Does Math Ruin this Game?

[laku]

perfect is a pretty big word imho. you see, while one solution has the highest ratio for production per tile other solutions have the fastest conversion from raw materials to final products. so thats the short answere to why there is no perfect base. it realy depends on what you want.

but i am pretty sure you were mainly talking about the perfect solution in terms of distributing items. well there is no perfect solution either, because you will always have unspent resources in your factory.
even if you are preventing idle machines from picking up resources and even if you are refeeding resources pilling up back into the system, you still have resources that are beeing transported and therefor idle and therefor messing with your ratio.
furthermore there is not even a perfect sollution for just producing a rocket (every hour or whatever) because from time to time you will need railways for example to connect a new or patch. so this means your output is nor predictable.

so you answere your question again ^^: the game is as coplex as you want it to be. even to easiest solution will involve some maths. if a solution is not good enought for YOU then YOU made yourself a new challenge that needs to be solved.
now have fun playing and dont worry about maths

[Nich]

Watching KOS play her new marathon map you can clearly see how punishing over building can be.

Let me ask you this would you build 8 smelting stacks to feed 1 gear machine? how about 1 smelter? No because clearly one is over feed and one is under feed. In the same way you can get is close by eyeball you can get it even closer with MATHS. This drive to perfection is what makes the game fun for me.

What really ruins the game is when others post their builds which are nearly perfect and there is nothing left for me to improve. Now I am stuck using that design as I can think of anything better. On the plus side you can then start looking at the factory as a whole and putting all the different builds together.

[Aru]

... What really ruins the game is when others post their builds which are nearly perfect and there is nothing left for me to improve. Now I am stuck using that design as I can think of anything better. ...

You could resolve to only use your own designs, that's what I do. It doesn't have to be better. My research stuff tends to be ugly and uses lowest-tier assemblers, cheapest approach, which means more of them, more complexity than necessary, takes longer to place it, less symmetry than there could be, ugly overall. Most people don't like that, understandably, but I do. (Point being, that everyone has different priorities and preferences in their designs anyway.) You can personalize your designs, and do everything from scratch. I endeavor to populate a blueprint book with designs that help me do everything from beginning to end. And I do the math myself.

Then you can come back and show them off later, or not, and other people could use some of them, or take inspiration, or not. It's not uncommon in the real world to pit independent teams of engineers on the same problem, disallow communication, then compare the different original, optimized results, take inspiration, and go back in for another round. Progress stagnates without originality, so feel free to be original. Now I'm thinking about the combinator setup to regulate nuclear power, I've picked assembler counts for making yellow science packs, full 0.15 science will be next, but I tend to be really slow making these things, I drag it out because I enjoy it. The only other of my designs that uses combinators is a cracking controller.

[Aeternus]

I think that the standard cookie-cutter ratios (e.g., 3:2 green circuits in non-marathon, furnace to belt compression, etc.) are just the tip of the iceberg.
P.S. Regarding your comment on a playstyle which involves no calcs - I personally very much dislike not planning stuff out because I find the reactive play-style to be extremely tedious once a base gets beyond a certain size (e.g., some of my red circuit assemblers aren't working because it looks like plastic is low --> add a plastic chem plant or two --> now green circuits are low --> add green circuit assemblers --> ...). I would much prefer to do some planning so I can quickly do an audit of my factory's performance using the production tab.

Same here, but I tend to have excess capacity from the ground up. I typically have more mines, smelters, green circuit makers and whatnot so that I'll always have an excess of base resources, which are then fed to more finely tuned higher production plants. I typically also make fairly large buffers of iron, copper and solid fuel (>120K). That way if a mine peters out, you don't suffer a cascading shutdown of production. Having a chain where every input is used in exact amounts to produce a certain output is efficient on paper - but it is only efficient while the entire system is fully supplied and the entire chain is intact. Buffers allow engineering for failure, and if you see a buffer value go down, you have time to investigate what the problem is without losing production.

As for which blueprints/designs are better... Try taking things in a completely different direction and see if you can find a different way to solve the same problem. Always running green and gears on your main belt? See if production locally isn't a better way to go, or the reverse. Enjoying belting things around your base? See if the Logistics system isn't a better option for some tasks. Eco unfriendly coal burner? Try fully powering with a solar array and shutting that burner plant down while the batteries have juice. Running end-to-end trains? Try a shared station setup and see if you can automate it. You may find a creative solution that noone has thought of yet. Or combine ideas of others, pick the best elements and combine them into something even better.

