Friday Facts #375 - Quality

[Mamizoi]

I am in love with this feature after giving it more thought so I thought i'd make an account just to rave about it.

I have never like how factorio is most efficiently played by copypasting blueprints made by some other fellow who's been playing the game for far longer than you have. That stank, I never liked that.
Quality either by intention or not basically removes that, given what items you own are high quality, you will have to build and adjust in real time to that. I cannot wait. No two perfectly optimal factories (outside of 100% legendary) will ever be the same again!

I was against this feature before, what, luck in my factory game? What ever happened to calculating to perfection!
But now that i think about it, staring at helmod or a spreadsheet was never.. actually fun..

I would just see i need however many dozen of a machine.. and place that many.. and then place the belts that reach the throughput... boooring
Now i'll have to dynamically decide where i put my best machines to maximise what i need at any moment in time, the factory will now shapeshift as it grows and i cannot wait for the spaghetti of recyclers.

I don't think I'll ever play another game once this expansion releases.

[Neutronium]

Does anyone know if you can recycle ore or plates, and if you can upgrade their quality from quality modules?

[Axs1]

Does anyone know if you can recycle ore or plates, and if you can upgrade their quality from quality modules?

New building: Recycler, gives a 25% chance of returning items created from some types of recipe (not furnaces or chemical plants), and doesn't return fluids; also has 4 module slots.

Source: https://wiki.factorio.com/Roadmap

[Neutronium]

Does anyone know if you can recycle ore or plates, and if you can upgrade their quality from quality modules?

New building: Recycler, gives a 25% chance of returning items created from some types of recipe (not furnaces or chemical plants), and doesn't return fluids; also has 4 module slots.

Source: https://wiki.factorio.com/Roadmap

Much appreciated. It seems like it still won't be too bad by running grinding cycles through stuff like electronic circuits, gears or copper cable, depending on what you need handled. Although with copper cable, since 1 plate -> 2 cables, I guess its 12.5% chance of yielding a plate.

[Axs1]

Does anyone know if you can recycle ore or plates, and if you can upgrade their quality from quality modules?

New building: Recycler, gives a 25% chance of returning items created from some types of recipe (not furnaces or chemical plants), and doesn't return fluids; also has 4 module slots.

Source: https://wiki.factorio.com/Roadmap

Much appreciated. It seems like it still won't be too bad by running grinding cycles through stuff like electronic circuits, gears or copper cable, depending on what you need handled. Although with copper cable, since 1 plate -> 2 cables, I guess its 12.5% chance of yielding a plate.

Yes, it is better to obtain the superior qualities only with recipes that return raw materials and thus avoid the entire complicated process just to obtain a specific final product.

[computeraddict]

Yes, it is better to obtain the superior qualities only with recipes that return raw materials and thus avoid the entire complicated process just to obtain a specific final product.

You are assuming quality modules are free and have no opportunity cost, no?

Using a quality module in place of a productivity module costs extra materials in lost opportunity. For every 1 quality product you could have produced ~4 extra non-quality products if you'd used an equivalent productivity module instead.

To boot, high quality quality modules are nutty expensive, and best used where the most material possible will be moving through the smallest number of machines, which is the final assembling/recycling loop (in much the same way that productivity modules are best used first in the machine that the most resources go through, the rocket silo).

[mmmPI]

To boot, high quality quality modules are nutty expensive, and best used where the most material possible will be moving through the smallest number of machines, which is the final assembling/recycling loop (in much the same way that productivity modules are best used first in the machine that the most resources go through, the rocket silo).

I disagree that the place where the most material will be moving though the smallest number of machine has to be the final step, this highly depend on what you are trying to get as an item.

Imagine you have your first 4 to 10 quality module, would you put them in the inserter assembly ? the roboport ? the armor ? the reactor ? the spidertron ? something for the space platform ?

There so many "last product", maybe you're better off keeping quality module to recycle green circuit( 2 steps of quality module ) and keep legendary iron plate and legendary copper, and legendary green circuit. Then from those material use productivity module up to the end no ? Will allow to make all the different item in trickle without investing 4 module per item.

[computeraddict]

Imagine you have your first 4 to 10 quality module, would you put them in the inserter assembly ? the roboport ? the armor ? the reactor ? the spidertron ? something for the space platform ?

