Ability to craft lower quality with higher quality

[AileTheAlien]

Change the quality system so that you can input items of different quality levels into a machine and the quality of the output is based on the quality of ingredients.

It would need to be structured such that the complexity and amount of the input determines the weight it has on the end product (i.e. something like a flying robot frame, which has a lot of intermediates and is a complicated item should have more weight then say, a copper plate). Same with number - a blue chip takes 20 greens and 2 reds. If I have 2 epic reds, 10 uncommon greens, and 10 common greens, the sheer quantity of greens would bring the end result to being closer to an uncommon than a rare, or even an epic.

I think the number of inputs would be a good enough way to do it, since most recipes don't have such a wide gap in the amounts of different inputs. IMO, it also seems fair logically, since if most of the input is green chips, isn't that where most of the room to make mistakes would come from? Put a powerful engine into a rust-bucket car, and you're still not going to win the race after you crash from shoddy steering.

Haven't even researched Epic and Legendary yet (or Recyclers, because quality recycling loops are kinda boring imo, I'll use them for spidertrons though).

Actually, recycler loops seem even more so like they side-step the complexity of dealing with different qualities, because you get a chance at upping quality during recycling, and not just when you retry the same recipe. So the way you'd deal with quality is just putting in a bunch of low-quality-recyclers right after each assembly line, and re-feeding all the ingredients back into those lines again. The 25%, 6%, 1% etc sizes of the different quality lines would change a bit, but probably not by much since it's fractions of fractions. Everything's then just a local loop, that eventually outputs your highest possible quality, ready for your trains.

So I guess what I'd actually want is just a way to dump items into the sea on Nauvis or Gleba, since Vulcanus has lava, Fulgora has recyclers, and Aquilo has very rare resources. Different lines of quality solved!

[dragon_gawain]

Asteroid collectors getting more arms I had not realized. Sure, that's useful. Cargo bays getting more storage as well. Same with the solar panels (though less valuable IMO). In general, yeah, the stuff for space is useful.
Accumulators storing more power - OK, where am I using that? I don't use accumulators on space platforms, so space isn't really an issue. Maybe a handful of them on Fulgora?
Crafting machines running faster? Why do I care. I can just put down another one. I'm not space limited. (again, the exception is in space, but even then, I can (usually) build more!).
Turrets having longer range.. sure, that's useful I guess, but like, meh? Yeah, I can see the use for Gleba, but also with enough infinite tech you can boost damage up crazy high so you don't even need it.. (but sure, the extra range is appreciated).
Power poles, yes, they're useful. I should've included them in my list. I had issues with quality poles, but the recycler more or less fixes those issues.
Quality inserters... unless you're boosting speed to crazy high amounts, you'll pretty much never need quality inserters. I've been using bulk inserters, and with the exception of foundries, no building of mine has been inserter capped. So like, just.. why?
Drills yeah, I totally forgot about them as well. Quality drills are OP as all heck.

So, sure, quality stuff for space is useful. But, I don't think I'd put quality mods in my main production line for these. For these things, I'd set up a small thing that I craft a machine, if it's too low quality, recycle it with quality mods, and try to tech it up the chain like that. I wouldn't be putting quality mods in my main production line (unless it's a full dedicated quality factory like my big science one).

So, to add to the list of useful quality stuff: asteroid collectors, poles, drills, solar panels.

Machines? Just build more. Solar panels when on land? Just build more.

Also, for the nuclear reactors, yes, they produce more heat, but the fuel consumption goes up proportionally, so they are not any more efficient. Same for the other power producing things (boilers, heat exchangers, heating towers, steam engines, steam turbines, the fusion stuff). Look at the math yourself if you don't believe me, but last I checked a constant consumption of 40.0 MW giving a heat output of 40.0MW was the same ratio as a constant consumption of 52.0 MW giving a constant output of 52.0 MW (looking at the nuclear reactor here). So it's just faster, not more efficient.


That being said, I will look more into the machines and see exactly what they do. I had made some assumptions about what some things would (for example, I has thought that the pumpjacks would just increase speed).
So I'll look into that and see if my opinion on quality machines changes much.

