#module discussion
1 messages · Page 1 of 1 (latest)
It becomes close, yeah
it has a limit, probbly somewhere around 7.5
It approaches some amount less than 8, depending on the length of the beacon tile you are tiling.
So if you’re tiling in the X direction
The limit as you tile in the X direction depends on the length Y of the build
Since you always have beacon endcaps in the Y direction
Yes, you increase both linearly
You cannot do that though
Why not?
You have to supply logistics along either the Y or X direction
You can’t expand that infinitely and maintain an 8 beacon tile
The roboport takes some space so it also reduces beacon efficiency
If you add roboports in the middle of an 8 beacon build it ruins the module efficiency
It adds new endcaps
In the middle
Threads are good for debates
yeah there were people asking questions so i moved it here
It lowers it yes
But you can abaolutley make it so it approaches 8
Its not effective for an actual game, because you would be building ludicrous sizes
approaches means it hits 8 so it wont hit 8, unless you have some other definition
Approaches means it gets infinitely close to 8 without hitting it
Like f(x)=1/x
Although it would be a different function here
I think you'll hit a limit but im not sure, stretching it to infinity?
It’s not 8.
But gets closer the closer you get to infinite
Sigh… okay. I’ll show you
Or rather
Show me a build you can expand infinitely in both directions
That is 8 beacon compact
If you don’t have one we can analyze any other beacon builds
And set up a limit ourselves
This is a thought experiment, I haven't actually thought what it would look like
I guarantee it won’t be approaching 8
The thing is, that infinitely expanding 8 beacon square is literally impossible
I mean for one thing map limits
But for another logistics….
If you assume logistics is no issue and an infinite map, it approaches 8, no?
Roboports too big, belts run out of throughput.
Your only chance is miners exactly where you need them
Refilling the beacon line as it goes
An entire science factory inside 8 beacon
And even that hits the map limits
So not 8.
If logistics is no issue and the map is infinite 8 is possible yes
Right I mean
Thats what I'm going off of.
Just typing quick 
I can try to figure out how to infinitely feed it while impacting the efficiency minimally
So maybe you can have them assembling different things
And have it loop every X tiles with an input of base materials
With sushi belts and miners where you need them
that sounds fun to build
I guess it’s possible
super fun... i might try that lmao
Its possible, albeit likely only in editor
But module efficiency though or beacon efficiency whatever we call this
It should really be multiplied by average machine uptime
I thought we are only computing machine/beacon
(Average Machines per beacon) / (average beacons per machine) * average machine uptime
Adding uptime might skew this a lot
I’m trying to come up with a new metric that includes both
itll make it a ton more complicated and reduce beacon efficiency
I’m trying to come up with a new metric that includes both
Or rather, it's going to make it a hell of a lot more complicated
Or rather, it's going to make it a hell of a lot more complicated
I want a metric that fully analyses the efficiency of you spending your resources on moduled designs
I think the equation I posted is the one…
A scale of 0-1
And then make the de facto most efficient machine?
With 1 being perfect
Yes…
Factories with module efficiency scores close to 1 are good for module transition
Because they are extremely cheap in module cost
You might be able to get full uptime If the chain of machines is perfect
The idea is to hit decently long 8 beacon designs I think.
And tile them to be squares
As long as you feasibly can
And then layer them with each other
With full uptime
You need a final output as well
This works a lot better with infinitely fast belts haha
Oh without a doubt
I won’t be approaching infinity length or width hehe
Just need enough to be far down on the module efficiency
See I think I’ll be upgrading these builds to blue belt or at least red belt
More length.
But their module efficiency is pretty good.
I think some repeating unit of science production would be key
And just assuming ore is under it.
Idk.
Since it's the only thing that theoretically actually uses up end products fully
Yeah you see why I was confidently calling the limit -> 8 impossible xD
I honestly don't think it's impossible.
I think I went down this thought experiment once and was like shit ima have to settle huh
Its possible, albeit incredibly complicated
If we ignore machine uptime, I think a design that freely spawns ore will be possible
Uptime makes it infinitely more complicated tho
So much more
But outside this infinite theory, it’s super useful for actually comparing these designs
Cause I think if we set up this calculation right we could make it a percent modifier to the payoff time…. 