I'd put them in the assembly loop for higher quality quality modules.

There so many "last product", maybe you're better off keeping quality module to recycle green circuit( 2 steps of quality module ) and keep legendary iron plate and legendary copper, and legendary green circuit. Then from those material use productivity module up to the end no ? Will allow to make all the different item in trickle without investing 4 module per item.

The first chokepoint for every item recycling loop is the recycler itself. You can bootstrap a quality process for anything and everything with just a quality'd recycler. At that point, there's no reason to feed anything that can be productivity moduled into it, since you are gaining quality by recycling and recrafting through two processes that can't accept productivity modules at all.

As for the "trickle of everything", you actually have *higher* start up costs to try to quality cycle intermediates than to cycle the final products. Quality cycling intermediates takes more recyclers and more recyclers per item type since for every, say, quality stack inserter you have to produce 15 green chips, 1 red chip, a blue inserter, and 15 gears in the target quality first. So you could put modules in one recycler and one assembler per tier you want to go up, or you could multiply that infrastructure by 32 times (more than 32 times since the red chip has a craft time greater than 0.5s) to produce the components in quality first at the same rate. And that's in addition to losing free materials from the productivity bonus.

To wit, it takes the same number of cycles through a quality recycling loop to produce any item in quality. As in, you could use the modules to produce a single quality stack inserter chain, single quality t3 assembler chain, single quality speed 3 module chain, single quality quality 3 module chain, single quality prod 3 module chain, single quality electric furnace chain, and so on for 26 more different product types and it would churn out more high quality products, for cheaper, than spending those same modules on a bunch of quality chains for intermediates that could only match the output rate of that first quality stack inserter chain.

[mmmPI]

Imagine you have your first 4 to 10 quality module, would you put them in the inserter assembly ? the roboport ? the armor ? the reactor ? the spidertron ? something for the space platform ?

I'd put them in the assembly loop for higher quality quality modules.

But where ? at the very begining first like the green circuit ? they are part of the loop ?
You mean in the assembly first or in the recycler first ?
And then which step of recycling is the better ?
What if recycling is much faster than production ?

If we consider tier 3 quality module, they are made from tier 2 quality module. It's not a given that one would prefer placing its first quality module in the latest step of the chain, instead of recycling the previous. Say recycling tier 1 module is fast ( 5 second ), and making tier 3 module is slow ( 60 second ). Then there would possibly be more material processed in the recycling of tier 1 module rather than in the attempt of making new tier 3 module.

That would mean, it would be better to place the first quality module to recycle every tier 1 module, speed , efficiency, productivity, and quality, BEFORE placing quality module in some assembly producing quality module higher than tier 1.

We have no info to be categoric saying one is better than the other.

As for the "trickle of everything", you actually have *higher* start up costs to try to quality cycle intermediates than to cycle the final products. Quality cycling intermediates takes more recyclers and more recyclers per item type since for every, say, quality stack inserter you have to produce 15 green chips, 1 red chip, a blue inserter, and 15 gears in the target quality first. So you could put modules in one recycler and one assembler per tier you want to go up, or you could multiply that infrastructure by 32 times (more than 32 times since the red chip has a craft time greater than 0.5s) to produce the components in quality first at the same rate. And that's in addition to losing free materials from the productivity bonus.

To wit, it takes the same number of cycles through a quality recycling loop to produce any item in quality. As in, you could use the modules to produce a single quality stack inserter chain, single quality t3 assembler chain, single quality speed 3 module chain, single quality quality 3 module chain, single quality prod 3 module chain, single quality electric furnace chain, and so on for 26 more different product types and it would churn out more high quality products, for cheaper, than spending those same modules on a bunch of quality chains for intermediates that could only match the output rate of that first quality stack inserter chain.

That is making strong assumptions with no informations i feel.

[Axs1]

Yes, it is better to obtain the superior qualities only with recipes that return raw materials and thus avoid the entire complicated process just to obtain a specific final product.

You are assuming quality modules are free and have no opportunity cost, no?

Using a quality module in place of a productivity module costs extra materials in lost opportunity. For every 1 quality product you could have produced ~4 extra non-quality products if you'd used an equivalent productivity module instead.