I'm sure part of my negative outlook on quality is that it's new and has some issues, but every factory has some growing pains, so I'm trying very hard to give it a solid chance. My opinion sways a bit every day, but I am warming up to the recycler loop as something fun.

[Tinyboss]

Accumulators storing more power - OK, where am I using that? I don't use accumulators on space platforms, so space isn't really an issue. Maybe a handful of them on Fulgora?

It's nice on Fulgora because making quality accumulators is REALLY easy, and you end up making huge fields of them. But a lot of people DO use accumulators on their space platforms, and it's a game-changer for designs that do. If you know your power output or solar production will fluctuate, you don't need to have enough panels (which take up a lot of room) to match the highest demand, only the average.

A great example is an automated route that includes Fulgora. You charge up your accumulators when you're near other planets, then they support your solar panels when they don't make enough near Fulgora. If you have a solar-only design that makes that trip, you can trade some solars for accumulators and take up less room. Especially since accumulator quality bonuses are much better than the ones for solar.

Crafting machines running faster? Why do I care. I can just put down another one. I'm not space limited. (again, the exception is in space, but even then, I can (usually) build more!).

That's certainly a perspective. But if "I'll just make more machines" is a good answer, then why are beacons and modules on your list of worthwhile quality items? Not only is it an additional bonus, it's MULTIPLICATIVE with the others.

Turrets having longer range.. sure, that's useful I guess, but like, meh? Yeah, I can see the use for Gleba, but also with enough infinite tech you can boost damage up crazy high so you don't even need it.. (but sure, the extra range is appreciated).

Quality turrets are WAY easier to get than enough infinite tech to trivialize stompers.

Quality inserters... unless you're boosting speed to crazy high amounts, you'll pretty much never need quality inserters. I've been using bulk inserters, and with the exception of foundries, no building of mine has been inserter capped. So like, just.. why?

Well, foundries for one. Also EM Plants and fast assemblers with recipes that take huge item counts. Also sushi belts or other situations where you need to manage the inserter stack size.

So, sure, quality stuff for space is useful. But, I don't think I'd put quality mods in my main production line for these. For these things, I'd set up a small thing that I craft a machine, if it's too low quality, recycle it with quality mods, and try to tech it up the chain like that. I wouldn't be putting quality mods in my main production line (unless it's a full dedicated quality factory like my big science one).

Yeah, I agree with this. I didn't try to do quality in my main line, either. I did it for iron ore, copper ore, and plastic, and saved out the quality items for personal equipment. Then I got to Fulgora, set up a large-scale scrap recycling operation, and did the "quality modules to make quality quality modules" loop for a few hours, so I could put epic Q3 modules in all my drills and recyclers. Now I can make nearly anything I want on Fulgora, at any quality I want. Of course, the higher the quality, the lower the quantity I can do. But uncommon is the new common for everything, rares are easy in reasonable quantities for most things, and a stack of epics is in reach for some of the stuff where it's worthwhile. I still have to pick and choose what I make for epic quality.

Also, for the nuclear reactors, yes, they produce more heat, but the fuel consumption goes up proportionally, so they are not any more efficient. Same for the other power producing things (boilers, heat exchangers, heating towers, steam engines, steam turbines, the fusion stuff). Look at the math yourself if you don't believe me, but last I checked a constant consumption of 40.0 MW giving a heat output of 40.0MW was the same ratio as a constant consumption of 52.0 MW giving a constant output of 52.0 MW (looking at the nuclear reactor here). So it's just faster, not more efficient.

Yeah, someone else pointed that out as well. The only benefit for reactors is potentially using fewer, which is only a thing in space and probably not a big deal even there.

[DavidEscott]

I very much agree with this, and think it should be as simple as a toggle on the machine to enable "use (and waste) higher quality items". As the item is inserted it can immediately get downgraded to the recipes quality level.

