(Long post warning) Rethinking quality — modular, configurable, no random chances

[Green Cat]

Disclaimer: i used the ai to rephare everything, but AI being Ai and me having comunication issues, there might still be grammer issues or similar....like issues with rations since this is just a suggestion and not a case study or whatever... anyways, sorry for any such issues

TL;DR

Quality in Factorio needs a full overhaul to remove gambling elements and make quality progression logical, modular, and player-controlled instead of random gacha-style chances.

What?

The way quality is obtained was implemented lazy via the 1% chances when it should have been higher for lower tiers. Yet, having the best chance to get lower quality, only happens when you are end game, especially post game. Meaning we can't even use quality unless we intentionally waste time. In that time, we could just ignore quality and unlock Epic and Legendary plus module 3, to actually have a chance to get higher quality, questioning why in between qualities even exist since we never got to use them.

Idea is that uncommon quality items should be accessible once you unlock quality 1, and not have lower odds than getting a legendary post game. (Assembling machine 2 has 2 slots, you start with 2% and can have 4% if lucky and 6% if very lucky, vs any legendary that is from the start higher.) So yeah, as I said, how to get quality was not implemented in a way for people to use quality, but to give an excuse for gambling.

Gambling is an element that many people don't want in Factorio. Do you enjoy waiting and waiting to see if an item randomly turns into higher quality? congrats (/s), you have a gambling addiction. Hope you stay away from gacha games since one day their server will close and you'll lose access to everything you paid for. And this is the point — no one wanted a gacha mechanism in Factorio.

How do we get quality? via a gacha mechanism of gambling, gambling, gambling, that of course involves recycling and getting rid of excess items. The vanilla method of getting quality is that of gacha games, not Factorio.

Any major overhaul mod shows exactly how to get higher mk tier of items. But you know what happens when you combine them with quality? you introduce gambling. Is the ratio the same? yes. Because you see, no one is complaining that it takes a massive amount of resources to get a better item. The issue is not that and never was that. The issue is that we are forced into gambling to get said items, followed by waste management, because of few people who did not seek help for their addiction and insted added it to Factorio, to then go on an pretend "there is no issue, everyone does it", and other mentality you will see and hear, when talking to people with gambling addiction, about their issues and the problems their addiction cause to others. Like ruining Factorio. But hey, use mods, because giving options and alternatives are not happening because how else can we justify that we have no chose but to gamble.... <- mentality of people with gambling addiction.....

Now that hopefully I covered those 2 points, I will come to the actual suggestion... yeah this was all just to point out the situation, blame to all the people with gambling addiction who love the current version.

1st change: quality to not be dependent on space age. Just like how the elevated rails is its own mod, the quality to also be its own, especially since this will allow to add a new tick list.

1st tick list — the very first one is actually the main one that, by default, is actually OFF. Why? because it will say something like "customise quality". Aka if it's off, we will have the current version of the game as it exists, no other settings can be changed (similar to how if you disable pollution or evolution, the sliders can't be adjusted).

If its tick box is checked, we can change the names of quality (sorry, please keep WoW terms to WoW and don't bring them to Factorio), the % of items, and where to unlock quality. Aka a drop list — you select the planet, including unlocked from start.

Same but for recycle, as in, where to unlock it and remove surface condition. Aka people will be able to use recycle even without space age, like overhaul mods that are not compatible with space age, but you still want quality in it and of course the recycle.

Plus, there is another reason why I want the recycle to be able to be accessed sooner. As I said, most of us paid for the expansion in order to have access and use the quality items, but NOT to be forced to gamble and told to use mods if we don't wish to gamble.

So here comes the second overhaul: you add a quality module? Common -> 100% uncommon. And as you can guess, this is kind of too much. This is where the recycle comes in.

The recycle to have a tick box. If this tick box is enabled, the recycle will work very similarly to how the Furnaces work. Just to be clear, I do mean to make it work just like the Furnaces exactly to remove gambling. Aka no, the 25% chance, keep it for yourself. My whole suggestion is about getting rid of gambling.

