Factorio Forums

www.factorio.com
https://forums.factorio.com/

Quality and Speed: How much waste does haste make?

https://forums.factorio.com/viewtopic.php?f=5&t=123768

Page 1 of 4

Quality and Speed: How much waste does haste make?

Posted: Thu Dec 05, 2024 5:59 am
by Romayne
What?
So, I've been plagued with the sight of discussion about people using speed modules in conjunction with their builds with quality modules in them. At first, and even at the time of writing, it just seems like a ridiculous idea. Ultimately, any influence of speed on a building will reduce the quality output, which means more input materials for the same output.

However, I do have just a lingering doubt about this being so simple as an open and shut case, as there are situations and reasons that surely must have evoked this kind of curiosity from people, such as not enough space or materials for a full-scale quality setup (reasonable, Vulcanus/Fulgora are terrible with free space for this)

Well, undoubtedly it's true that a faster building will produce more items, but the real question I want to answer here is what are the differences in design, stats, and use cases for designs which do and do not use speed modules?
Methodology/Constraints
Naturally I want this to be as objective as possible, and to do so this requires adding on constraints that can make the application of this information situational or inapplicable. That's kind of the gist of things here, but oh well. Let's assume:
- Production goal is the highest quantity of legendary electric engine units possible
- There needs to be SOME kind of target, and I think EEUs are good here as it's a relatively far up ingredient that doesn't benefit from much of what SA has to offer that would bias towards either. Not the +300% prod of rocket parts and/or plastic, not the insane buildings, etc.
- Unlimited resources (raw resources to get it finished. molten metal & lubricant, along with prod technology. No intermediaries, no final products, of course. and only normal inputs, not uncommon/rare/epic/legendary. that defeats the point.)
- This favors the haste part, since surely if you have all the materials you want the waste doesn't matter. Right?
- 200x200 space to get production done
- This enforces the spacial kind of constraint that this design approach seems to be derived from
- All resources (infinity chests, infinity pipes) must come from some unload zone on the left, all production must happen to the right of that zone.
- This forces the constraints to be more like Freeplay mode, as materials tend to come from one side with designs rather than from infinity chests and pipes in the midst of things.

As for the buildings, I've made a few starter designs to kick things off. I'll measure the effectiveness of builds by a few metrics, so as to allow the greatest comparison between them:
- Input/Output ratio: The fractional ratio of each input raw ingredient to the output ingredient, from their (items or fluids)/minute.
- Space usage: Roughly what tile area was used by the design to produce the materials. This will be the square area (farthest extremities in the NE/NW/SE/SW corners).
- Actual output rate: How many EEUs the design outputs per minute.

Overall, this should compare the resource efficiency vs. the throughput density of these kinds of designs, and see how the two differ, if much at all, or if greatly so.
Design 1: Demo/Standard (no speed)
Image
- Molten Iron: ~39,737/m (~974.0 molten iron : 1 Q5 Electric Engine Unit)
- Molten Copper: ~7,188/m (~176.2 molten copper : 1 Q5 Electric Engine Unit)
- Lubricant: ~10,851.8/m (~266.0 lubricant : 1 Q5 Electric Engine Unit)
- Q5 Electric Engine Unit: ~40.8/m
- Width: 97 | Height: 69

Design 2: Demo/Standard (legendary speed 1 modules)
This design scattered normal beacons with 1 legendary speed 1 module and 1 empty slot around the design.

- Molten Iron: ~44,079.7/m (~1057.1 molten iron : 1 Q5 Electric Engine Unit)
- Molten Copper: ~7,205.3/m (~172.8 molten copper : 1 Q5 Electric Engine Unit)
- Lubricant: ~13,563.8/m (~325.27 lubricant : 1 Q5 Electric Engine Unit)
- Q5 Electric Engine Unit: ~41.7/m
- Width: 102 | Height: 70
Image



So interestingly, the total EEU output increased by 2.2%, for a space increase of 6.6%, and an increase of material cost (for which I will only check lubricant for now) of 25%.

Also, if you're wondering 'but what about T3 Q5 speed modules instead of T1 Q5?' I tried this and it decreases the output of Q5 EEUs by somewhere around 25%.

