Advice for chronic refactoring/rebuilding?

[Expunge]

I've played probably 10 or so maps in factorio now, all of which I've deleted because I feel like my designs have gotten away from me. The furthest I've made it so far is blue science at 60spm. I have a hard time building on top of things that I don't fully understand (re-did my oil processing so many times before I was satisfied that I understand it) so I tent to rebuild/restart often. I really want to launch a rocket, and maybe someday do a 1kspm base or something, but I feel like I haven't found a design that I like (extensible, efficient, aesthetically pleasing)

I enjoy experiencing the game as the developer intended (balanced, stable, as straightforward as factorio can be) and I want to be able to earn achievements legitimately, so I want to avoid using mods/"cheaty" map generation settings (I do turn off cliffs though)

I seriously spent an hour last night trying to generate a map that I felt compelled to play and ended up just getting frustrated and going to bed.

Any ideas/inspiring designs that you guys could share? Thanks!

[eradicator]

Read FFF 236.
Realize that the game was meant to be chaotic.
Never start a new map again.
Start building modular outposts connected with trains.
Don't like your current $green_circuit_outpost anymore? Just build another "nicer" one 200 meters north of the old one.

Restarting a map only makes you lose progress, doing all the things again you already did with no sense of "getting ahead". All megabases "grow" over time and take hundrets of hours to complete. Most first build a small "ugly" base to construct the buildings, and then build the megabase from scratch somewhere else on the map.

The more you play, the better you design, the more ugly are *all* of your old designs. By restarting you're essentially punishing yourself for getting better.

The urge to "start from scratch" is something we all have to fight. Don't lose to it!

[Expunge]

FFF 236 was an interesting read. As for never starting again, I have been trying really hard to do that for the last two games or so. Planned out my map in creative mode as far as I could. Made it to bots for the first time and it quickly became obvious that the design wasn't compatable with bots whatsoever. Tried to fix it with personal bots and got exasperated when I ran out of room/got confused and then I nuked it. It's hard to plan a base around a concept that you have no idea how to utilize yet (I have yet to make a single train even..) and it's frustrating when things are a confusing mess. Sometimes these little problems can take hours to figure out if you haven't figured them out before. I guess nuking the base from orbit every so often isn't exactly time efficient either..

[Deadlock989]

Experimenting in creative mode is good advice. I've only made one "mega base" that I consider truly successful and admirable, and it was completely modular and all designed in creative first. I had about 1000 hours of play under my belt by that point. When (inevitably) you realise there are flaws in your designs, stop, go back, tweak your blueprints, return to the game. Mind you it's a bit chicken and egg because if you don't know how to play the game in the first place, you won't get much done in creative either.

I also played without biters and cliffs for most of my "learning time". It takes away the stress of being under attack and you can explore the mechanics of the game at your own pace. You can return to biter worlds when you have more confidence and have mastered at least the basic factory elements to the point where, e.g., you can knock out a green circuit facility without engaging more than a neuron or two. Ignore any macho bullshit that says otherwise.

Also, get to understand ratios. Ratios are the natural predator of spaghetti.

Finally, the best way to learn is to make mistakes, so you are doing it right, except maybe the nuking part. Your first twenty or so bases will be pants. Accept that with tranquility and move on. The reason this game is like crack and has such excellent replay value is because the scope for tweaking and improvement is as near to limitless as makes no odds.

[eradicator]

I also played without biters and cliffs for most of my "learning time".

I actually played with biters for my "learning time", but they were probably less aggressive back then. Cliffs didn't exist either :D. My first ever "need to restart the map" was because i never expanded, and by the time i needed to i didn't have any ammunition left to push back the biters. Now that i understand the game well i don't play with biters anymore, all they do is steal time i could be spending on building more factory! And if a friend wants to play MP with biter i force them to set the starting area to maximum size.

@Expunge:
Nobody can design a bot base before they have bots. People with bot bases build small belt bases just to get to the bots, then completely rebuild with bots only. For a beginner a bot/belt mix is probably best. If you're running out of space...leave more space! The map is large, space is not a rare resource. The first thing you should do is *massively* overproduce belts and inserters, so you never *ever* run out of them, that makes spacious building much more comfortable.

And trains: Do the tutorial *now*! They're not difficult. But everybody (even me, having played a lot of train games) played the first 50 hours or so without just because "oh but belts work fine so far...". But once you try trains you'll never want to go back. And that particular knowledge will prevail even if you nuke the base ;).

[BEEFE]

And trains: Do the tutorial *now*! They're not difficult. But everybody (even me, having played a lot of train games) played the first 50 hours or so without just because "oh but belts work fine so far...". But once you try trains you'll never want to go back. And that particular knowledge will prevail even if you nuke the base .

God yes, trains are so so good. I still use lots of belts, often ludicrously long belts trickling over to an established train stop, but nothing beats trains in terms of throughput vs resources and time invested. It's cool as heck to watch them chumble around, and if you invest in a passenger line, trains are fantastic for fast travel. Especially with the new "temporary stop" mechanic with ctrl-clicking on a piece of rail.

[Ranger_Aurelien]

I think I will agree Factorio is intended to be somewhat organic. There are a few random elements in play for each world -- resource placement, the growth of biters (see the World Generation for default interval windows).