To boot, high quality quality modules are nutty expensive, and best used where the most material possible will be moving through the smallest number of machines, which is the final assembling/recycling loop (in much the same way that productivity modules are best used first in the machine that the most resources go through, the rocket silo).

Sorry for the confusion, I thought I had explained myself clearly.

My point was that it is better to obtain the raw materials in higher quality, because if you do it with the final product you would have to create a system for each product you want to manufacture, while with the raw materials, you manufacture any item and you would use less assembly machine.

To boot, high quality quality modules are nutty expensive, and best used where the most material possible will be moving through the smallest number of machines, which is the final assembling/recycling loop (in much the same way that productivity modules are best used first in the machine that the most resources go through, the rocket silo).

I disagree that the place where the most material will be moving though the smallest number of machine has to be the final step, this highly depend on what you are trying to get as an item.

Imagine you have your first 4 to 10 quality module, would you put them in the inserter assembly ? the roboport ? the armor ? the reactor ? the spidertron ? something for the space platform ?

There so many "last product", maybe you're better off keeping quality module to recycle green circuit( 2 steps of quality module ) and keep legendary iron plate and legendary copper, and legendary green circuit. Then from those material use productivity module up to the end no ? Will allow to make all the different item in trickle without investing 4 module per item.

I consider this solution to be better, because by having the legendary raw material and applying the productivity modules, you directly eliminate the process of creating the legendary product.

I will proceed to try to give a clear example:

Objective: Create Solar Panel

Note: only legendary modules will be used. Iron plates and copper plates are considered raw materials. The cost of a legendary product is assumed to be 56 times more expensive than the normal one.

Ingredients: 5 copper plates, 15 electronic circuit, steel plate.

First with the solution proposed by computeraddic that only applies the quality modules for the final product as shown in the video of Friday Facts #375 - Quality.



Note: Only replace the electronic circuit with the final product.

the recipe you need to multiply by 56 times due to the cost to obtain the legendary product.

5 copper plates ----> 280 copper plates
15 electronic circuits ----> 840 electronic circuits

• 45 Copper Wires ----> 2520 copper wires

• • 22.5 copper plates ----> 1260 copper plates

• 15 Iron Plates ----> 840 Iron Plates

5 Steel plates ----> 280 steel plates

• 25 iron plates ----> 1400 iron plates

Raw materials consumed:
Copper plates = 1540
Iron plates = 2240

If we apply the productivity modules to each ingredient (except raw materials) we obtain:

5 copper plates ----> 280 copper plates
7.5 electronic circuits ----> 420 electronic circuits

• 22.5 Copper Wires ----> 1260 Copper Wires

• • 5,625 copper plates ----> 315 copper plates

• 7.5 Iron Plates ----> 420 Iron Plates

3.33... Steel plates ----> 186.66 steel plates

• 16.66... iron plates ----> 933.33... iron plates

Raw materials consumed:
Copper plates = 595
Iron plates = 1353.33...

now we proceed with mmmPi's solution:

5 legendary copper plates ----> 5 legendary copper plates
15 legendary electronic circuits ----> 15 legendary electronic circuits

• 45 Legendary Copper Cables ----> 45 Legendary Copper Cables

• • 22.5 legendary copper plates ----> 22.5 legendary copper plates

• 15 Legendary Iron Plates ----> 15 Legendary Iron Plates

5 legendary steel plates ----> 5 legendary steel plates

• 25 legendary iron plates ----> 25 legendary iron plates

Raw materials consumed:
Legendary Copper Plates = 27.5
Legendary Iron Plates = 40

If we apply the productivity modules to each ingredient (except raw materials) we obtain:

5 legendary copper plates ----> 5 legendary copper plates
7.5 legendary electronic circuits ----> 7.5 legendary electronic circuits

• 22.5 Legendary Copper Cables ----> 22.5 Legendary Copper Cables

• • 5,625 legendary copper plates ----> 5,625 legendary copper plates

• 7.5 Legendary Iron Plates ----> 7.5 Legendary Iron Plates

3.33... Steel plates legendary ----> 3.33... steel plates legendary

• 16.66... legendary iron plates ----> 16.66... legendary iron plates

Raw materials consumed:
Legendary Copper Plates = 10,625
Legendary Iron Plates = 24,166...