Here's cryoplants for the best possible case for quality, where the relative decrease in quality is minimal due to the 8 module slots available.
Design 3: Cryo/Default
Consumption:
- Refined Concrete: 2,276.5/m (227.7 refined concrete: 1 q5 cryoplant)
- Processing Unit: 1,138.3/m (113.8 processing unit 1 q5 cryoplant)
- Superconductor: 1,138.3/m (113.8 superconductor: 1 q5 cryoplant)
- Lithium Plate: 1,138.3/m (113.8 plate: 1 q5 cryoplant)

- Width: 68 | Height: 32
- Q5 Cryoplant: ~10.0/m

Image
Design 4: Cryo/Speed (legendary speed 3 modules)
Consumption:
- Refined Concrete: ~6825.3/m (~437.5 refined concrete: 1 q5 cryoplant)
- Processing Unit: ~3412.7/m (~218.8 processing unit 1 q5 cryoplant)
- Superconductor: ~3412.7/m (~218.8 superconductor: 1 q5 cryoplant)
- Lithium Plate: ~3412.7/m (~218.8 plate: 1 q5 cryoplant)

- Width: 68 | Height: 40
- Q5 Cryoplant: ~15.6/m

Image

(Editor's note: Blueprint strings not provided for these designs as it brings the post above its character limit.)

The result?

- 25% increase in space
- 56% increase in output
- 92.27% RELATIVE increase in resource usage per Q5 cryoplant, 300% increase in TOTAL resource usage

I only have found evidence against any level of speed in quality setups. Please submit your designs if you have evidence suggesting otherwise.
EDIT: So far one post about plastic with speed being a notable space saver, see below.

Re: Quality and Speed: How much waste does haste make?

Posted: Fri Dec 06, 2024 11:02 am
by ichaleynbin
Image
Which of these three setups produces the most legendary plastic per second?

Re: Quality and Speed: How much waste does haste make?

Posted: Fri Dec 06, 2024 11:15 am
by ichaleynbin
inb4 "What about the rest of the setup?" you do the same thing at all tiers and get a 800%++ increase across the board. Roll common plastic the same way.

It's very simple. In the extreme version, would you rather roll a 1% dice one million times, or a 50% dice one thousand times? The 1% is expected to succeed more times than you even roll the 50%, at 10k. 37% of 178 is 65, more than four times as many successes as you even roll dice without the speed beacon.

It's not even remotely close, you apply this at every step of the chain and it's not about equal, plastic cryoplants have truly absurd gains on per second outputs. This particular example is over an 800% increase.

Re: Quality and Speed: How much waste does haste make?

Posted: Fri Dec 06, 2024 3:53 pm
by Romayne
Hi, I see where you are coming from but please provide an actual working design.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 8:54 am
by RockPaperKatana
Just a suggestion, don't block everyone who told you you are wrong. Your hostility means nobody cares about your opinion and nobody wants to spend any effort to give you counterexamples, even though it's so simple and obvious.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 10:19 am
by mtzu
120/s of Coal consumption, no speed (doesn't completely fit on one screenshot):
grafik.png
grafik.png (475.69 KiB) Viewed 1314 times


~ 120/s of coal consumption, 1 Legendary Beacon with 2 Legendary Speed 3 Modules per Machine:
grafik.png
grafik.png (1.25 MiB) Viewed 1314 times


My save has Plastic Productivity 1 unlocked, but that's just a multiplicative factor on all plastic output, so it doesn't really matter, when we compare both designs. Both Designs consume roughly the same amount of coal.

The no speed design:
- takes 100 Cryoplants and 800 legendary quality 3 modules to build.
- outputs 120 / s * 2 * 1.1 * 0.496 = 131 quality plastic (uncommon or better) per second.

The 2nd Design:
- takes 9 Cryoplants, 72 legendary quality 3 modules, 3 legendary beacons and 6 legendary speed 3 modules to build.
- outputs 120 / s * 2 * 1.1 * 0.375 = 97.9 quality plastic per second.

Since the chance to skip a quality tier is always 10 %, independend of the quality chance, the ratio of uncommon:rare:epic:legendary plastic is in both cases 900 : 90 : 9 : 1. I will always tank a 25 % productivity decrease for a 10 fold decrease in build size. 800 legendary quality 3 modules are expensive, even in late game.