That would enable us to build a production line with quality modules to try and upgrade intermediate ingredients. Then that line can flow into the next stage and a few assemblers at the start of that stage can grab and utilize the higher quality ingredients they can use while letting the excess quality ingredients pass on to be used in lower quality recipes.

There shouldn't be any bonus to the lower quality recipe, but it would free up having to set up a whole recycler setup to remove excess quality ingredients, which frankly seems a bit odd.

[Rykuta]

1/ productivity-quality exploit. During development there was an option for any quality recipe where a base rolled quality was lowest quality of provided ingredients. There was however exploit when using productivity modules: a machine could be fed normal quality ingredients that would "charge" the bonus crafting progress and when bonus progress was close to finishing a higher quality ingredients were given causing the bonus productivity to give an additional set of products of the higher quality.

2/ weighted quality. If a blue chip would be crafted, which ingredient would have higher weight? Greed chips since there are 20 of them needed or red chips since they are more complex? Also what would it mean that quality is 60% normal 40% uncommon?

I suppose the solve for #1 one would be simply to keep track of the lowest quality ingredient used in any recipe up until the productivity bonus material is completed.

The more complex way (at least in my head) to handle this would be to probably keep track of productivity results of each independent quality output as an independent progress bar. So if you wind up needing 10 crafts to get a bonus craft, and 8 of those crafts wind up with common results, and two with uncommon, you would have a 20% filled uncommon bar and an 80% filled common bar. I /think/ number wise this works out to about the same output as the current implementation of quality?

==================================
I think I made a post at some point about my thoughts on how #2 would probably work though I do not remember which thread it wound up in.
Basically the way I imagined quality working initially was that every quality part would contribute to the quality of the output based on the quantity and not so much the complexity of the part.

So if a recipe required 4 of Material A and 2 of Material B to craft; then all that matters for determining the end quality would have been how many x/6 quality components you would have.

If 2 of the 6 were uncommon and 1 was rare, you would first roll a 1/6 chance to be rare, if that fails you would roll 3/6 to be uncommon (the rare now counting as a potential uncommon) and then if both fail, you wind up with common at 100% since all your parts can be classified as common as their lowest.

If 4 parts were uncommon and 2 rare; you roll the 2/6 to be rare and then you are just guaranteed to have an uncommon output.

In this scenario the percentage of the components is what we are using to determine quality rather than the complexity of the individual components.

Quality modules would simply add an additional roll at the end to upgrade the tier of the output. You could make it so that it only applies this to the lowest quality ingredients in the craft; or just make it output based.

Though answering the question of "Why pick this method instead of some other method for weights? What about complexity of the parts"? I think that I'm not actually sure it matters what methodology you use to weigh each component. You could try weighing them based on some function that takes raw material cost and total crafting time of sub-components and spits out a probability where if you have 100% of all parts being a certain rarity it gives you a 100% chance to at least roll at that level. You could also have it where uses how many intermediate steps are involved in a particular recipe chain to weigh each component. You could also just give each component its own "complexity score" that modifies how it is used in weighting recipes (and the calculation is some quantity * weight function).

Ultimately what is important here is simply that mixing component qualities actually affects the output quality in some fashion. The "How" isn't nearly as important so long as it is something the player can intuit (or just see the probabilities in like a tooltip or something) and design solutions around. Basically just pick something that mathematically works out to how many "rares" from a balance perspective you want getting pumped out of a particular recipe given some particular split of inputs.

So why did I choose the quantity of the parts rather than any sort of complexity weighting? To me it just seemed like the simplest thing to implement and the simplest thing for users to understand at a glance. It also seemed the most in-line with the current implementation as far as how many parts of each quality it would output. (But also my probability math is definitely rusty, so I could be wrong on that point).

[Tinyboss]

I very much agree with this, and think it should be as simple as a toggle on the machine to enable "use (and waste) higher quality items". As the item is inserted it can immediately get downgraded to the recipes quality level.

People (very possibly including me) will screw this up, I guarantee it. Copying assemblers or copying settings from one to another, for example. Or setting up a requester chest wrong, and consuming all the legendary <whatever> in the whole logistic network, when you were only trying to use rares.