Hmmm
No that’s just Straight up the uptime already
I lied
Yeah no need include it
A universal formula for moduled efficiency?
Having it separate is more useful
If you want to properly measure efficiency, we also need to measure output vs power×space
(Average Machines per beacon) / (average beacons per machine), and average machine uptime should tell you everything you need to know?
Ehhh power,
Ehhh space 😛
that should bea seperate calculation
Very cheap concerns tbh
one for module efficiency,
one for power (if you want)
one for space (if you want)
Also the same builds that optimize module efficiency optimize those… usually
Like you can just list 5 things I think
Output rate, power cost, space, module efficiency, average machine uptime
Space density is one metric that we can measure as
Output per second/(power used×space used)
not for power but for space yeah
The issue with that is it only matters for comparing designs of the same end material
I would just call that a secondary statistic, if you look at the 5 I posted you can easily get that from those 5
You can see it on factorio lab for instance
If you need 4.873 machines
And you have 5
Uptime of 4.873 / 5
Assuming full output right
Like when you’re output limited by a belt or whatever
Bots would be good for machine uptime
So I think the best builds for your module transition would be those where the uptime * the module efficiency is highest
Which makes things interesting
You can sacrifice one or the other if one or the other is particularly strong
hm so maybe have it not output a full belt is best?
Oh I see
Slightly under ratiod gives us best uptime then
So for example 55 machines if we need 56
That way if there any hiccups, we still have full uptime
Yes, this can be better. Assuming the modules are significantly more expensive than the belts
This might not be good inside like… tier 1 module + blue belt
The best application for this imo is making beacons, modules, and assembly machines
Have an infinite assembly line that is assembling itself
Put down infinite blueprint, and it makes the resources as it goes abd builds it
Fuck I'm gonna make that
😄
Massive blueprint that you build with 1 assembly machine, 1 construction bot, 1 miner
nitro ‘Perfect module efficiency’ Rick
And as it expands, it builds itself
It replicates building a factory the way you would
But.. all pre-planned
And you make the bare minimum
Then just time-lapse the rest
Imagine doing that until rocket
I’m kinda trying to preplan a factory like that from beginning and time lapse it
With very efficient designs so the time lapse is nutty
On deathworld hehe
Just a bp of the entire factory, but spaced perfectly so it can only build in the order you want
Oh I see
Nah, ofc you space it out properly
Or better yet, you can time it with power switches and combinators
So after x time, the next area opens up
And it builds that
It would be 1 bot all the way to blue science lmao
Might take a while
My sushi castle kinda works like this
But its not really the same
Yeah you’d have to have like micro builds or something
I think there’s a mod like this
Removes players characters
You start with bot
And you gotta play sandbox like
That's not quite what I'm thinking
That does seem interesting tho
I'm just saying a massive factory BP that is designed to build itself
In vanilla
I’m kindaaa working on this? I mean its more like for practical usage but that’s how it will go down
Like the top middle block is super compact and just makes robots off 2 robot frames asssemblers
And then this whole thing will have roboports
And it’s super tight so the robots can feasibly build it
So I’m hoping this whole thing can be finished in like ~ 8 hours. Or something.
In a new deathworld
Something like that, but starting way earlier
And you can see the bigger base
The bottom left corner is a mall mostly
So this base will build a bigger one on an even larger scale in the next 20 hours or so
Yeah I guess it would be similar to that
Added realistic time estimates xD
But instead of starting with robot production, you do it from scratch
The first ones were more hope and dream
All the way to robot production, and then all the way to rocket
Yea, for this base it’s just the player doing it by hand
But I mean if there were robots I guess they could do it… slowly
Hmmm I could time the roboport inserter
You’d want like a bare minimum robot prod chain
Yeah sushi belts would be good for this
For seeding new networks and taking refined products network to network
Segment into areas, and when an area is finished, it places a combinator down that sends a signal to add a roboport to the logistics network
So start the next area
You can check for it being done via construction robots not working for like 2 min straight
Also works