long since started using buffer tanks and extra steam engines for overloads handling. Current ratio is 2 engines to 1 boiler, I put 3 engines and 2 tanks to 1 boiler. It can handle very serious power drops and I can attach data cables to tanks and make alerts.

A train coming from any cardinal direction immediately makes a 90 degree turn, and then makes another 90 degree turn if it has to exit in cardinal direction, and on top of that, it makes a 90 degrees turn for every 90 degrees of counterclockwise travel in the mesh.

here's a better one

If you don't care to gather wood, just set it on fire.

this kind of thing carries over into nuclear as well because unlike old school boilers nuclear reactors just keep burning fuel even if you dont need it. so if you want to be frugal, hook a circuit up to a tank, and have it send some fuel to your reactors whenever it gets below a certain point.

so serbian on 15.16, anyone got a torrent?

I can get you a copy. I think the latest is 15.18 right now. Which version you need and what OS?

Just lookin' for whatever we're on on the serb
windows 8

Pretty sure that's 15.16.

mega.nz/#!IxJXzKrT!zTRxCwq3Dxizqz6KgXCPRMvZCfnQclbZ6uOZSwr27bY

Poison capsules are better for clearing forests. I usually carry shitloads of the things from as soon as I research them. Pretty sure burning the trees leaves the burnt husks around that you still have to clear afterwards. The ol' deconstruction planner tends to work wonders if you have the inventory space to spare for the piles of wood it'll bring in though.

Why the fuck do we have blue science but no trains

My beef with Bob's is that it doesn't really make the game "harder", just more tedious. It's less like having a bigger set of Legos, and more like requiring you to first assemble your Legos from several iterations of smaller parts before you can start building stuff with them.

Yes, I have all my trains leading to a giant pit in the ground where it drags all biters to Hell.

In all seriousness, it's kinda unified; to be honest I pretty much only standardized it in my last base. I can't remember the number of tiles between my rails, but I suspect it's two because I stuff power poles and turrets between the rails, and I seem to remember needing to space them out to allow room for signals. Anyway, all trains go on the right side; all intersections are counter-clockwise roundabouts (though sometimes, if a lot more trains are going in a certain direction than the others, I'll add a bypass on the side of the roundabout); all trains are one engine and two wagons. The main sections of the network are double rails, and spurs heading to a single ore patch are singles. Only one train services each ore patch if I can help it, because setting up two-way traffic on a single rail line is a pain. Each wagon is loaded or unloaded by six inserters that go into an iron chest, then another inserter, then the belt system. I had a standard for wait times, too, but I forgot what it is.

Toward the very end I had a standard for the concreting and hazard concreting around the stations, too. Looked pretty cool.

We had all resources nearby for the time being so it wasn't too important. If there's anyone playing it's usually around that 1600GMT by the looks of it

I wonder if it's possible to send a signal when turrets start firing so I can have a big sign that says "FUCK OFF" light up every time the biters come knocking at my walls.

With gun turrets you can just trigger it when the ammo inserters start moving.

Connect your laser turrets to your normal network not directly but passing through an accumulator, then you can check if its charge drops below 100%. Not a perfect method since any normal power spike will also trigger it and it won't work in the first place if you rely on solar/accumulators. Maybe you can countercheck with a second accumulator embedded in the main network and not trigger the sign if that accumulator is also below 100%.

The alternative is that you power your defenses with a completely separate grid (using nuclear or steam engines), so you can read power spikes centrally in some accumulator or steam storage tank.

If it's early on, I just research 3 grenade damage upgrades (which pushes grenade damage to over 50), and bomb trees. Coal is usually pretty abundant early on, so is iron. Later on, I use a personal roboport and deconstruct.

Speaking of steam engines, has anyone tried the "transport wagons fulla steam" system for remote locations yet?

The pumps need power to get the steam out of the wagons and into the engines. This means that once stopped, your outpost will need manual restarting and to start at all, you'll need some initial power anyway. And this means you'll need a boiler, fuel and, potentially even worse, a water pump. Sure it sounds cool, but the disadvantages are severe and I guess you wouldn't gain much.

All you'd need is 1 boiler and any fuel. Just stick it onto one of the steam pipes until the pumps get electricity, which should be almost immediately, then you can remove it. You could even include it in the train stop blueprint or whatever, and grab a piece of fuel from the locomotive if you forgot.