[Edit] Or throw all of that out of the window and load up Bob's mods. I guarantee you, you'll be starting from scratch in terms of figuring out what's what

[Nich]

... What really ruins the game is when others post their builds which are nearly perfect and there is nothing left for me to improve. Now I am stuck using that design as I can think of anything better. ...

You could resolve to only use your own designs, that's what I do. It doesn't have to be better. My research stuff tends to be ugly and uses lowest-tier assemblers, cheapest approach, which means more of them, more complexity than necessary, takes longer to place it, less symmetry than there could be, ugly overall. Most people don't like that, understandably, but I do. (Point being, that everyone has different priorities and preferences in their designs anyway.) You can personalize your designs, and do everything from scratch. I endeavor to populate a blueprint book with designs that help me do everything from beginning to end. And I do the math myself.

Then you can come back and show them off later, or not, and other people could use some of them, or take inspiration, or not. It's not uncommon in the real world to pit independent teams of engineers on the same problem, disallow communication, then compare the different original, optimized results, take inspiration, and go back in for another round. Progress stagnates without originality, so feel free to be original. Now I'm thinking about the combinator setup to regulate nuclear power, I've picked assembler counts for making yellow science packs, full 0.15 science will be next, but I tend to be really slow making these things, I drag it out because I enjoy it. The only other of my designs that uses combinators is a cracking controller.

"Resolve" You my friend are much stronger then myself.

I am cursed with curiosity as to what other people are doing. Once I see a design that is better then what I have currently thought of I will spend hours trying to improve it but in the end the reason I like it so much is because I find they are generally optimized for just what I want. I just personalize it a little. It is still a blatant rip off. I was luckly and made a variant of the Red circuit octagon before I saw a slightly better version on the forums. I currently can not watch KOS past episode 26 because I have not done nuclear yet in my own game and this will be my first attempt at something new in a while.

[mophydeen]

I try to make my blueprints ratio and ocd correct (not perfect).

What ruins the game for me? building blueprints without robots in vanilla.
That's why I add sl-extended to the save file which gives personal robots.

Currently I almost got all important blueprints updated for 0.15 except the nuclear stuff.

I think you can't play factorio without math.

[Stede]

No. The math is awesome. I've always played computer games with a notebook & scratchpad on my desk. I've been known to write custom scripts and functions 1000s of lines long in Excel to visualize data and sandbox things around. It's been awhile, but I've run SAS simulations, as well.

Factorio is one of my favorite games of all time.

[Shokubai]

I'm a former teacher, and love math.

But I'm wondering, for those of you who've got the ratios memorized (or are re-learning them now) or have them on paper just to your side, does it get too boring to simply place everything in the perfect ratio for each game? It seems, and even more so with the blueprint-saving update, that all you have to do is plop objects down, maybe changing the shape based on the geography.

I wonder if I want to learn all the ratios, or if i'd rather just have to deal with the consequences of not knowing the perfect combination of inserters, belts, and furnaces?

Or, does the math make it more fun?

This is one of those personal questions that feels different for everyone. I personally took a long break late in .13 After investing hundreds of hours I wanted to let it rest given the expected improvements in 15. Part of my reason for the break was this scenario. I had initially become obsessed with the perfect ratios for every process. It was fun at first...the planning and figuring out for myself but it killed replay-ability for me because it felt the same every time.

With the launch of .15 I made the concerted effort to pay attention to the ratios but not get hung up on them. The effect has been that I am much more focused on the specific challenges of my map settings then with getting things perfect. I find that I am spending orders of magnitude more time on exploring, building and problem solving than on getting 200 u/m red circuits. When shortages do happen somewhere I address it quickly and with more of an eye toward creativity than perfection.

Today, I am going to put mining drills on a 22mil iron ore node i found and setup shipping. Ill have to take my FARL train home to refill track stuffs and the couple of cars of extra track, assemblers, drills, etc.. and if it fills up from the waiting requester chests I'm going to pull out and go expand my refining station to handle the new input without giving it another thought. It will be an exercise in restraint to not go into my base and check the flow of everything there...BUT that small difference will save me hours of work perfecting something that isn't really broken.

[Frightning]

What really ruins the game is when others post their builds which are nearly perfect and there is nothing left for me to improve. Now I am stuck using that design as I can think of anything better. On the plus side you can then start looking at the factory as a whole and putting all the different builds together.

Honestly, I've not really had this problem because most other people's designs don't fit all of the requires I have for my own designs (I like designs that I can dynamically scale up as needed, and that have virtually no upper-bound for how large they can be made).