[grim]

I think machines should accept items of any quality higher than the quality of the assigned recipe.
This would mean that by default, if quality isn't specified, a recipe would accept inputs of every quality.
The quality of the output item would simply be the lowest quality of the input items.
When setting a recipe there could be a tick-box to allow explicitly demanding all input should be the exact same quality as the recipe, allowing both systems. Or maybe even the full set of greater than or less than comparisons to allow things like "less than or equal to epic quality".

Problems with the current system.

Researching epic/legendary quality can cause working infrastructure to no longer work.
Researching epic quality will break all factories that have only been designed to deal with uncommon and rare qualities, due to belts clogging with epic quality items that have no infrastructure to handle them.
I cannot overstate how bad I think it is that researching something can cause a working factory to stop working, or that you're forced to redesign every single part of your factory that deals with quality just because you want to research something.
When you research productivity modules, old parts of your factory become obsolete and (probably) should be upgraded, but simply researching the new stuff doesn't make the old stuff cease to function at all, it just makes it inefficient. The same should be true for epic/legendary research.

Putting a module into a machine should not have the ability to break parts of a factory.
Putting a quality module into a machine will cause belts to clog with different quality versions of the same item that are unusable without redesigning the factory.
The penalty for not using the module correctly should be a lack of efficiency rather than stopping the production chain entirely. I imagine such a harsh penalty for messing up causes many players to just never engage with quality at all.

A higher quality item can sometimes be less valuable than a lower quality item.
Uncommon intermediate products are often less valuable than normal quality ones because they simply can't be used. If they were usable in lower quality recipes then they could at worst be inserted into the science production chain, but instead they often get recycled, quartering their value.
I know that very advanced factories can deal with excess medium quality intermediates, but I still think one should not need such a complex system just to deal with items that are meant to be "higher quality" than normal items, and such an advanced factory would still outperform dumping excess intermediates into normal quality science production.

[rldml]

At least, it should be an option you may turn on in an assembler just to use every ingredient as if it all have the same quality.

[BlueTemplar]

Duplicate :
Ability to craft lower quality with higher quality

(Welcome to the forums, please use search !)

[coolkau]

After having played with quality for a while I think:

There needs to be a way to downgrade quality:

• Yes boxes fill very slowly, yes they are cheap... But it feels wonky that first you pay for the high quality item and then you cant use it for something cheap and you need to pay for storage?!?
• Trashing: There are no "quality greater than" logistics requests so you constantly have your inventory filled up when you rebuild parts of the factory. Then if you want to push stuff back to storage automatically you need to set up 5 requests per item.
• Development: I like to play wit designs to make them more efficient. It would be nice to just wipe all quality of the outgoing belts until you are sure its all working as expected... specifically without taking out all the modules... I don't want to run after all my qualify items down my main line, I might not have caught if I get it wrong initially. Just a simple safety net

Though I agree by now that given how quality is implemented the downgrading should be explicit / a manual process. Uncontrolled mixing and using various qualities in machines would cause lots of unexpected issues given qualities do not stack, e.g. limiting stacks in boxes, train full condition, ... it really gets a mess very quickly. So they need to be handled with care anyway... Additionally -- considering lowest quality determines the output -- there would be a lot of wasted qualify items even with simple recipes such as steel... (1 in 4 iron plates is low quality, all is low)

Furthermore, having to think carefully about the quality supply chain is actually adding to the complexity. It is quite fun.

Though there are a few things which are really lacking and adding grind:

• Mixed circuit conditions. (Enable station if more than 4k of any quality is in chest)
• Mixed logistic requests (Trash all items with quality less than/ more than)
• Mixed blueprints: Just use the best quality available, update to best quality available (for bulk creation like smelting or miners... ). Slowly scaling up quality is really annoying otherwise...

[BlueTemplar]

Allow Quality Downgrade (=> mod)

[Drundia]

So why did I choose the quantity of the parts rather than any sort of complexity weighting? To me it just seemed like the simplest thing to implement and the simplest thing for users to understand at a glance. It also seemed the most in-line with the current implementation as far as how many parts of each quality it would output. (But also my probability math is definitely rusty, so I could be wrong on that point).