Converting it into normal raw material is:
Raw materials consumed:
Legendary Copper Plates = 10,625 ----> 595 Copper Plates
Legendary Iron Plates = 24,166... ----> 1353.33... Iron Plates

Conclusion: The two ways consume the same amount of raw materials, but the process defines which is cheaper, for the first case it only works for the manufacture of a single item, while the second, although the cost is the same, It can be applied to any recipe without the need to make a recursive system for each product you want to make.

Note: I was surprised by the result, I mistakenly believed that mmmPi's method was cheaper. If I have made any mistakes, I would appreciate it if you would let me know, thank you.

[mmmPI]

Note: I was surprised by the result, I mistakenly believed that mmmPi's method was cheaper. If I have made any mistakes, I would appreciate it if you would let me know, thank you.

There was spreadsheet from bicycleEater on the forum with the costs associated with the probabilities and that yielded a result similar to what you are saying, that the 2 method "cost" the same of raw material, so i'm not so surprised it's not "cheaper" it would be multiplication in different order

Edit : it's in the other thread viewtopic.php?p=593811#p593811

However, the "cost" is not only in raw ressources. There is 2 different notion of time too.

1) the time for the factory to generate the goods. ( more what computeraddict argued about but we don't know how much time is required to recycle)

2) the time for the human to set up the factory. ( that's a personnal choice of having blueprints of loops ready for everysingle item )


I think if you are going to use a lot of quality item the 2) will force you to produce high quality raw material, rather than a loop for every single item, i'm thinking modded game as the extreme case, where you have more "finished product" by that i meant "building that are placed on the map as part of the factory". That would just be too long/ too much of a footprint to have recycler for every single product, some not active all the time which would maybe never produce enough item to pay for the module they host.

The cost is that too, if you have module in intermediate, you can guarantee a higher uptime from them, whereas for every single "finished product", it forces you the player to know which item do you consume the most, if you create a loop for roboport of higher quality, you can't use them if you don't also have the substation loop maybe, multiply that with inserters if you plan a city block blueprint and so on.

That's the same tradeoff when you have furnaces for iron plate that are dedicated to green circuit, but you need steel, and not green circuit, so half your furnaces for iron plate are idle, and you are not happy because they could be helping making more steel but no they are dedicated to green circuit

[Khagan]

To boot, high quality quality modules are nutty expensive, and best used where the most material possible will be moving through the smallest number of machines, which is the final assembling/recycling loop (in much the same way that productivity modules are best used first in the machine that the most resources go through, the rocket silo).

Exactly so. We are all familiar with the priority list for productivity modules. (OK, one can quibble with some of the individual numbers, but the concept is sound.) Quality modules will, to start with, have just the same sort of list, except that your choice is not restricted to intermediate products. In fact, you probably want to avoid intermediate products, precisely because you want to use productivity with those instead.

So what is the best choice of end product to make in your assembler stuffed with rare and expensive quality modules? You want a recipe that processes lots of raw materials, very quickly. Satellites are clearly a contender for top choice. Spiders? Nuclear reactors? Assembler 3s are a less ambitious but probably still reasonable option (only half a second to make).

[mmmPI]

To boot, high quality quality modules are nutty expensive, and best used where the most material possible will be moving through the smallest number of machines, which is the final assembling/recycling loop (in much the same way that productivity modules are best used first in the machine that the most resources go through, the rocket silo).

Exactly so. We are all familiar with the priority list for productivity modules. (OK, one can quibble with some of the individual numbers, but the concept is sound.) Quality modules will, to start with, have just the same sort of list, except that your choice is not restricted to intermediate products. In fact, you probably want to avoid intermediate products, precisely because you want to use productivity with those instead.

So what is the best choice of end product to make in your assembler stuffed with rare and expensive quality modules? You want a recipe that processes lots of raw materials, very quickly. Satellites are clearly a contender for top choice. Spiders? Nuclear reactors? Assembler 3s are a less ambitious but probably still reasonable option (only half a second to make).

But satelites of high quality are of no use are they ? Only to recycle them ? then you make new satelites and you recycle them again and again to farm quality material ?
That's the weirdest example i could think of x), spider nuclear reactor assembler 3 all seem pretty terrible choices to mee because you are not going to use that many of high quality, thats doesn't seem like the item i would want to get, once you have your 50 reactor built, you will not need more in the game will you ? ( even assembler you need less than modules ) would you deconstruct the reactor loop ?
To me it feel like temporarily allocating your modules in a place that will not further boost your module producing ability as much as if you were to place them in the assembler 3 producing quality module ( of hopefully higher quality), which even that, i'm not sure it's THE faster ROI.