What I'm interested of is the efficiency, how long will a fluid train full of steam last in terms of power? How large of a mining outpost is it viable for, assuming you're using laser turrets to defend it? How many tanks do you need in the outpost itself to hold the steam until a next load can arrive?

Actually a boiler won't generate steam without water, will it. You can't unload barrels without power either.

You can just plop down a solar panel or two to get the pumps going though.

1 accumulator, 1 arithmetic combinator, 1 power switch and a few wires is all you need. Just make the power plant large enough that production is always higher than consumption, and wire it up so that power flows to the system if there's a steam supply train at the station or there is more than 50% charge in the accumulator. When there is enough steam in the system, the accumulator stays topped off at 100%, and once the steam runs out, the charge dips below 50%, at which point the power is cut until refueling arrives. What is tougher is determining when a station actually needs more steam, unfortunately, I don't think it's feasible to achieve without wiring a circuit line anyway, so the point of wireless bases is lost.

1 unit of fluid is 30 kj IIRC. So a 4-car train arriving every 5 minutes would be 30 MW.

You'll probably need to feed it constantly, so I don't think any kind of logic wiring would be useful anyway.

Ideally you should attach steam tanks to the back of your mining trains, and re-load them with steam when you're unloading the ores. Trains don't seem to have a clear limit on how many wagons/tanks they can have, I tried a while ago and I drew around with a train that had like 15 cargo wagons.

One mining operation (for iron, copper and coal) takes 90kW * (1/0.525)s ≈ 171kJ. One fast inserter operation to transfer the ore (let's take two fast inserters, for a buffer chest): 2*46kW * 0.417s ≈ 38kJ. Total of 209 kJ, rounded up to 210. So you need 7 units of steam per unit of ore mined. I hope I am not wrong on these calculations. Anyway, one cargo wagon is 8000 units of ore, and one fluid wagon is 75000 units of fluid. Theoretically, a train with an equal number of cargo and fluid wagons should be able to maintain a self-sustaining remote facility.

i wonder if it makes more sense to ship fuel and water

Water gets turned to steam 1:1 so no, but you could ship water to a coal outpost (for no real gain).

Then again you could run that entirely without power if you wanted. Burner stack inserter when?

I'd assume you have your outposts fully loaded up on E1 modules, so the numbers are a lot kinder.

After trying alot of stuff, I just started using two-way extremely well-signaled rail systems with bypasses. Turns out, my factories never are THAT much large to warrant the level of fuckery Ive seen in the average base.

You could set it up so if the power is off, a burner inserter just yanks a single stack of fuel from the train engine to kick-start the system.

does upgrading hands holding capacity make them more energy efficient when moving multiple items at a time?

In the sense that it takes fewer movements to move the same number of items yeah.

...

...

That can be fixed with a solar panel and a accumulator that are connected only to the pumps.

I was kind of hoping degenerate mods like that would never make it to factorio. How disgusting.

There's a mod for diesel generators and stuff that doesn't need water so you could use that with burner inserters to unload coal, get the generator running and power the pumps to get steam/water.
Or just not run out of power in the first place.

...

I remember some user saying a while back that the inserters have really strange power consumption levels at different points of its work. Something about how when it is grabbing stuff it uses more power than when it is swinging its arm and then the least amount of power (but not 0) while idle.

how hard of a hit to FPS do these massive factories cause?

I haven't built one where i notice any significant performance drop yet. Everything I read says it won't be a GPU bottleneck, it'll be ram speed

Honestly I have only noticed lag around my electric pole/laser turret blueprints after I've placed them and they've been built. No idea what would cause that but it's probably my computer shitting itself more than anything. Millions of items on belts moving around and being picked up by inserters through my entire factory and there's not a single bit of lag, stuttering, or fps drops. It's probably some of that fucking slavic coding magic because this game is so well optimised.

it's pretty impressive. it makes me wonder if it could be done in opencl or something on the gpu, but it would be pretty insane and nobody would use it

if i'm going to spend money on a game i think factorio deserves it, i don't see any sjw faggotry going on and the coding required to make this and make it run as fast as it does is impressive.