Let's pick red science as an example. You need 1 gear assembler for 10 red science assemblers, but it would contribute the massive 50% to the output quality. Picking highest ROI ingredient for quality will be cheap. On the other hand bumping up quality in recycler is expensive. That's why ignoring quality of ingredients may be a good solution. After all you still need to do fine work to get fine results, fine ingredients are not enough. In real world applications there are many cases where the lowest quality component ends up being a bottleneck and so the final product is only as good as its worst component.

Accepting ingredients of higher quality is solving the problem that recycler is a) expensive and b) a separate technology. Advanced Oil Processing comes with all cracking recipes and you can flush storage tanks anyway. In case of quality there is no way to fix bad ratio of different qualities until recycler is unlocker. And this is the only cases where game punishes you for overproduction. Think about it: punishment for overproducing high quality products.

[BlueTemplar]

Yes, also why I think quality is unlocked too early.

It's not as bad as it would be with petroleum products though, as there are many more things that can be made from most items.

And just making them in the right ratios (with some buffer) is much easier than with the petroleums, especially if you get the idea of not doing cracking, or maybe even just not doing logical wires on cracking (like most players probably fail to do).

(And if you're into extreme solutions, shooting chests is not much different than flushing tanks.)

[Ranakastrasz]

Hmm. If quality was unlocked only on fulgora, then since you get the recycler there you have the built in void option. Honestly, given the surprise 500 science price, it feels like it was intended to have an extra prerequisite anyway, and putting it on fulgora makes sense.

If it was unlocked on fulgora, the planet of way too many byproducts, and an item voider, it would be far less painful.

[BlueTemplar]

I see the logic in that, but that might be too late (especially with Fulgora being the hardest to reach of the secondary planets), and might also make Fulgora feel like the mandatory 2nd planet ?

(Think how Elevated Rail might have been locked to Fulgora, considering how it's so much more important there than other planets — but wasn't.)

[VertebreakHER]

1/ productivity-quality exploit. During development there was an option for any quality recipe where a base rolled quality was lowest quality of provided ingredients. There was however exploit when using productivity modules: a machine could be fed normal quality ingredients that would "charge" the bonus crafting progress and when bonus progress was close to finishing a higher quality ingredients were given causing the bonus productivity to give an additional set of products of the higher quality.

2/ weighted quality. If a blue chip would be crafted, which ingredient would have higher weight? Greed chips since there are 20 of them needed or red chips since they are more complex? Also what would it mean that quality is 60% normal 40% uncommon?

I've been heavily considering this since it seems like the most direct development insight on why an Any Quality recipe or "down-binning" isn't implemented, and I believe its fairly solvable.
I'm considering the differences between "down-binning" (the output is always the lowest available input.)
And "mixed-chance" (the output is a probability determined by the input.)

1/ Assuming mixed quality recipes affect Chance for Quality rather than down-binning, the productivity-quality exploit can be handled by assigning a single value to the output chance of quality. This value would move towards the quality of the last crafted item, or average out the quality level of all crated items, until the productivity bonus progress resets. A value of 1.1 would represent normal quality with a small chance to produce uncommon quality. A value of 5 would be a guaranteed legendary. So in a case where an arbitrary number of normal quality ingredients were used, with one instance of legendary quality ingredients, there would still be potential benefit but not in an exploitable capacity.

Another solution which would apply easily to both mixed-chance and down-binning is to simply make the productivity bonus output the lowest quality that was crafted during that "charge". (Personally I like this option better. Its not as theoretical and easily works with other solutions.

2/ Weight quality in the case of down-binning is a non-issue (as you would simply output the lowest quality in the input.) But in arguing for mixed-chance I believe there's potential for something very interesting.