I've had good success with a basic belt system supplemented by robots for balance (0.17), getting a number of worlds to rocket/space science..
1) Early game, start a basic red+green science with stone furnaces.
2) have scouted enough for a good wide belt area, leave single sideways belts to preallocate a belt area (let's call it south to north).
3) Develop science up on one side (say East), and mostly circuits on the other side of the belt (west), initially with underground belts under a few cliffs for now. Send science DOWN the belt on sides of belts to save time and seamlessly be inserted into a science matrix. As soon as useful science is up, set up one assembler making the new thing and putting the output into a box.
4) further back past the beginning of the belt a number of screens, set up the refinery operations as its own octopus, with the belt "backing up" toward it steel, iron, copper (For batteries, cliff explosives, ammo, etc), and sending up acid, lubricant, plastic, batteries, cubes, etc, into the belt.
5) mining and processing further out away from the main base deliver in more belt lanes, keeping away from the main factory area (deliver by train, belt, etc). (Note iron plates have stacks of 100 vs ores 50, so that saves more space in trains/boxes)
6) once robots are mass produced, replace all the "one off" boxes with passive (red) chests, and use them to keep critical assemblers working at full speed. Also find where you have effectively stopped lines and have filter inserters put the excess into passive (red) chests from the very end of lines, or front-splillover with splitters (output priority=mainline, with excess to passive box). For example pretty much all science pack source materials are useful to have in the logistics network!

https://wiki.factorio.com/Science_pack


More high level:
Keep refining away from the manufacturing plant. Set aside a large area for oil refining away from both plant and furnaces (ideally nearish to water and coal).
I visualise the belt as the trunk of a tree and one side's large branches each make one science pack. To simplify the plan, I start by putting down an assembler with the pack, then look at the requirements, then plunk down several more blank assemblers a few squares away (not worrying about connections at all at this point!), and program THEM with the requirements. I look at THOSE requirements, then repeat. It should generally come down to belt-delivered items or things we already have (blue science's oil octopus notwithstanding). I then set up all the belts to get a trickle of ONE of each assembler to get all the components running (with space to increase to dozens at each step), but this will ensure the recipe is right and I have all the pipes and belts set up. I also double-check the grid power level before scaling up.


With Factorio don't be afraid to not understand as you build. See my above where you wade through all the science dependencies and try to get the things the science calls for, then try to scale it up. Learn! More bottlenecks will then appear as supply/demand on your mines/assemblers works itself out. Power browns out so inserters can't grab. Bolster everything up, and in the meantime, labs have researched all the way to the end of the next sciencecolour! Open the next science; repeat!


P.S. As blue science is the focus here, note that you need 2x as many blue assemblers as all the others to get a steady flow (see the link above) : 5:6:5:12:7:7

[eradicator]

2) have scouted enough for a good wide belt area, leave single sideways belts to preallocate a belt area (let's call it south to north).

I prefer a west<->east main-bus, simply because screens are 16:9, so you can see more of the bus without scrolling.

[Zavian]

If you get to a point where you realise that you want to scrap your existing base and build a new one, you can just walk a couple of screens to the north/south/east or west of your existing base and start building again. That is always faster, easier and less effort than starting from scratch again.

[Ranger_Aurelien]

2) have scouted enough for a good wide belt area, leave single sideways belts to preallocate a belt area (let's call it south to north).

I prefer a west<->east main-bus, simply because screens are 16:9, so you can see more of the bus without scrolling.

Thinking about it I'm drawn to north-south Belt so each of my science "branches" perpendicular to the Belt fits well on a 16:9 screen. Flask assemblers perpendicular to the Belt start the branch (closer to the Labs) which dictate the requisite assemblers "north" of them sending their output to the west and then feeding the flask assemblers eastward (So I can extend both flask and material lines eastward with simple CTRL+C, CTRL+V).

[Expunge]

Little update. Spent a few hours in creative mode today to get a better idea of what I'm going for. 120spm seems attainable with a modest bus/starter-ish base that I've put together

Here's my plan for this base: https://kirkmcdonald.github.io/calc.htm ... +//zh/0Tc=

Here's a screenshot of where I'm at. Kinda going for a cityblock Nilaus style base I guess. I'll fill the base as needed by adding smelters and assembly blocks. As I need more throughput to continue building I'll upgrade in place. Eventually switch to some sort of train network bringing material in from the side. Since the base is somewhat modular, perhaps if I get the itch to nuke it I'll just work on one block instead of starting the whole thing over.

huge base image

https://i.imgur.com/SDO1Kf7.jpg

Thoughts?

[Horus773]

wow, this is such a great topic, I have to contribute. Long time lurker, first post

I am currently in the same mindset where I look at my base and compare my creation to all the awesomeness I see here and on youtube and I feel humble to say the least.

My first base was to learn the basic, along with the tutorial. My 2nd was ok but a huge spaghetti and I kinda stalled mid-game. In my current 3rd base I installed all the Bob mods to see how it would go. Awesome new content but there is so much new stuff to learn and no wiki, my base is also a huge spaghetti but with better spacing and organization. I also play with trains which is really fun to say the least.

So I’ve yet to finish the game with a space launch but I have to admit that I now accept spaghetti base as a stepping stone in learning and expanding a base. It cannot be perfect and aligned from the start unless you only use blueprint or know how to build to integrate blueprint. Which is another stepping stone in the game. I only started using them recently because of my train expansion, they were super usefull for that.

Have fun and continue building factories

Anyway, I’m totally hooked on this game and my sleeping habits are all off. Lack of sleep is catching on me. It’s been a while since I enjoyed a game this much. The possibility and replayability are huge.