Idea is: you don't set any recipe. So any item put in the recycle will automatically be used just like how the Furnaces do with ores. This means that stack inserters are even more valuable, plus, combinators via "item is x stack or more", "set filter", and "hand size". See? no gambling and we already use more advanced stuff that the game offers.

So, going forward, to explain why I wrote those 3 terms: the recycle to work like a Furnace that is crafting steel plate. Meaning if you insert one or two iron plates, nothing happens. This is exactly where the "same ratio" but no gambling is. As in, we will need more items (seriously, there are mods that show exactly this).

As I said at the start, quality was implemented lazy. Why? because each item should have had their own ratio. Why is it just as difficult to get an uncommon burner inserter as it is to get an uncommon turret railgun?

Idea is that you put an item in the "tick box on" recycle, it will automatically detect what item it will process. Aka the same item but higher ratio.

For this, we need first of all quality modules. And here is one of the first major changes. There is no % chance. No, what the quality modules will affect is similar to productivity, as in, it will multiply the current item's "value" as an ingredient.

Let me give a simple but exaggerated example of what I mean. We add a quality module 1 that has 1%. Now the recycle will need 100 normal items to give 1 (100%, no gambling) uncommon item. Due to adding more quality modules, we now have 50%. This means you only need 50 to get 1 higher level. And of course, at 100% — 1 normal = 1 uncommon.

However, all this was regarding items that you want to increase the quality via themselves, like iron ore to iron ore. As in, for recipes, things will be implemented differently.

Idea is that when playing, we don't see all these recipes any different from how we see them now. So, if I select a normal quality item, I will need to provide normal quality. If I select uncommon recipe, I need to provide uncommon ingredients.

However, if a quality module is added, how many of said ingredients I need to provide will be increased with the % of quality. And the other change is that despite me selecting a normal quality recipe, the result will be 100% a tier higher (no jumping or otherwise).

So, to give a recap: new recycle tick box on — I put in an uncommon gun turret, it will NOT destroy it, it will ask me for more. So like in case of 2%, I will need to give around 98 uncommon gun turrets, but the result will be a 100% Rare Gun turret.

This way we get rid of any kind of waste management exactly because many don't want gacha mentality in Factorio. You need more or something better? make more, and use more! And like this, we will never have unwanted quality items sitting around until we unlock epic and then legendary exactly because we can upgrade any of them, and we do need more of them for this.

This means we will actually actively build more and more quality items. Especially if the devs realise they added quality for the enemies but the enemies never are quality... as in... we will need quality turrets and more due to the enemy having quality boost.

Any recipe that exists won't be chance on its own. But when we add quality modules, how many we need is increased depending on the quality %. And the end result will be 100% one higher quality item, no junk, no scrap, no nothing because the whole idea is to remove gambling and gacha from Factorio.

We already have waste management forced on us on Gleba, and excess items we need to get rid of on Fulgora and space platform, so yeah, too much is too much. And just to point out, quality on legendary recipes has the same effect as it has right now: nothing, but if you have a mod for higher tiers, then you'll get those.

So, to point out: everything keeps the same ratio, meaning that it's not game-breaking or easy to get higher quality. As mentioned, no jumps. Meaning with a 1% chance, you'll need 10,000 normal items to get 1 rare. And of course, 1,000,000 normal to get 1 legendary. But again, this is for 1%. Since the higher the % is, the fewer items you will need.

As an example, via mod or whatever, 100% quality on the "tick boxed" recycle: 1 normal item, via recycle it 4 items, it will become 1 legendary. Eliminating any gambling element, still making quality hard to access but easier the higher the quality %.

Eliminate any stockpile of quality due to all of them being consumed exactly to get higher rank, including no "ok I need to destroy this turret and now I have its materials I need to waste manage" situation exactly because the new tick box recycle converts same item to itself, exactly to remove the waste management aspect we are forced to do on Fulgora.

And as a reminder, from the start, I said all this is not enabled by default, but you need to tick a "customise quality" that allows you to change how the recycle works, how quality works, the names of the ranks, heck, even add more ranks, and who knows what else. I mean, many mod makers showed how this PAID feature could have been implemented.