Edit: I corrected a small mistake in the quality ratios.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 11:25 am
by ichaleynbin
RockPaperKatana wrote: ↑Sat Dec 07, 2024 8:54 am Just a suggestion, don't block everyone who told you you are wrong. Your hostility means nobody cares about your opinion and nobody wants to spend any effort to give you counterexamples, even though it's so simple and obvious.
Exactly this. OP, if you has just listened to me instead of blocking me on discord, you could've avoided this entire embarrassing thing when I showed several examples. It's funny because you kept saying "I haven't seen any examples," lol yeah because you blocked the people who disagreed with you and were showing you examples. If you had read my messages you could've seen several.

It's not even remotely close and I showed you that with envelope math before you posted this thread. I'm not going to put a bunch of work in to prove something to your standards when a back of the envelope calculation can show an 800% difference. I don't know what's optimal, so far it seems one legendary Speed 1 module is excellent, but speed beacons are in an entirely different league.

Prove? Setup? No, ~800% per machine. It's not even close, why would I waste more of my time when you can't see the obvious?

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 11:31 am
by ichaleynbin
Romayne wrote: ↑Fri Dec 06, 2024 3:53 pm Hi, I see where you are coming from but please provide an actual working design.
I posted videos of the comparative designs I am using in my actual base, on discord. Unblock me and apologize to me, and I'll relink it to you there.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 4:15 pm
by Romayne
You can add it to this post or create a new post if you'd like.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 4:20 pm
by Romayne
mtzu wrote: ↑Sat Dec 07, 2024 10:19 am 120/s of Coal consumption, no speed (doesn't completely fit on one screenshot):

grafik.png


~ 120/s of coal consumption, 1 Legendary Beacon with 2 Legendary Speed 3 Modules per Machine:

grafik.png


My save has Plastic Productivity 1 unlocked, but that's just a multiplicative factor on all plastic output, so it doesn't really matter, when we compare both designs. Both Designs consume roughly the same amount of coal.

The no speed design:
- takes 100 Cryoplants and 800 legendary quality 3 modules to build.
- outputs 120 / s * 2 * 1.1 * 0.496 = 131 quality plastic (uncommon or better) per second.

The 2nd Design:
- takes 9 Cryoplants, 72 legendary quality 3 modules, 3 legendary beacons and 6 legendary speed 3 modules to build.
- outputs 120 / s * 2 * 1.1 * 0.375 = 97.9 quality plastic per second.

Since the chance to skip a quality tier is always 10 %, independend of the quality chance, the ratio of uncommon:rare:epic:legendary plastic is in both cases 1000 : 100 : 10 : 1. I will always tank a 25 % productivity decrease for a 10 fold decrease in build size. 800 legendary quality 3 modules are expensive, even in late game.
It's good to know there's an application out there where this can be seen as worth it for space concerns. It's unfortunate to hear that the total output still tanks, but at least there's an upside here. I'm interested to see how far this idea can be taken with other intermediary products, and whether this kind of efficiency is limited only to cryoplants -- though I don't imagine it is so much as it merely is on a lesser scale for other buildings.

In any case, thank you for this submission :)

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 4:44 pm
by mtzu
Well, the more quality modules you have, the less your productivity is relatively decreased by adding a speed beacon. So the cryoplant definitely is the best building to do this with.

Before I knew about the concept of asteroid rerolling I used a Lava -> Molten Iron -> Iron Plates -> Upcycle loop to create legendary iron and I used speed beacons for the Molten Iron -> Iron Plates foundries, since you really do not care about the productivity of these buildings but only their output rate.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 4:57 pm
by mtzu
I wished, that quality actually worked like the corresponding page in the tips suggested. If the quality chance actually affected the fixed 10% chance to skip quality tiers, then it would probably not make a lot of sense to add speed or even productivity into any quality builds. But as the quality system doesn't work the way, it is described ingame, you will always benefit from adding speed beacons to your "base producers", as long as you do not care about the increase in consumption of the used resource.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 6:35 pm
by aka13
Is asteroid rerolling still a thing? I hoped it would be removed.

Re: Quality and Speed: How much waste does haste make?

Posted: Sat Dec 07, 2024 10:20 pm
by coffee-factorio
ichaleynbin wrote: ↑Fri Dec 06, 2024 11:02 am Image
Which of these three setups produces the most legendary plastic per second?
None of the above.

I built models to see what a single plant could actually do. Based off 3 scenarios:

Plastic with no beacons and quality.
Plastic with 1 beacon and quality.

Plastic with full productivity modules and quality in the drills
with a recycler scheme.