They've really optimized the FACTORY elements to reduce lag (stuff like making items on belts maintain the same values until/unless they hit something or change belts; so the vast majority of the items in motion aren't actually being actively calculated at any one time; and also changing the way trains calculate the path to their destination), but you'll still take major hits on your FPS due to biters if there's a lot of them.

As I understand it, if a biter gets into loading range and then you exit the area, the biter in question is stuck in a weird limbo where they don't move but are still being "checked" by the game constantly to see if they can move and how they were previously moving is stored in active memory. So all the biters you see on your map are constantly having an impact on those calculations. If you run at full speed through heavy biter areas and get many thousands of them chasing you, only to get stuck out of loading distance, you can turn your game into a slideshow.

The developers actually said that they initially wanted to do some good software, not a videogame, but it somehow ended up turning into factorio.

A lot of indie games seem to start out like that. Kerbal Space Program was a software toy the devs made up to test some physics simulator they wrote. I'd imagine Dwarf Fortress was just a personal project for a long while.

Orphan Finder mod.

At least in 0.15, they've optimized belts significantly. Article related.

steamcommunity.com/games/427520/announcements/detail/514936124362888985

I thought that got delayed to 0.16?

How goes the rails?

Anyone else having issues with long reach? It's set to 125, but when I load a save it randomly no longer works and I'm unable to get it working for that save again.

...

Are there any mods that make the game more PvP focused by adding more weaponry, artillery, etc? Just building the factory without a real set goal gets boring, wish there could be team vs team games.

Unless it's in a chest somewhere.

Someone was apparently trying to connect to earlier and got mod mismatch errors, you have to download the mods in the link and go into mods settings in Factorio and enable only the the mods included in the rar, all other mods have to be disabled

make sure you keep the "base" mod enabled, it's literally the game. factorio will warn you if you try to disable it.

Is there a limit to map expansion?

Reminds me of hyper efficient SupCom MassFab/Power Station Installations.

So when do biters become amphibious?

Never. That would basically break the game at this point. Also, it would be a major change, and I'm pretty sure the devs have decided there won't be any more major changes.

can mods do this? it doesn't seem that far out to have ships and aerial vehicles, and associated sea and air biters.

you already have flying drones, seems like the code for air is already there anyway

Yeah, the 32-bit integer overflow. But they say it's like dozen of hours of travel, even with max train speed.

Literally just remove the hitbox. It's not that much of a "flight mechanic"

just started playing
reminds me of starcraft if it was production focused, and i didn't expect the first person top down by the looks of screen panning trailers
if any base building game can straighten out my shit planning/design skills it's this, **that or low iq is forever*

I'm launching a dedicated peaceworld for building fun or general whatever, if you have any map gen requests or a good seed I havent launched the fresh server yet. ETA about 5 mins while i configure the .jsons

Alright, for anyone who cares

Server name is "Peaceworld" in the server browser
pass is: ayylmao
pic related is the mods

Server is a headless dedicated, so no worries about it going down, feel free to dig in.
No biters, only building.

forgot to mention i turned off authentication so it /should/ be pirate friendly.

How am I supposed to purge without ayys?

I just played for a while. I didn't do all that much since it was empty, but I'll be back if you keep it running. Coal production is up and there's a bootstrap steel production area that needs to be torn down quick.

OpenCL would only be useful if you wanted to support billions of bots, as few other things in the game would fit that model. A bot has an arrival time and the need to do potentially many very simple independent updates of its position with constant velocity. Those updates can be done on the GPU and the arrival time kept in some ordered datastructure manged on the CPU.

Next time also name it "Peaceworld" but make sure to crank up natives all the way up to Starship Troopers.

i was reading through their forums and it seems like they've discussed this at length they're. It boiled down to:

1) OpenCL / CUDA is underdeveloped and a pain in the ass to work with. Writing the code and testing it would take forever and the time is better spent optimizing for multiple cores.

2) CPU/ram -> GPU/ram lag time is huge, this would slow down the game anyway. There isn't enough shit to send down the pipe to make it worth it sending it every frame.

3) Data Structures don't match and factorio would have to change all kinds of things to conform to OpenCL / CUDA, and OpenCL doesn't match CUDA either, you'd have to do both.

I'm going to look around for this discussion about Dwarf Fortress, I'm sure it's been had and FPS death is a bigger complaint there I think.