Satelite to only recyle them though seem like how you'd get the fastest the material to craft many high quality module. That seem like a weird build to have say 20 assembly making satelites only for recycling, but that's a no brainer for high quality solar pannel or accumulator i think x)

[Axs1]

Note: I was surprised by the result, I mistakenly believed that mmmPi's method was cheaper. If I have made any mistakes, I would appreciate it if you would let me know, thank you.

There was spreadsheet from bicycleEater on the forum with the costs associated with the probabilities and that yielded a result similar to what you are saying, that the 2 method "cost" the same of raw material, so i'm not so surprised it's not "cheaper" it would be multiplication in different order

Edit : it's in the other thread viewtopic.php?p=593811#p593811

However, the "cost" is not only in raw ressources. There is 2 different notion of time too.

1) the time for the factory to generate the goods. ( more what computeraddict argued about but we don't know how much time is required to recycle)

2) the time for the human to set up the factory. ( that's a personnal choice of having blueprints of loops ready for everysingle item )


I think if you are going to use a lot of quality item the 2) will force you to produce high quality raw material, rather than a loop for every single item, i'm thinking modded game as the extreme case, where you have more "finished product" by that i meant "building that are placed on the map as part of the factory". That would just be too long/ too much of a footprint to have recycler for every single product, some not active all the time which would maybe never produce enough item to pay for the module they host.

The cost is that too, if you have module in intermediate, you can guarantee a higher uptime from them, whereas for every single "finished product", it forces you the player to know which item do you consume the most, if you create a loop for roboport of higher quality, you can't use them if you don't also have the substation loop maybe, multiply that with inserters if you plan a city block blueprint and so on.

That's the same tradeoff when you have furnaces for iron plate that are dedicated to green circuit, but you need steel, and not green circuit, so half your furnaces for iron plate are idle, and you are not happy because they could be helping making more steel but no they are dedicated to green circuit

Yes, I find it much better to get the legendary raw material and then craft, as you said, for a recipe that needs red or blue circuit it takes a lot of time and imagine crafting 56 times that amount, while the legendary raw material can be obtained from recipes that are very quick and make the red or blue circuit directly, which saves 56 times the time

[mmmPI]

Yes, I find it much better to get the legendary raw material and then craft, as you said, for a recipe that needs red or blue circuit it takes a lot of time and imagine crafting 56 times that amount, while the legendary raw material can be obtained from recipes that are very quick and make the red or blue circuit directly, which saves 56 times the time

I agree with the reasonning, if you use productivity first and then quality the quantity of material that one has to process is first increasing and then sharply trimmed at the end. Which forces player to manage a lot of transformation process. 1 => 2 => 4 => 8 => 16 => 32 => 64 => 1
Whereas if you use quality first, and then productivity, you sharply reduce first, and then you multiply, 1=> 1/64 => 1/32 => 1/16 => 1/8 => 1/4=> 1/2=> 1

The end result is the same, but if you divide by 64 first , you only have fraction of material to process in your factory. Plus you could decide to allocate the high quality material for any any item with very little extra infrastructure.

But recycling satelittes ... tempting too, though i suppose the satelittes will get modified

[Axs1]

Yes, I find it much better to get the legendary raw material and then craft, as you said, for a recipe that needs red or blue circuit it takes a lot of time and imagine crafting 56 times that amount, while the legendary raw material can be obtained from recipes that are very quick and make the red or blue circuit directly, which saves 56 times the time

I agree with the reasonning, if you use productivity first and then quality the quantity of material that one has to process is first increasing and then sharply trimmed at the end. Which forces player to manage a lot of transformation process. 1 => 2 => 4 => 8 => 16 => 32 => 64 => 1
Whereas if you use quality first, and then productivity, you sharply reduce first, and then you multiply, 1=> 1/64 => 1/32 => 1/16 => 1/8 => 1/4=> 1/2=> 1

The end result is the same, but if you divide by 64 first , you only have fraction of material to process in your factory. Plus you could decide to allocate the high quality material for any any item with very little extra infrastructure.