I mean, is that not why we always hear "use mods"? because others are able to show that this could have been implemented way differently. But no, we have a 1% chance, because we need gambling... so you either gamble from the start of the game, or waste hours and hours, or just advance the game until you are almost done with the game so that you actually have a chance of getting quality items... yep... very nice... what can I say.

By the way, how about changing the "next quality has only 1% chance"? Like what I wrote. Instead of 100 normal items for 1 uncommon, how about 10 normal items for 1 uncommon item even at 1%? And legendary to need even more items exactly because they are the highest. Aka the ratio of how many lower quality items you need for the higher quality to change according to the rank of the end result quality item.

Why?

Because gambling and gacha mechanics don't belong in Factorio. The quality system was implemented in a lazy and random way that rewards luck instead of engineering, planning, and automation — which is the very essence of Factorio.

This overhaul removes gambling completely, keeps ratios logical and balanced, allows more control through configuration, encourages building more instead of recycling randomness, and fits the game’s design philosophy. It gives value to modules, automation, and ratios instead of random chances, and removes unnecessary waste management forced by the current system. Plus no more everywhere you look, recycles, recycles, recycles, and more recycles. At least if there was recycle mk2 and mk3 with better speed, more modules, heck, more % default quality (aka the building itself has % quality) exactly because they are used for quality farming. But no, slow, very slow, and low chance, because you wanted quality? want you got, was gacha gambling, where it takes days to get that shinny ultra rare whatever you won't have access once they close their servers so stop forcing WoW and gatcha elements in factorio and seek help for you addiciton insted of justifing that we need gambling elements in Factorio or else it won't be "fun".

[h.q.droid]

I think AI has butchered your article. You keep saying getting rid of gambling. But what you propose seems... identical to the current "gambling" mechanism. It's not the grammar, it's the main content AI has removed.

[crimsonarmy]

This suggestion seems to remove a lot of the challenge of quality. It might be an interesting mod to try out but I don't see why it should be in the base game.

[crimsonarmy]

I think AI has butchered your article. You keep saying getting rid of gambling. But what you propose seems... identical to the current "gambling" mechanism. It's not the grammar, it's the main content AI has removed.

I think the idea is to move from 1% chance to upgrade items to every 100 items one is upgraded. Think about it kind of like if productivity modules gave an x% chance to make an extra item rather than filling up the extra production bar.

[Tertius]

Quality in Factorio needs a full overhaul to remove gambling elements

Quality in Factorio is about statistics, not gambling. It's not gambling, it's statistics. Statistics is about large numbers, and if you use large amounts of items for quality manufacturing, you get reliable, stable and predictable output.

That doesn't mean implementation of quality in Factorio is satisfying. There is space for improving the way to manufacture quality items, but the general approach with statistics is fundamentally workable. In my opinion, creating the required upcycling loops can be improved with tools from the game engine, for example with an upcycling blackbox that inputs large amounts of normal items and outputs a quality item once in a while that's essentially an assembling machine with 100-1000 seconds processing time for one item. I find constructing upcycling factory lines by hand for every item I want as quality is extremely tedious, and in general it's always the same. I lost interest in crafting quality items and started a new map instead.

[quineotio]

TL;DR

Quality in Factorio needs a full overhaul to remove gambling elements and make quality progression logical, modular, and player-controlled instead of random gacha-style chances.

I don't think the problem is the "gambling" aspect - it works fine for uranium processing. But rather it's the number of potential outputs you have to account for any time you use a quality module in anything. It quickly becomes unwieldy.

The system as it stands works ok if you put quality modules into miners and just manage the outputs once - then you build stuff from quality components. And it's ok if you want a small number of quality items to accrue naturally from your mall. But everything in between that becomes tedious.

For example, if you want a quality assembler, you can build something that recycles non quality output, and there are ways to make more efficient systems. But If you want to do the same with every recipe you have to build many similar setups. It's very repetitive.