OpenCL works on nVidia cards that's one and CUDA is shit anyway that's two.

It's not underdeveloped, it's just that you can't throw what is essentially a traditional supercomputer at most problems and expect improvement. You have to have a problem that fits, an approach to solving it that fits, and a lot of skill. It's why I suggested scaling bots - that would fit, but I don't think supporting a billion bots would improve Factorio. In any case, there are vector extensions in modern processors that I guarantee you go completely unused outside of core system libraries in virtually every program and would be a better place to start if you really wanted to try scaling bots as they are far less limited.

No, the problem is that unlike CPUs, GPUs are massively vectorized processors. That means that it only runs efficiently if all cores in a cluster can execute the same operation simultaneously. If they have to execute different operations, only those cores that need this operation will run and the rest will have to wait until that completes. In bad case scenarios that winds up with worse performance than on CPU.

I'm familiar with the issues, I used to work at SDSC in the '90s although now I'm in networking.

For the bot example I gave, a straight-forward approach would need to keep the source and destination coordinates of a bot's chosen path together in arrays so that you can use vector multiplies and adds on the CPU to get the current position per tick. The vast majority of time spent calculating bot related stuff will be these per-tick calculations of position, some infrequent decision-making on arrival/departure/out-of-power that will happen at a known time in the future so can be separately scheduled in a priority queue of some kind, and the renderer/biter/player checking for nearby bots.

You can impress your friends and do that last part /four dimensionally/ by doing intersection tests on the line segments rather than the position as a filter to see if at any time the bot might be relevant and worth checking the position of. Since you only need to update this filter data structure when a new path is chosen rather than per tick, you can use trendy math things that make you look like you're worth 300k starting like Hilbert curves rather than grind the living fuck out of something like a quad-tree which would need updated per tick for every bot's position even if nothing was interested in it. If you're really being fancy, you can also use this filter to decide what bots are worth calculating the position of at all, as bots nothing could interact with don't need a position.

For the OpenCL version replacing the vector math part in the first paragraph (leaving the other two parts as-is), the limiting factor is transferring arrays back and forth from the GPU and your tolerance of code that is not easy. To make it scale, since very little math is being done just calculating the position of bots with constant velocity on a line segment, the array transfers have to scale better than the operations you're performing so eventually at some number of bots it's going to be faster. The straight-forward way is to scale the source arrays by breaking them into blocks, always add new bot paths to the lowest id block that has free space, and not updating blocks on the GPU that have only had bot paths removed from them until the blocks contain no paths. It's perfectly normal to do math on missing values and ignore the result although this bothers a lot of programmers when they run into it, it's just a cost associated with fragmentation. This gets source arrays scaling with the rate of new bot paths rather than total paths.

Not sure how would you use hilbert curves for that application.

Cache friendly update order of quadtree, maybe?

They're used in Hilbert's dynamic R-trees for doing intersection tests on line segments.

The hilbert curve is used for clustering of R-tree, what does it have to do with line segment intersection test?

The cache-friendly way is first to last memory address wise.

putting the free in freeplay

So I got my basic factory going and now I'm starting ripping at huge fuckoff smelting facility. I'll upgrade belts to blue later.

fugg

Also expected throughput is 1.5 million iron plates per hour on red belts and 2.28 million iron plates per hour on blue belts. There's a mirror of this facility on top, too, it'll do copper.

Bretty gud famalam

Maybe I should have beaconed.

At least it looks pretty neat on the outside.

Your ore is not even going halfway through the complex, what are you talking about beaconed?

They already updated since you last patched, so we're incompatible.

Holy shit.

Can we see the rest of your factory?

how is it pirate friendly if you don't give a hostname?

That's because the mines are running empty. There's 70 furnaces in each row, which should process exactly 80 ore per second, but the trains are arriving irregularly now so they aren't getting fed properly.

Here's a map. The whole thing is basically a roundabout. Refinery is up north, nuclear stuff east and science south. Main bus for actually making stuff is north of the coal patch, and the complex that makes circuits is east of that.
Rockets are made in the center. I thought the complex would be huge but beacons basically brought it down to nothing so now there's a ton of empty space.

Nigger blue belt's throughput is 40/s that's half of your expected capacity, you will need 2 ore belts and 2 output belts per each row of furnaces.