But recycling satelittes ... tempting too, though i suppose the satelittes will get modified

I am missing a detail that I did not consider, and that is that if you work with legendary raw materials in the final product you can use productivity if applicable, then you reduce the cost of the legendary items by half

[EustaceCS]

^ which once again confirms that Quality feature is a great decision which stimulates creativity on one more level.
The problem is, while the developers basically did teased more or less precise numbers, exact Quality effect list is still due to appear.
And, to a degree I am aware of, ANY information about Quality <-> Liquids interaction (including one at the wiki) is just a speculation.
And, given time to release, numbers ARE subject to change.
So it's probably not worth the effort to crunch exact numbers yet.
But it probably is worth the effort to sum up possible approaches.

From what I can see here, we discussed:
- "as intended" approach of mining resources and filling up manufacturing chain with quality modules to a degree where needed Quality level of end product can appear reliably in satisfying quantities
- "almost as intended" approach where at some point of manufacturing chain ALL products of insufficient quality are recycled
- "as secretly intended" approach where at each step of manufacturing chain ALL products of insufficient quality are recycled
- "not as intended" approach where ALL raw resources are being upgraded to maximum quality available through some Recycler + completely irrelevant recipe combo, and only then used for manufacturing actual products of maximal quality with 0 hassle whatsoever

There's one more variable we (technically, you) overlooked - energy consumption (which translates into either extra space for solars, or extra infrastructure for nuclear, or both).
And it have some very funny implications which might actually simplify the gameplay for casuals like me by ALOT.

I as a really lazy player am considering 4th approach.
Because it has two hidden benefits.
1. One more reason to do ALL refining on-site. Attempt to get highest Quality raw resources possible allows not only starting manufacturing best goodies without making main factory overly complicated, but also utilizing less trains / wagons to haul raw materials around.
2. Since Quality upgrading structures take extra space anyway, unless Energy consumption with Quality modules would get nerfed to death, it might be reasonable to supply mining sites with autonomous energy sources like solar panels too. One less reason to drive around with trusty Jeep and a personal roboport placing long-range electric poles all over the place!

[Khagan]

To me it feel like temporarily allocating your modules in a place that will not further boost your module producing ability as much as if you were to place them in the assembler 3 producing quality module ( of hopefully higher quality), which even that, i'm not sure it's THE faster ROI.

I agree that what you will want when first starting out on the quality journey is more high-quality quality modules. But making modules is slow. So if you only have enough high-quality modules for a few assemblers, an indirect approach is perhaps going to be more productive.

So it's probably not worth the effort to crunch exact numbers yet.
But it probably is worth the effort to sum up possible approaches.

Even 'sum up' is surely premature; I suggest 'start to consider'.

[computeraddict]

But where ? at the very begining first like the green circuit ? they are part of the loop ?
You mean in the assembly first or in the recycler first ?
And then which step of recycling is the better ?
What if recycling is much faster than production ?

If we consider tier 3 quality module, they are made from tier 2 quality module. It's not a given that one would prefer placing its first quality module in the latest step of the chain, instead of recycling the previous. Say recycling tier 1 module is fast ( 5 second ), and making tier 3 module is slow ( 60 second ). Then there would possibly be more material processed in the recycling of tier 1 module rather than in the attempt of making new tier 3 module.

That would mean, it would be better to place the first quality module to recycle every tier 1 module, speed , efficiency, productivity, and quality, BEFORE placing quality module in some assembly producing quality module higher than tier 1.

We have no info to be categoric saying one is better than the other.

There are 20 t1 modules per t3 module. Even not paying attention to the other chips that go into each tier upgrade, if we just assume that recycling time is either constant or proportional to the crafting recipe that it is undoing, the answer is still to focus on uncrafting and recrafting just the final product with the highest bonus quality modules available. A t1 module has 15 seconds of crafting time compared to the t3's 60, meaning there's five times as much t1 module crafting time as there is t3 crafting time. Once you factor in the 30 red and blue chips for each t1->t3 chain as well, there's really just no reason to touch anything less than the t3 step itself for starters. It requires more than five times less quality modules for the same throughput unless the uncrafting recipe times are completely customized, but that seems highly unlikely.