The easiest way to handle determining the "quality weight" of each item is I believe by setting it manually in the recipe itself. (This would mean manually setting an Assembler 1 to 0.6 circuits, 0.2 iron plate, 0.4 iron gears, or some other arbitrarily decided ratio.) This mean not leaving things at the whim of a problem prone generalized solution, and would allow development to tune individual balance of Any Quality recipes in cases where some are obviously more beneficial than the other. Which brings me to an interesting point:

Rather than using this to try and weight quality as evenly as possible, it could be used to purposely lopside the weight of a particular item in the recipe. If we say for instance that the output chance of an Assembler 3 heavily favors the quality of speed modules over Assembler 2 (or vice-versa), and then expose this information to the player through user interface, the player has the opportunity to consider this outcome when asking themselves where they want to put their quality modules and an additional layer of gameplay is added to the quality system that potentially makes it all the more engaging.

[Rykuta]

So why did I choose the quantity of the parts rather than any sort of complexity weighting? To me it just seemed like the simplest thing to implement and the simplest thing for users to understand at a glance. It also seemed the most in-line with the current implementation as far as how many parts of each quality it would output. (But also my probability math is definitely rusty, so I could be wrong on that point).

Let's pick red science as an example. You need 1 gear assembler for 10 red science assemblers, but it would contribute the massive 50% to the output quality. Picking highest ROI ingredient for quality will be cheap. On the other hand bumping up quality in recycler is expensive. That's why ignoring quality of ingredients may be a good solution. After all you still need to do fine work to get fine results, fine ingredients are not enough.

I'll try to explain what I mean when I say that it seems most in-line with the current implementation.

In the current implementation, in order to get a 100% guaranteed uncommon with a 10% chance to roll into a rare (for our science example); you have to successfully roll two 10% chances (the plate and the gear). Alternatively, you just have to roll a 10% on the end science product. Note that rolling two 10% in a row is actually harder than just rolling the end product. But this is totally worth it right? Because now with the guaranteed uncommon you have a 10% chance to roll a rare. Well, the odds of sequentially rolling two 10% rolls successfully is 1% right? Which as it turns out is the same as the odds of rolling that blue without any of the other modules being uncommon.

Basically, from what I can see, the current implementation doesn't care about the complexity of your end product; its all or nothing. You either have all the parts you need to upscale, or you don't. The odds even out to be basically the same with either path you take to get that higher rarity. Infact arguably it seems like having downstream components producing rarities has a near negligible increase in the quality of the end products.

I actually went ahead and tested this a bit in sandbox mode, I made a building contraption that outputed inserters where every ingredient in the chain had the same quality modules (as well as the end product). And I compared it to a control group of just modules on the end product and let it run until the common output box was filled for both. What I found is basically that the quantity of end products was the same for both +- a few inserters.

So to me, in our red science example you can think of it as instead of being all or nothing, now it's just a gradient scale of probability. Two rolls at 10% being successful in a row means 100% chance for uncommon. One roll means 50% instead of 0%. Since getting that 50% chance required you to roll 50% of the input items at 10%. It seeeemed to me like this all evened out still. However, I recognize this is probably where my ability to do probability math reaches its apex and I feel like I'm probably missing something here. Regardless that is where the source of my suggestion came from.

The reality of course is that any number of solutions is perfectly fine by me. Its all just math at the end of the day. Maybe having 99% of an item's required components caps out at giving a 50% bonus to encourage you to still set up offramps to get that 100% into an end product for 100% chance. Maybe it just balances out the /down-the-line/ chaining logic. Maybe we do use weights to accomplish that too. Any of these will do fine.

Ultimately the goal is to just make it so that putting quality modules on intermediates is worth doing and more accessible than it currently is.

[Ranakastrasz]

Sounds about right. Same way productivity is the same, however you distribute it. Except in the costs of the modules themselves, so certain steps are better.

[Koub]

[Koub] Merged into an older thread with the same suggestion.

[AssaultRaven]

Sounds about right. Same way productivity is the same, however you distribute it. Except in the costs of the modules themselves, so certain steps are better.

Not quite. Unlike with productivity, you can take quality a step further by having parallel production chains with quality modules that kick upgrades to a higher chain at intermediate steps.