General good practices?

[Purplekief]

I'm a new player that's completely fallen in love with Factorio, but since just finishing my first factory and seeing the absolute mess I made I figured I'd come ask for some input.

I understand the base principles of most items in the game, I still got some research to do on train signals but I'll get there eventually. I'm more looking for the DO's & DON'Ts of Factorio. Let me give an example.

Through my factory I run a 4 lane bus, as I heard was good practice. Something with throughput and stuff. What I did not know is that there was a correct and incorrect way to split from that 4 lane. I used to split from every late seperately using a single splitter and looping my bus to the left/right when it was time to split the middle lanes. (I did them in order to keep them somewhat balanced)

Apparently you should run 4 splitters diagonally acrosss the bus and use that 1 lane left from the splitter. I'd love to hear more about stuff like this, I've already researched a lot but would like to hear from some experienced players.
Best regards.

[impetus maximus]

do play a couple of play throughs and don't watch guides.
do make spaghetti. if you look at the game's trailers, making spaghetti is the correct way to play.

seriously though... play, and try and learn on your own. if you jump right into guides, or using other players layouts/blueprints you are doing yourself a HUGE disservice.
a lot of the magic of the the game is learning on your own, and having those beautiful 'ah HA!' moments.

welcome to the forums.

[Koub]

One of the best things in Factorio is replayability for improvement.
If you get everything the optimal way the first time, what will be left to do ?
I have over 600 hours of Factorio behind me, and I still play like sh*t .
The one advice I have is : think big, start small.

[Selvek]

The 4 lane bus you have heard about isn't necessarily the best or only way to go - it has its disadvantages like any other build. And the " correct" way to split off from it is far from clear cut, and depends on the situation.

Best way to play factorio is blind

[c0bRa]

Mainbusses are nice for beginning and mid-game. Lategame your mainbus will be too small to get all of the amount to the factories. When you really want to use the mainbus in endgame, you have to refill them in the middle. Especially Iron, copper and green circuits are very intense in use. Also (my) train setups are ineffective in lategame.

For balancing MxN Lanes you should print out that picture here (or just make the BPs ):

[Jap2.0]

On the spaghetti vs. main bus and a few other things (note that this is directly copied from one of my old posts, so it might not be entirely relevant, but most of it's still good advice, source.

A few things:

• You don't need a main bus. I cannot stress this enough. Some people prefer to use a main bus design, but it is by no means necessary. It is certainly more organized and some people like a main bus design, but because you don't have a main bus doesn't mean you're doing it wrong. Once I tried to build a base with a main bus and I simply couldn't. I was bored with it before I even had green science fully automated.
• Whether you expand you factory and try to improve it or start over is entirely up to you. If you try, you can probably remedy most of your problems (except probably implementing a main bus now, but again, you don't really need one. If you really want, you can rebuild a large portion of your base when you get bots.), but if you think it would be more fun to start over, start over. The opinion of the forum shouldn't change your mind, you should go with which way will be more fun.
• Yellow is a very advanced science, so you probably want to make blue next. (If by any chance that was a typo, purple and yellow science are on a similar tier, so make whichever one you want more first.) Most technologies (maybe even all) which require yellow also require blue.
• You will always need more of something. Right now you need more iron. If the patches you are mining right now aren't fully tapped, fill them up with mining drills. If they are, then belt more iron in, or use trains. There are several mini-tutorials relating to trains in-game, and if you need further help, ask on the forums. Then you should probably double or triple your smelting. One belts is not enough, and iron usage will increase rapidly. You will have to do the same thing to copper eventually. You will almost always be building more production or needing more supply, so this won't be anywhere near the last time you'll have to do this.
• There are a few ways to increase the supply of iron to your assemblers. You can put another belt (or two) next to your existing belt, or if you don't have enough space, merge the belts later along the line when the belt becomes nearly empty.
• For oil, I recommend having a tileable design (if it's in a blueprint, that's even better) like this:
  
• No matter what, having fun is your main concern.