As for recycling potentially being faster than crafting, it's very possible that this is the case. In that case, though, you would just have unmoduled assemblers feeding into a quality moduled recycler until you could afford to start filling the assemblers with quality modules, too. (Realistically you should just have some q1 quality modules in everything to start and replace with higher quality quality modules as they become available, as the math on doing a recycling loop with only a single machine with quality modules is absolutely atrociously wasteful.)

Going over some other items that you might want in quality, they appear to all have much shorter crafting times than their components, especially ones that involve anything more time consuming than gears and green circuits.

[allocation of quality modules based on known crafting times]

That is making strong assumptions with no informations i feel.

What information do I lack? We know the crafting times of all the current items, we know the crafting times of all of their components, we know the probabilities of quality upgrades. Unless the recycling time is truly something wacky, we've got all the information we need to allocate quality modules based on a given set of priorities. Trying to produce a trickle of every intermediate through recycling loops is going to take longer to produce anything of note than the t3 quality module production line is going to take to churn out the first set of 75 t3q1 quality modules.

snip

Sorry for the confusion, I thought I had explained myself clearly.

My point was that it is better to obtain the raw materials in higher quality, because if you do it with the final product you would have to create a system for each product you want to manufacture, while with the raw materials, you manufacture any item and you would use less assembly machine.

I understood you and found no upsides to your method. It doesn't even have laziness going for it, as only producing intermediates doesn't significantly reduce how many builds are required, as there are 10+ components that you can produce interesting final items from. (Iron, copper, steel, batteries, explosives, copper cable, iron sticks, iron gears, low density structures, stone bricks, etc. And if processing units, concrete, and electric engine units always are reset to q0 because of the inclusion of fluid, you'd have to bootstrap quality of anything produced from them directly anyway.)

My point was that with your method you either have no production volume or wind up using more modules than if you just made a build for each thing you were interested in. If you hold the number of modules available constant, your method produces dozen times fewer finished products than just dedicating them all to one device, then manually switching to another later. Each module has to do more work to produce a quality finished product the earlier in the chain it is. It's just like how productivity modules have to do multiple times more crafts to pay for themselves the earlier in the production chain they are.

If you hold the production volume constant, your method requires dozens of times more modules. In both instances your method requires stupendously more raw materials on an ongoing basis, because you're throwing away four free productivity products for every quality product that is produced.

snip

I consider this solution to be better, because by having the legendary raw material and applying the productivity modules, you directly eliminate the process of creating the legendary product.

I will proceed to try to give a clear example:

Objective: Create Solar Panel

Note: only legendary modules will be used. Iron plates and copper plates are considered raw materials. The cost of a legendary product is assumed to be 56 times more expensive than the normal one.

snip

Conclusion: The two ways consume the same amount of raw materials, but the process defines which is cheaper, for the first case it only works for the manufacture of a single item, while the second, although the cost is the same, It can be applied to any recipe without the need to make a recursive system for each product you want to make.

Note: I was surprised by the result, I mistakenly believed that mmmPi's method was cheaper. If I have made any mistakes, I would appreciate it if you would let me know, thank you.

You didn't count how many modules were being used in each process. The "without the need to make a recursive system for each product you want to make" is a goal without a good reason for it. The builds for each are going to look nearly identical, just like how each chunk of current mall looks very same-y.

That's the same tradeoff when you have furnaces for iron plate that are dedicated to green circuit, but you need steel, and not green circuit, so half your furnaces for iron plate are idle, and you are not happy because they could be helping making more steel but no they are dedicated to green circuit

Where modules are allocated is not the same tradeoff as where furnaces are outputting to. Allocating furnace output is a horizontal transfer of resources. Allocating module placement is a vertical one. Moving modules up or down the tiers of production changes their efficacy. Changing where a furnace outputs to does not change its efficacy.