[Nich]

Other then robot designs I cant really think of anything that is infinitely expandable. Belt bandwidth runs out surprisingly quickly. I was working on a train concept but after 8ish wagons it still hit an upper limit

[Peter34]

Like Hannu says... in 0.14 I built 14/10 boiler/steam-engine ratios because it's easy to remember and do. But exact solar/accumulator ratio, especially when mods change the day/night lengths and solar output? Nah, I just guess and build extra accumulators or solar if necessary.

For input resources like iron supply, trying to calculate it is pointless because so many parts of the factory might or might not be running for reasons you really can't guess, so just try to increase supply if the belts aren't saturated, or go and troubleshoot why something isn't getting built...

That's how I play too. In most cases I just saturate the Belts and use those as buffer.

(EDIT: Same for fluid productions. Solo I have 10 Storage Tanks each for Crude, Heavy, Light and Petrol Gas. In MP I might have 15 or 20 Tanks for each. I usually have 2 Tanks for Acid.)

Some "math" things I do do:

1. Try to craft 1 or often 2 of each Potion type per nominal second. That is, I'll have 5 or 10 Machines making Red Magic Science Potions, and 6 or 12 making Green Magic Science Potions, etc, and in 0.14 and earlier I'd have 12 or 24 crafting the Blue ones.

2. Electronic Circuits are 3:2 Machines. It's just dumb to do it any other way.

3. Boilers/Steam is 1:20:40. Because then shit works. (Actually, of course, now with 0.15 I probably start with 2 Offshore Pumps and one Belt of Coal, going in through a very few Boilers, each Boiler having 2 Steam Engines, and then I add more Boilers and Steam Engines as the game progresses, until I hit 20 & 40. But I do know the ratio.)

4. When making red circuits, I try to have one local Machine making Copper Wire per 8 Machines making the red circuits. And I do have one Belt dedicated solely to Copper Wire, vs another Belt split evenly between Electronic Circuits on one lane and Plastic on another.

(Disclaimer: It is possible that the ratios for 2 and 4 have changed in 0.15. I don't know yet. But if I find out, then I'll adapt my approach.)

5. I know that in theory, 12 tier-2 Furnaces can saturate one side of a Belt with Iron/Copper Plate, so I go for 10 per side, so that I won't build up a dead buffer.

6. I used to use that neat grid setup with a Substation 16x16 then "filled" in with Accumulators and Solar Panels, which ends up yielding a 3:4 ration, but I think it was supposed to be closer to something like 4:3, but anyway I'd also plonk down some Substations surrounded solely by Accu's. Now Substations are 18x18 but the principle still holds. Plonk down some Solars and some Accu's, and keep doing that to shift away from needing the Steam setup, and if there are brownouts during the night, or from excessive Laser Turret use, plonk down more Accu buffer. So no real arithmetic here, in this #6 item, at all.

That's it! Hardly requiring a PhD in arithmetic, let alone mathematics. No. "Math" isn'r runing this game.

[iceman_1212]

Other then robot designs I cant really think of anything that is infinitely expandable. Belt bandwidth runs out surprisingly quickly. I was working on a train concept but after 8ish wagons it still hit an upper limit

Modular designs as part of distributed production (i.e., no need to have all smelting in one place, all GC production in one place, etc.) are infinitely expandable and very easy to connect up using trains.

[Frightning]

Other then robot designs I cant really think of anything that is infinitely expandable. Belt bandwidth runs out surprisingly quickly. I was working on a train concept but after 8ish wagons it still hit an upper limit

Yes, I tend to prefer robots precisely for that reason, though trains can also scale pretty insanely well.

[Peter34]

That's it! Hardly requiring a PhD in arithmetic, let alone mathematics. No. "Math" isn'r runing this game.

Actually, I lied! Sorry! There is one more ratio of which I am keenly aware, and which I take into account when I play. That is the 1:1 ratio between a Furnace producing Iron Plates and an equivalent-tier Furnace producing Steel Plates.

For instance, I know, based on this ratio, that if I have 20 Furnaces producing Iron Plate and route their total production into another 20 Furnaces producing Steel Plates, then it all matches up, since making a Steel Plate requires 5 Iron Plates but making a Steel Plate also takes 5 times as long as making an Iron Plate (again: same tier Furnaces).

Still doesn't require a PhD in arithmetic. But it is a 7th item, in addition to those 6 that I listed.

[nevniv]

I feel like it only ruins the game for me if I cheat and look up other people's ratios. I only know a couple ones just like Peter, and the rest I just constantly tweak because I'm not sure what is optimal.