[quyxkh]

as I heard was good practice

Heh. The Factorio designers have an abiding love for subtle quirks. The game's built to reward an eye for detail. If you don't already have a bagful of designs using the equipment you see these "best practices" videos of, turn away. Play What Can You Do WIth a Shoe? with it. Get inventive. Play with it.

You'll see I'm not the only one saying this.

[The Eriksonn]

I like to make lots of trains with each production in differet places: smelting, Circuits, red/green sci,blue/Purple sci, modules...

I know it is not at all the fastest or most compact and such, but it is fun and i like that

[hoho]

When building stuff, leave PLENTY of space between different parts of the factory unless you like to rip up and redesign your base all the time when you need to increase production of something.

You can have more than one "main bus" - just dedicate some area for some specific things and have specialized production areas feeding it. E.g have a separate production line for green circuits, science and other stuff. Don't be afraid to build them as separate outposts if your base is getting really big.

Automate construction of almost everything. When you have two belts with iron, green circuits and gears you can line them with a bunch of assemblers that build majority of the most commonly used stuff (miners, assemblers, inserters, ...). Being able to just pick up a stack of them greatly speeds up your ability to expand things. Of course, later on with logistic bots it's even easier to set up.

This is my personal preference but I pretty much never bother with exact balancing of belts. If machines earlier in the bus end up consuming things too much so stuff later down the line won't have enough, I just take it as a sign of having to expand resource generation. Of course, if you want to buffer up all sorts of things that won't really work.

[Frightning]

As someone with around 800 hours into the game so far (and far from having done everything I want to do in the game too). I have learned some important principles from many times of starting over and/or running into unexpected difficulties. There are a few general principles that seem to me to separate what I see from veteran players (including myself) from those who are new at the game:

Veteran players lay things out in much more deliberate fashion, that is to say that they are planning ahead, and expecting the inevitable need for certain setups to be expandable to meet increasing demand as the game goes forward. Part of this is knowledge from experience: they know which things are going to experience significant demand and when that demand tends to really kick in in terms of the tech progression of the game. Along the same line, veteran players also know which setups they are eventually going to rebuild in a different form. Generally, there is an (up to) 4 stage process to that, because of how the tech progression works:
-Initial manual setups: Everyone does this at the very beginning, and you don't stay at this stage for long. These are generally reworked into the first automated setups as soon as the relevant tech is researched or enough Iron ore has been mined and smelted to make enough belts to automate transport of materials.
-Belt based setups: Once you have belts, you generally will use them for a very long time, they aren't really replaceable until you research Logistic chests, which is High Tech science, and not something you will have for a long time.
-Train systems: Trains are the best method available for transporting large amounts of materials over large distances, hence most veteran players start moving bulk materials by rail as soon as they need to move them over non-trivial distances (e.g. where running a belt becomes an expensive investment and/or lacks the throughput required). This usually starts with ore, but can easily extend to smelting and some other oft-built components (like gears and green circuits). This also lends itself to modularization (which, in the long term, can be a very wise thing to do if done thoughtfully).
-Logistic Bots: Fussing with belt setups for complex, late game recipes can be annoying at times, and messy. Bots remove all the clutter and inelegance while also boasting superior throughput over short distances, which makes them ideal for moving items around within individual factory blocks and for handling a wide variety of recipes in a simple and efficient manner. They can completely replace belts if so desired, but at the cost of non-trivial energy usage. The compactness and neatness of logibot based systems lend themselves nicely to endgame beaconized setups, which are also spectacularly efficient for making large amounts of complex items at minimum material cost (prod+spd module setups).

A good way to help yourself improve at the game is to think about scalability of various setups and how much of something you might actually need down the road. It's also worth thinking about when to transition setups from one stage to another. For instance, I often make a fairly temporary setup for Processing units as soon as I research the tech so that I can make a few when the need arises (particularly since they aren't hand-craftable), a setup which I will later dismantle when the time comes to build a properly scalable automated setup to meet regular demand for Processing units (e.g. to feed High Tech science and/or tier 2+ module production).

Discovering all the little 'tricks of the trade' to making such expandable setups is part of the fun of the game, imo. (I'm still learning things occasionally, even after 800+ hrs playing the game).