When I tried out doing a toy version of just green chips to have something to come to the thread with, it seemed that there where some issues in numbers being reported because I could get 6300 rate off an impractical toy. But off 10 buildings I could both beat that rate and as OP suggested I would get better material use doing it.

OP, you're right on by the way.

So what I thought then is: arrange the recyling so it's done in front. Then use best productivity to boost the scenario and see what comes of that - it's hard to calculate what .18/1000 and .31/1000 rate is through recyclers but it's easy to build toy builds in the map editor to practically test the thing.

Image



This one does 10 plastic per minute off 1 building over an hour run by using productivity instead of quality.



By using quality at the point of the lab you get 1.1 item per minute and 6 drills.
Coal use over 1 hour is 180 items/minute off a yellow belt. For the other one you get a nonsense number of 1.4K items a minute, but it underfills a yellow belt visibly even though I'm using 8 drills. I tried upgrading to red belts (I might as well as follow my own advice and make sure I'm not making an i/o mistake on an impractical toy build). But for an infeed of 3 machine energy I'm seeing that 1.1 items/minute number consistent. Another person reported 90 IPS plastic off 100 buildings, if they were reading off the production chart those units are in minutes. I would welcome a practical check and a sanity check on this toy, but that doesn't seem like it scales right.

Does this invalidate using speed modules for quality? Not yet.

This is what happens:

Image



I had to upgrade everything and then turn all my belts green. And add an entire second belt. To get, 4.7 ipm out of it for 2000 coal input. So the best choice was: 8 drills and 1 productivity modded beacon, 8 super speed drills and 1 beacon each did half that rate. 6 drills and no beacons did 1/10th of the best rate.

I'm sorry that I can't give you better than practical toys, but OP is right :D

Re: Quality and Speed: How much waste does haste make?

Posted: Sun Dec 08, 2024 1:19 am
by Khazul
No need to mix them.

You use pure quality via highest quality quality modules you have in recycling/reprocessing steps and use highest quality productivity modules + speed beacon in the production steps. The only place I tend to mix depending on need is miners and for low levels of opportunistic passive quality byproduct - ie where I dont really care, but a few quality products are useful.

There are highly scalable ways of getting some basic legendary resources (iron ore, coal, cooper, steel etc) at end game. The challenge is getting to that stage of the game in the first place as you need a lot of science for maxing out productivity while in the background you keep cranking up you quality of modules and eventually some production buildings for the speed.

The two modules to crank the quality up ASAP are obviously quality but also productivity as high productivity can at least partially offset losses in recycling/reprocessing. If you are pure quality only in the recycling/reprocessing steps, then it is easy to massively scale that up as its just one machine and not a complex production chain. Eventually you crank up the quality of building including beacons and also speed modules, then you end up with fast production chains that dont need very many buildings and productivity that largely offset the losses in recycling to upcycle for more obscure parts.

Some may argue the end game is broken, but to get to that stage in the first place to be able to abuse it is a massive undertaking, so breaking it is your reward IMHO :)

Re: Quality and Speed: How much waste does haste make?

Posted: Mon Dec 09, 2024 2:58 am
by RockPaperKatana
coffee-factorio wrote: ↑Sat Dec 07, 2024 10:20 pm
ichaleynbin wrote: ↑Fri Dec 06, 2024 11:02 am Image
Which of these three setups produces the most legendary plastic per second?
None of the above.

I built models to see what a single plant could actually do. Based off 3 scenarios:

Plastic with no beacons and quality.
Plastic with 1 beacon and quality.

Plastic with full productivity modules and quality in the drills
with a recycler scheme.

When I tried out doing a toy version of just green chips to have something to come to the thread with, it seemed that there where some issues in numbers being reported because I could get 6300 rate off an impractical toy. But off 10 buildings I could both beat that rate and as OP suggested I would get better material use doing it.

OP, you're right on by the way.

So what I thought then is: arrange the recyling so it's done in front. Then use best productivity to boost the scenario and see what comes of that - it's hard to calculate what .18/1000 and .31/1000 rate is through recyclers but it's easy to build toy builds in the map editor to practically test the thing.

Image



This one does 10 plastic per minute off 1 building over an hour run by using productivity instead of quality.