Does anyone know if there's a place to get information to calculate the proper ratios on our own? I know assemblers have crafting speeds when I mouseover them but things like how many items per second an inserter moves don't seem to be anywhere in-game that I can figure out. Still somehow people still know this stuff and factor it into their designs

[beltan]

Howdy all.
I'm glad the discussion is going. I forgot a bit about the post, so am going thru all the responses.

As someone mentioned, watching KOS (whom I love to watch), I can see how even those perfect ratios can be daunting.

I've decided that I'm going to not worry about the ratios. Other than a mall, I'm not going to download any blueprints. I'll simply create them as I go. As another person posted, each game is different, so I think the set up should be different.

It is extremely satisfying to see all machines producing efficiently and to see belts/trains/robots flowing fully and smoothly. But I'm having fun also running and fixing fires. ☺

I don't mind learning the ratios, but, as someone mentioned, it is better to 'discover' them on my own, rather than getting the perfect ratio from a blueprint or from an online post.

It would be great if there was a mode that changed all the assembly times for every object: a random amount for each item. Maybe for related materials, the relationship stays the same, but the times are different, so more (or less) machines would be required to complete the product.

[Frightning]

I feel like it only ruins the game for me if I cheat and look up other people's ratios. I only know a couple ones just like Peter, and the rest I just constantly tweak because I'm not sure what is optimal.

Does anyone know if there's a place to get information to calculate the proper ratios on our own? I know assemblers have crafting speeds when I mouseover them but things like how many items per second an inserter moves don't seem to be anywhere in-game that I can figure out. Still somehow people still know this stuff and factor it into their designs

The Factorio wiki has most of the useful not-shown-in-game information that I've made use of on it's relevant pages. Figuring out ratios is entirely a matter of basic arithmetic and the item recipe. Take Plastic bars as an example: (as of 0.14.23) it costs 1 Coal, and 3 Petroleum gas, takes 1 second, and produces 2 Plastic bars. Chemical plants have a crafting speed of 1.25 (going to assume no modules here, but it's not hard to incorporate module effects, just makes the arithmetic messier and the sort of thing I'd actually want a calculator for rather than just doing it in my head on the fly). So, 1 Chemical plant produces 1.25 times 2xPlastic bars every 1 second, or 2.5 per second. It consumes 1.25 Coal and 1.25 times 3=3.75 Petroleum gas per second to do so.

The general formula for calculating production speed is:
i=c*(1+s)*(1+p)*r/t

Where i is items per second, c is (base) crafting speed of relevant machine (player character has c=1 btw), s is speed bonus, % as decimal, p is productivity bonus, % as decimal, r is 'recipe product count' (i.e. how many of the item does the recipe actually make, it's 1 for most recipes, but some have different values, e.g. 2 for Plastic bars, 4 for Land mines, etc.) and t is recipe time (e.g. 1 second for Plastic bars, 0.5 for Copper cable, 10 for Gun turret etc.)

without modules, the 1+ terms are just 1, so it simplifies to:
i=c*r/t
The c values for assembling machines, follow a nice pattern:
assem1 c=.5=2/4
assem2 c=.75=3/4
assem3 c=1.25=5/4
That is 2,3,5 (over 4)...handy for using a mix of assembling machine types and knowing how to compare their throughput rates. Calculating r/t is easy just looking at the item recipe, so you can take that result and just adjust by the relevant c value above to get throughput from given assembling machine.

[Serenity]

For instance, I know, based on this ratio, that if I have 20 Furnaces producing Iron Plate and route their total production into another 20 Furnaces producing Steel Plates, then it all matches up, since making a Steel Plate requires 5 Iron Plates but making a Steel Plate also takes 5 times as long as making an Iron Plate (again: same tier Furnaces).

This allows you to directly insert from the iron into the steel furnace. So you can go ore -> iron -> steel -> output. Basically the same design as a normal smelter line - only wider.

[Jürgen Erhard]

Accounting for random spikes in demand is fine but biters are pretty regular, they don't care if it's day or night. Therefore over time it's fairly safe to assume you'll be getting a similar ratio of biter attacks between day and night as your own factory's power consumption so the best ratio is indeed 0.84.

That being said, Steam Engines are much more powerful now and nuclear is a thing so I probably won't even bother with solar any more xD Now that we have this awesome blueprint library thing and I don't keep having to look up what the design is.

I'm getting back to solar *so* soon... even with more powerful steam (and it's nice indeed) and much more powerful (but wasteful) nuclear... solar is a no-brainer. Plunk down and forget. Both other sources need constant feeding. And, as someone further down said, belt *are* limited.