I have 2 input belts for each row. Except steel since that needs a furnace to make iron and then another one for steel.

Very impressive. Very efficient. Perhaps a little TOO efficient.

Broke ass thread, come on.

Got the most recent torrent? Already bought the game but on a new computer after my old ones hard drive went to shit and don't feel like it right now.

Anyone hosting a server? Possibly with the robot army mod?

i bet u nubs don't even build computers
github.com/justarandomgeek/factorio-computer

...

Kinda boring, tbh.

ya i don't know, it's cool that he did it but it's someone with clearly too much time on their hands.

I still haven't messed with any circuits, i read about them, but I can't really think of anything I'd use them for at the moment.

Turn factories or inserters on/off? An empty hole on the belt would do that. I could use them to auto switch between solar/accumulator and steam generators, but doesn't that already happen too?

It seems like a really cool thing, I just can't of anything. I haven't gotten late game yet, but it seems like the most useful thing would be to balance oil products I'm not using.

This shit is only ever impressive to people with zero knowledge of computing and circuitry. 99% of the groundwork involved has been completed by other people, so the majority of these kind of projects essentially boil down to just following, step-by-step, other people's instructions.

You could print the wiring diagram and complete something like this in less than an hour on a breadboard, but because it was assembled in a slow and clunky manner using a video game I guess it's suddenly some sort of amazing feat? Honestly, I'd say it's more impressive to see someone produce the resources necessary to do this.

i noticed that, when i was looking at this, it's not like anyone invented their own, they're rebuilding old chips

It's overly complicated, but it feels right.

Trainworld is comfy, since expanding to new resources is always an adventure that requires thought and preparation, and you get extremely large forestfires.

Am I the only autist who names certain locations? I am sure the Xenos got their own names for them, so I like to plaster human names all over the map.

Just take it from the Xenos.

You can also get a combat shotgun and just blast your way through the things. Also has the advantage of not leaving black stumps behind and not spreading accidentally.

i haven't tried trainworld yet. I was trying to do a runthrough on default but it seems so damn easy it's cheating, so back to deathworld

You can ramp up the amount of enemies and resuources required anyways. Trainworld just makes expansion rather difficult, since you will have to
A) Fight through hordes of xenos beasts to get to the next resource patch
B) Build an elaborate train system to keep your outposts supplied with energy and ammo, while also getting the raw resources to your main smelters/assemblers.
The only thing trainworld makes easier is the fact that lakes are less common and can be quite massive once you find one. This means that building around them or using them as a natural barrier is easier.

Trainworld also disables xenos building new bases so expansion is a lot less of a hassle than you painted it as.

what gets me is don't xeno's attack rail and power lines? I haven't ran rail in deathworld yet, but just to expand to a nearby resource i pretty much have to build a wall of turrets/walls all the way there.

You can always just set the resource density and patch size to the same value as in trainworld, you know?

The more evolved ones do iirc, I think the ones below 0.5 factor don't

I always start to go cross-eyed when I'm wrangling with building a blue economy. I've done it, but it's always a massive clusterfuck. I got two questions for the turbo-autists here:
1. Is there a kind of sorter that auto-filters a belt based on input? For example, a belt with all kinds of random shit on it, and I tell it to move all plates to the left side and all gears to the right side - and then another one right after that further sorts things out so you can filter trash off of your belts?
I know I can do this with selective grabbers, but belts+splitters looks neater.

2. Is it possible to fit an entire blue factory, from raw resource to final product, in a single non-zoomed-out screen? It doesn't have to produce a ton, as long as it doesn't starve out.
Playing this game has given me an OCD need to make things as compact and minimalist as absolutely possible, so I'm kind of thirsty for ideas here.

you could try something like this
wiki.factorio.com/Tutorial:Circuit-network_Cookbook#Sushi_Belts

You fucking degenerate piece of shit should be hung.
To answer your questions though:
Yes, vanilla blue economy is surprisingly easy. Pic related. Blue arrows are the "raw" input (I count copper and iron plates as raw), and red arrows are the output. Input generally "flows" from right to left, the output goes from left to right.

Any servers up?