By using quality at the point of the lab you get 1.1 item per minute and 6 drills.
Coal use over 1 hour is 180 items/minute off a yellow belt. For the other one you get a nonsense number of 1.4K items a minute, but it underfills a yellow belt visibly even though I'm using 8 drills. I tried upgrading to red belts (I might as well as follow my own advice and make sure I'm not making an i/o mistake on an impractical toy build). But for an infeed of 3 machine energy I'm seeing that 1.1 items/minute number consistent. Another person reported 90 IPS plastic off 100 buildings, if they were reading off the production chart those units are in minutes. I would welcome a practical check and a sanity check on this toy, but that doesn't seem like it scales right.

Does this invalidate using speed modules for quality? Not yet.

This is what happens:

Image



I had to upgrade everything and then turn all my belts green. And add an entire second belt. To get, 4.7 ipm out of it for 2000 coal input. So the best choice was: 8 drills and 1 productivity modded beacon, 8 super speed drills and 1 beacon each did half that rate. 6 drills and no beacons did 1/10th of the best rate.

I'm sorry that I can't give you better than practical toys, but OP is right :D
You are missing the point and the comparison wasn't about quality modules in drills at the first place.

Re: Quality and Speed: How much waste does haste make?

Posted: Mon Dec 09, 2024 4:19 am
by Khazul
aka13 wrote: ↑Sat Dec 07, 2024 6:35 pm Is asteroid rerolling still a thing? I hoped it would be removed.
Why it that any different to recycling and building ladders? If anything it lacks the option for a productivity gain step as is the case with many make a component something and recycle it ladders so you basically have to brute force it via huge numbers of crushers.

I use asteroid reprocessing to get a couple of specific raw materials to legendary and I am not even sure there are other routes for one of them.

Re: Quality and Speed: How much waste does haste make?

Posted: Mon Dec 09, 2024 7:41 am
by mrvn
Khazul wrote: ↑Mon Dec 09, 2024 4:19 am
aka13 wrote: ↑Sat Dec 07, 2024 6:35 pm Is asteroid rerolling still a thing? I hoped it would be removed.
Why it that any different to recycling and building ladders? If anything it lacks the option for a productivity gain step as is the case with many make a component something and recycle it ladders so you basically have to brute force it via huge numbers of crushers.

I use asteroid reprocessing to get a couple of specific raw materials to legendary and I am not even sure there are other routes for one of them.
You mean you have quality modules in the asteroid reprocessing and loop it to get legendary asteroid chunks and then you make legendary iron ore?

You can always recycle iron ore with quality modules to get 25% * 10% = 2.5% better iron ore and 75% loss. It exists, it's just not as good an option.

Re: Quality and Speed: How much waste does haste make?

Posted: Mon Dec 09, 2024 1:22 pm
by Khazul
mrvn wrote: ↑Mon Dec 09, 2024 7:41 am
Khazul wrote: ↑Mon Dec 09, 2024 4:19 am
aka13 wrote: ↑Sat Dec 07, 2024 6:35 pm Is asteroid rerolling still a thing? I hoped it would be removed.
Why it that any different to recycling and building ladders? If anything it lacks the option for a productivity gain step as is the case with many make a component something and recycle it ladders so you basically have to brute force it via huge numbers of crushers.

I use asteroid reprocessing to get a couple of specific raw materials to legendary and I am not even sure there are other routes for one of them.
You mean you have quality modules in the asteroid reprocessing and loop it to get legendary asteroid chunks and then you make legendary iron ore?

You can always recycle iron ore with quality modules to get 25% * 10% = 2.5% better iron ore and 75% loss. It exists, it's just not as good an option.
I was using it for making coal and calcite.

Re: Quality and Speed: How much waste does haste make?

Posted: Mon Dec 09, 2024 9:15 pm
by Dictator
Quality quality modules can handle some haste due to higher base quality, but you can forget about skipping whole tiers of quality.
Legendary speed modules create more haste with same waste (quality reduction)

Here are quality chances with legendary quality modules.
obraz.png
obraz.png (283.31 KiB) Viewed 778 times
Cryogenic plants get 8 modules, so 50% quality chance.

One beacon with one legendary module reduces quality chance by -6.25% chance while tripling speed.
Crusher doubles quality+1 production, while cryogenics almost triple it.

Even with two legendary speed modules affecting it cryogenics can generate quality very fast.
obraz.png
obraz.png (255.96 KiB) Viewed 778 times
So there is net positive combination of quality quality and speed module, where lower chances are more than countered with increase of speed.
All times are UTC
Page 1 of 4

Powered by phpBB® Forum Software © phpBB Limited