Say you have an iron gear assembler and a roboport assembler. A quality module in the roboport assembler is doing MORE THAN FOURTY-FIVE TIMES as much work when it's working as if you stuck the same module in the iron gear assembler. Even if it's idle 97% of the time, it's still better off in the roboport* assembler than in a 100% uptime iron gear assembler. (Once you factor in the red chips and steel that go into a roboport, the roboport assembler could be idle 99.9% of the time and still be a better use case for the module than the iron gear assembler.)

snip

Yes, I find it much better to get the legendary raw material and then craft, as you said, for a recipe that needs red or blue circuit it takes a lot of time and imagine crafting 56 times that amount, while the legendary raw material can be obtained from recipes that are very quick and make the red or blue circuit directly, which saves 56 times the time

You have it exactly backwards. By putting the modules on the precursors, you take 56 times the time at the step that takes more time. There is no recipe that requires more time for the final step than for the sum of the preceding steps. Even tier 3 modules with their 60 second craft time have less than the sum of their immediate ingredients. Sure you can build a green chip in 0.5 seconds, but if you need 10 of them that's 5 seconds of crafting time. And a lab only takes 2 seconds of crafting time, to give just one example. So you just had your modules work for 5 seconds (plus another 5 seconds for the gears and 1 second for the belts) to make the precursors for a finished product that only takes 2 seconds, wasting 9 seconds of module time.

It gets worse for other buildings. Probably the absolute worst offender for your approach would be a nuclear reactor. A final build time of only 8 seconds, but the 500 red circuits, 500 concrete, 500 copper, and 500 steel that go into it have a cumulative craft time of nearly 4.9 hours.

Whereas if you use quality first, and then productivity

You can't. The number of steps that you can use productivity on is finite. Adding quality at the end does not replace a productivity step. Adding quality early on replaces a productivity step, and that productivity cannot be recovered.

[Axs1]

Note: I was surprised by the result, I mistakenly believed that mmmPi's method was cheaper. If I have made any mistakes, I would appreciate it if you would let me know, thank you.

There was spreadsheet from bicycleEater on the forum with the costs associated with the probabilities and that yielded a result similar to what you are saying, that the 2 method "cost" the same of raw material, so i'm not so surprised it's not "cheaper" it would be multiplication in different order

Edit : it's in the other thread viewtopic.php?p=593811#p593811

However, the "cost" is not only in raw ressources. There is 2 different notion of time too.

1) the time for the factory to generate the goods. ( more what computeraddict argued about but we don't know how much time is required to recycle)

2) the time for the human to set up the factory. ( that's a personnal choice of having blueprints of loops ready for everysingle item )


I think if you are going to use a lot of quality item the 2) will force you to produce high quality raw material, rather than a loop for every single item, i'm thinking modded game as the extreme case, where you have more "finished product" by that i meant "building that are placed on the map as part of the factory". That would just be too long/ too much of a footprint to have recycler for every single product, some not active all the time which would maybe never produce enough item to pay for the module they host.

The cost is that too, if you have module in intermediate, you can guarantee a higher uptime from them, whereas for every single "finished product", it forces you the player to know which item do you consume the most, if you create a loop for roboport of higher quality, you can't use them if you don't also have the substation loop maybe, multiply that with inserters if you plan a city block blueprint and so on.

That's the same tradeoff when you have furnaces for iron plate that are dedicated to green circuit, but you need steel, and not green circuit, so half your furnaces for iron plate are idle, and you are not happy because they could be helping making more steel but no they are dedicated to green circuit

I haven't reviewed the excel he made.

According to what he wrote, the cost is reduced if you apply quality in each step of manufacturing the product, however, it requires many quality modules and 5 assemblers for each ingredient which makes it take up a lot of space.

That would be more appropriate for when you have a large base and even so it doesn't seem like such a good idea to me, because you would do it for each recipe.

Based on the cost that I mention, I am not convinced that it is below 56 times more and I also did my calculations and they are higher

Here:

I have a question, how have they carried out the calculations to obtain that the cost of creating a legendary item is 56 times more or 116 according to a correction they made.

I have tried to do the calculations and it doesn't get me anywhere close to those numbers.

These are my values obtained (Quality modules - Legendary):

Q 2 ------- Q 3 -------- Q 4 --------- Q 5
1.50 ------ 5.45 ------- 19.66 ------ 70.77
1.74 ------ 6.09 ------- 21.70 ------ 77.87
1.74 ------ 5.85 ------- 20.61 ------ 73.68
3.48 ------ 11.00 ------ 34.42 ------ 106.51
3.48 ------ 10.37 ------ 30.59 ------ 89.09 -------------------- (I think these are the closest)


Note: I used google translate, If something is not clear, let me know.