Post your spaghetti.

i've been doing this same thing, i've been thinking about breaking out coils and gears though and running them through the bus. it seems like it would cut down on the number of factories needed, in deathworld / expensive recipies it gets stupid.

same thing with steel. it's not used as often (for as far down the tech tree as i've gone), but either steel goes down the bus or coal to make it does.

1. Never mix 2 or more things in a single side of a single belt. One belt should only carry a max of 2 items; one on each side. You can manage sides easily without filter inserters.
2. Don't be afraid to move items long distances with belts. It's much easier when you have plenty of space to route your items.
3. Add all circuits to your main bus, since they're needed for many things but usually a pain to craft locally. (green sucks up too many resources, red requires plastic, etc)

Just put coal down the bus. Coal is more versatile, since it can be used to make plastic, grenades, and a whole lot of other stuff.

I agree on everything but the green circuits. They may be needed for a lot of things, but some user in the previous thread told me a neat little trick to make them locally.
Basically, you just plop down one green circuits assembler and have two copper wire assemblers feed DIRECTLY into it. Then just run an iron plate line past on one side, and a copper plate on on the other and you are done. You can output the green ones using either long inserters, or have the iron plates go along on the same band as the copper. You can see it in .

There's 2 problems with crafting green circuits locally.

The first is that it takes 2.5 items to craft 1 circuit, or 7 for 1 with expensive recipes, plus it takes both iron and copper every time. You'll need way more extra belts carrying iron and copper than if you just had a belt for green circuits.

The second problem is that they're needed in such massive amounts everywhere that you probably need equally much iron and copper for green circuits as for everything else combined.

Huh, I got very few green circuits assemblers actually. You can easily feed something like ten labs on blue, red and green science with just around ten or so green circuits producers. I am having more trouble feeding enough steel to my engine assemblers though, since I have too small of an iron to smelter throughput.

So how would a pirate get mods ?
Now that the modpage ist access restricted

Got back to it after long time.
Do these conveyor tracks still help vs spitters in recent version?
I placed them but they still remove my walls sometimes.

Also game seem to became more demanding or I lost some of my autism, dont even feel like moving past green tech. Also did those 'military" potions even exist? Just did all green stuff and started dicking around with some trains out of boredom.
That dude still needs no food, some agrocultural bots or canned alien factory would be nice.

Now I think whole idea of these science-potions is retarded. I think it's because this whole industry server no purpose, so these factories serve as artifical demand for produced goods. I guess whole survival setting was a mistake.

Do people find it easier to produce steel on site or sending through the bus? I've started doing it on site after finding the timing matches up well with some other needed recipes, and I don't like how a metric fuckton of iron gets reserved when the steel on the bus gets backed up

i didn't know the mod page is access restricted, that blows

I'm just giving them my 20 sheckles at some point, i think it's worth it

I also think it's well worth paying for, but I can't support an early access game out of principle at this point.

As long as it is your goal to build a factory you require to do research.

Never_pay_more_than_20_dollars_for_a_video_game.jpg

It's pretty much a complete game tbh. They deliberately kept it off steam for years with a zero bug tolerance policy. They are pretty damned based, nevertheless, keeping mods behind a paywall is weak.

Even if the mods weren't behind a paywall it would be difficult to get them going off a pirated copy with all the frequent updates. From reading the forums it seems like these seemingly small updates break mods frequently.

Seems like you completely missed the point.

I don't think I had to link my throwaway factorio account to my game to download mods, it's bullshit that you need an account to download mods though.

Not wanting to automate shit in Factorio is like not wanting to shoot shit in FPS.

play deathworld and if that isn't enough ramp up biter difficulty, and its' definitely survival. the xenu's are the reason for the industry. i don't get why people play these peaceworlds or with biters completely off, then it turns it into a total sandbox with no point. with biters sufficiently difficult the point is clear, if you don't get a factory up and get more advanced then the biters will push your shit in.

also this five lane bus with coils gears and coal is turning into total spaghetti

This thing will turn into massive, irredeemable spaghetti as soon as you need to expand iron throughput. Actually why the fuck are wires on the bus when green circuits aren't? Little reason to bus wires for anything when you can just locally manufacture 'em.

Don't know how long you've been gone but at one point they changed biter behavior to encourage targeting belts themselves.
And yeah, military potions are 0.15 stuff.

i dunno with expensive recipes i'd rather have the 40 copper cable factories i'd end up needing in the same place. i can locally manufacture the circuits. it seems like if i do copper cables in place i'll end up with more factories sitting there not doing anything when it backs up.