Another example is if you want to be maximally efficient, which means quality modules in everything. Just to analyze what it would require to build blue circuits:

1. Iron + copper, plus a way to sort 5 levels of quality and recycle excess to prevent jams.
2. Copper wire, same as above, but also 5 different machines for all 5 qualities, and managing any copper that results from the recycling process, which needs to feed back into the first step
3. Green circuits, same as above but with 2*5 outputs from the recycling process, outputs also need to be sorted into 5 piles
4. Red circuits, same as above, also balancing copper wire between red and green circuits, and 5 times plastic sort
5. Blue circuits, same as above, now managing green circuits and red circuits and their components from the recycling loop

And then you need to continue that pattern for anything you build with blue circuits, and everything else. The complexity is exponential but the benefit is relatively small. And you need to consider all 5 outputs even before you unlock all quality tiers if you expect your builds to continue to work after you unlock them.

So basically, I think quality is very flexible in theory (you can do anything with it), but in practice it's quite limited because of its complexity. It's also not particularly useful on Gleba because of spoilage - at least not for science.

[Hurkyl]

So basically, I think quality is very flexible in theory (you can do anything with it), but in practice it's quite limited because of its complexity. It's also not particularly useful on Gleba because of spoilage - at least not for science.

Quality bioflux has a long lifespan, plenty of time to accumulate small batches and then spin up the other materials necessary for combining with them.

But I assert it's not particularly useful for a completely different reason: you get more science value from productivity than you do quality. I think quality science would work better when thought of as a sink for common/low quality stuff when you direct higher quality to other purposes.

The complexity is exponential but the benefit is relatively small. And you need to consider all 5 outputs even before you unlock all quality tiers if you expect your builds to continue to work after you unlock them.

It's not exponential. All of your green circuit assemblers of different qualities don't need their own dedicated red circuit assemblers for output -- they all feed into the same group of red circuit assemblers.

So you only multiply by 5 once, not once for every step in the production chain.

[quineotio]

Quality bioflux has a long lifespan, plenty of time to accumulate small batches and then spin up the other materials necessary for combining with them.

But I assert it's not particularly useful for a completely different reason: you get more science value from productivity than you do quality. I think quality science would work better when thought of as a sink for common/low quality stuff when you direct higher quality to other purposes.

The problem is that the amount of science you get diminishes over time because of spoilage, so in order for quality ag science to be useful you have to build 1,000 (a rocket load) fast enough to overcome spoilage loss, or alternatively design a way to send less than a rocket load, which while not particularly difficult is not worth the effort - it's easier to just expand normal production.

The productivity being better than quality is a different issue, but I don't like it either.

It's not exponential. All of your green circuit assemblers of different qualities don't need their own dedicated red circuit assemblers for output -- they all feed into the same group of red circuit assemblers.

So you only multiply by 5 once, not once for every step in the production chain.

It may not be technically exponential depending on how you look at it, but in practice it's way more complicated for every step. Simplistically you could say that using quality makes things 5 times more complicated, or ten times if you include the additional recycling setup, but the complexity of the routing increases a LOT, and you have more areas where you need to account for jams.

Combine this with the difficulty of knowing beforehand how much production/logistics you'll need before the system is running and it's a nightmare. What I mean is, for example, how much excess of each quality of green circuit will be produced when you account for red circuits being recycled, and blue circuits being recycled into green and red, and then the red into green? Because your ratios WILL be off. Even with consistent demand that would be difficult to calculate - and you need some leeway regardless because of the variability of quality outputs. And the amount of excess green circuits will also determine how many excess copper wire and iron you have, and the amount of excess copper wire effects copper plates.

So in practice, a system that works perfectly fine at a certain level of complexity might break once you expand it, depending on the particular ratio of ingredients for different recipes.

So maybe you can quibble about whether "exponential" is the correct word, but that's only for blue circuits - imagine trying to build your entire factory with quality in every machine. It's like trying to build 5 factories that overlap, with the additional complexity of managing jams and recycling loops. It's not like a normal factory where you can let things sit at the end of the belt - every output of every quality needs to have somewhere to go at all times or else the entire system can jam. Have you ever tried to do it? It's far harder than you might assume.