How physician can one get?

I thought this world was going to be cooler.

What is the best expensive recipes green circuit setup?

This is the best I came up with. Copper on the outside so I can run more belts there if needed.

that actually looks pretty awesome.

Nope, looks pretty good to me.

can you do a 1 way stop? i keep getting a path error. do I have to loop it?

It either needs a loop or an engine on the other side. Trains won't go backwards unless you drive them manually.
As far as I know it's always best to only put engines on one side of a train, because having engines on both sides drastically reduces their top speed.

this
You don't have space for a 1 way stop in that because you need 2 locomotives, and your train stop is too far down.

Try something like this.

Not necessarily the top speed, but the engine on the one side has to pull the inactive engine on the other side so you sacrifice either some cargo wagons or speed & acceleration.

It's not like the game makes you.

How long does it take to fill that thing up?
actually, anyone got some more efficient method than just 5 fast inserters on each side feeding into boxes, which feed into the wagons?

thanks, that whole deathworld game was an abortion anyway, next game i'll do it right.
I'm not sure if outposts are even viable in deathworld that outpost was getting hit pretty fucking hard, xeno's dropped 3 spawners right outside of it. I've read expanding your main wall might be better, but there's now way in hell that's happening easily once the game's been going a while. Clearing out one base is hard enough with with turret creep and grenades, 20 of them would be really bad.

that's pretty cool, I assumed there was a cap.

is there actually a cap on anything? (beside your own PC)

I don't think so even the multiplayer doesn't have max player cap.

...

I dont think I have ever seen anything more autistic

You sure?

SIX fast inserters per side feeding into boxes, with stack inserters feeding the wagons. Also make sure to split the throughput somewhat evenly so there isn't one inserter hogging half the product if production goes slow.

I recall reading somewhere that a backwards-facing engine counts as two cargo wagons of weight, although I don't see it on the wiki.

Your train stop in that picture doesn't have enough space for the train. Make sure you look at the white boxes and yellow arrows that appear on the tracks next to your train stops when you're placing them to tell you where the train and its carriages will go and what direction it will be facing.

I have a taxi train that doesn't have any carriages that almost reaches 300km/h. It's fucking insane. My other trains only ever have two carriages.

Just tested it. Turns out it weighs slightly less than two but more than one.

Using different fuels also changes acceleration.
It goes Wood

When does this game launch and is there additional content planned after launch?

I don't want to get it if I can enjoy a more "complete" experience on a first playthrough at a later date

i haven't read about any more major features planned. apparently there's going to be some artillery wagon for the train, and a spider bot or something. the art / sprites are also getting redone, higher res or something. They're also working on better performance.

i think any other real major "features" can be done by mods. the modding capabilities with this are extensive.

Short answer, yes.

Next, try 5 engines all running rocket fuel.

Trains have a speed cap I believe so multiple engines mainly help for long trains.

Looks like it does. The trains move faster with more engines but their speed is the same so long as the ratio of engines to carriages is the same.

There is a speed limit of 298.1km/h that you'll never break no matter how many engines you have though. One train with rocket fuel can reach this speed just like four trains with rocket fuel. Acceleration is faster with more carriages though two engines without carriages is just as quick to reach its top speed as four engines.

...

that's good to know actually, had no idea how that worked

So it's like 90% complete? Is it worth getting right now then or should I wait. What would you do in my situation knowing what you know about the game?

I generally like waiting a year+ after release so that additional content is added and/or bugs are ironed out.

Ideally I'll get the ultimate or GOTY edition. I don't like coming back to games that get new content added, I prefer to play through something thoroughly and then move on.

If you're a patient person then waiting until release may not be a bad idea.

Factorio is a very tightly designed game, everything has a very definitive place in the game and nothing really feels like extra padding (except maybe sound players and concrete).

That said Factorio was originally supposed to release in 2014 or something, so the 2017 release date may not be accurate.

They watch the forums closely and fix reported bugs quickly. Also, they'll probably change the endgame again before release but it looks like that's nearly where it's going to be.

If you're interested just get it now because it takes a while to get into and you may not want to play again immediately since one game lasts so long. By the time you try a second time there will probably be another big patch. Anyway, when I first started there were no trains, so it was exciting to come back and try them out, and the quest to get a small factory started up went much more quickly since I knew what I was doing. I mean, unless you watch tons of tutorials your first run is going to be a wreck, so you may as well get some practice.

But this is solid advice too.

Neat. Thank you trainanon.

More like 99%. The problem they will face with the release version is that the game has no definitive ending so far. You can infini-research mining efficiency, but as soon as the techtree is finished there is no real reason to continue expanding, and the game kind of fizzles out being fun in an anti-climax. I guess they will do something to change this, make a final tech that says "leave the planet" or "interplanetary travel", which could start new game+, where you keep what's in your inventroy but nothing else. Dunno though. The mid game lasts for days and is incredibly fun.

What could new game+ give you that you can't get just by making a new map though?

One of the most fun things to me is the early game where you try to scrap together enough basic materials to automate basic materials.

They were at some point talking about stuff like letting players build and design the actual rocket you launch and then either having a space phase of the game, or having some trading options. No idea if they still have plans like that, probably not.

wat
i thought .15 meant 15%

Nah, just means it's the 15th major version.

Trading sounds neat. You could launch rockets and sell some of your produce to a whole simulated (possibly) online market in exchange for resources or other materials. Launching more and heavyer equipment would make the rocket insanely huge and thus make you more reliant to build large shit yourself while advanced stuff you don't need a ton of (like blue circuits) be imported because you don't want to bother with that crap.

The issue is that the "endgame" has always been kept accessible for the more casual players. But by the time the Rocket Defense of Rocket Silo or whatever else the implement has been built, the really hardcore players have these GIANT AUTISM FACTORIES and nothing to do with them. Launching the rocket is supposed to be the end of the game. The Orbital Space Laser melts all the biters planetwide: you win! Game over. But that's just not good enough for some players.

Basically, you've got a game that's about building a car so you can drive across a bridge. Except some people have instead build a car that can challenge the land-speed-records, and then complain there is no racetrack to test it on.

it's the same fate as dwarf fortress or rimworld. what's dwarf fortress's end game? fps death, same as factorio. the challenge is supposed to be the biters, but the people building these massive autism factories turn them off because "muh UPS".

I don't know how the biters solve the end game problem. They could get incrementally more difficult, requiring more and more resources to defend your autism fort, requiring you to continue to expand for more resources or wipe all the biters off the huge map.

i don't know, i haven't come anywhere close to end game, this is why i haven't really played anything else but deathworld.

maybe some kind of leveling system for bonus's to everything, and then scaling difficulty settings to go along with it.

maybe some kind of system domination or something. dwarf fortress and rimworld both have worlds on top of the maps your in. the rocket becomes a 1 way trip to another planet or something in the system, have a bunch of planets in the universe or system or w/e it is, and "winning" becomes conquering all those planets instead of the map your on

Wouldn't be any different than getting in a car and driving off to somewhere far from your base. There still wouldn't be any reason to build more shit then because it would essentially be building the same thing over. I don't see how they would implement an endgame to this game because it is a sandbox game without a story. The main attraction is the sandbox mode, not the campaign. I quite liked the story mode when I last played it years ago. Not sure if it has been updated now that you can launch rockets instead of just building the placeholder rocket defense tile.

The space layer of the map was considered a while back. They had a picture of it somewhere and it basically went as far as build a platform in space and then continue doing factory things but without resources to mine. I get the feeling he scrapped it because he couldn't think of a way to make it different enough to what is on the ground layer while also having it similar enough that it fits with the game. I was thinking something like an orbital ship yard that you had to deliver parts to with your rocket launches but it has the same problem: these are nothing items that don't do anything except be built.

I guess what I'm saying is that objectives are nice but it's a sandbox game so its strength isn't in them.

maybe they should turn this into an always online game and then you could earn factorio coins every game for launching the rocket, etc. you could then use the factorio coins to purchase DLC, or you could buy them separately if you don't want to grind. they could even offer the mod developers a 3% cut to have their mods on the DLC shop.

...

Best thing you can do is just make a campaign with strong scenario design.

Given the ease of the editor and the general competence of the community, I'm surprised player-made custom scenarios, in the style of the campaign missions, aren't extremely popular.