There's also the additional challenge of how difficult it is to make changes because of all the interlinked routing. With a normal quality base it's relatively easy to move things and expand, but if you have a bunch of quality routed everywhere it's very impractical to make changes.

[Hurkyl]

Combine this with the difficulty of knowing beforehand how much production/logistics you'll need before the system is running and it's a nightmare. What I mean is, for example, how much excess of each quality of green circuit will be produced when you account for red circuits being recycled, and blue circuits being recycled into green and red, and then the red into green? Because your ratios WILL be off. Even with consistent demand that would be difficult to calculate - and you need some leeway regardless because of the variability of quality outputs. And the amount of excess green circuits will also determine how many excess copper wire and iron you have, and the amount of excess copper wire effects copper plates.

So what you need is a system that can balance itself. When I've done quality up the production chain in the past, I usually try to engineer it so that the factory can use backpressure to dynamically figure things like that out.

So maybe you can quibble about whether "exponential" is the correct word, but that's only for blue circuits - imagine trying to build your entire factory with quality in every machine. It's like trying to build 5 factories that overlap, with the additional complexity of managing jams and recycling loops. It's not like a normal factory where you can let things sit at the end of the belt

My conception of holistic quality is that recycling loops should be fairly targeted. IMO the main point of doing quality along the whole production chain is to avoid the 25% cutdown of recycle steps.

To oversimplify, if you need 10 quality rolls to get the desired quality, if you use a recycle loop you have to do the craft-recycle iteration five times, paying the cutdown each time. But if you can get two quality rolls as you go up the chain, you only have to craft-recycle four times. Of course this has to be balanced against the cutdown of using quality modules instead of productivity modules; if that balance goes the wrong way you might not even want to be using quality modules at those intermediate steps!

- every output of every quality needs to have somewhere to go at all times or else the entire system can jam. Have you ever tried to do it? It's far harder than you might assume.

I've done a lot of quality up some production chains, but admittedly I've not done a whole factory design.

(and last time I was running numbers I came away with the impression that productivity was best until you get to the specific loop you want to upcycle on so I stopped thinking too hard about it)

[quineotio]

I've done a lot of quality up some production chains, but admittedly I've not done a whole factory design.

(and last time I was running numbers I came away with the impression that productivity was best until you get to the specific loop you want to upcycle on so I stopped thinking too hard about it)

The difference between a simpler setup and full quality isn't big enough for it to matter in a regular playthrough. And as you said, productivity invalidates a lot of the potential gains. There's only so many quality equipment/machines you need, and then the only purpose of quality is science, which is only worth doing at the steps where you can't use productivity. But you can't only do those steps because you need all your ingredients to be quality, so it's all or nothing.

So it's only something you'd do if you wanted a specific challenge, and personally I think it's too tedious to bother with.

[h.q.droid]

The difference between a simpler setup and full quality isn't big enough for it to matter in a regular playthrough. And as you said, productivity invalidates a lot of the potential gains. There's only so many quality equipment/machines you need, and then the only purpose of quality is science, which is only worth doing at the steps where you can't use productivity. But you can't only do those steps because you need all your ingredients to be quality, so it's all or nothing.

So it's only something you'd do if you wanted a specific challenge, and personally I think it's too tedious to bother with.

Just to add that another problem with the whole-factory design is the speed module incompatibility. Like for military science all 3 intermediates can use quality, but actually putting quality there requires hundreds of assembly plants, on top of the messy output to sort. It's just not worth the tiny gain.

[designdev]

So basically, I think quality is very flexible in theory (you can do anything with it), but in practice it's quite limited because of its complexity. It's also not particularly useful on Gleba because of spoilage - at least not for science.

Quality bioflux has a long lifespan, plenty of time to accumulate small batches and then spin up the other materials necessary for combining with them.

But I assert it's not particularly useful for a completely different reason: you get more science value from productivity than you do quality. I think quality science would work better when thought of as a sink for common/low quality stuff when you direct higher quality to other purposes.

The complexity is exponential but the benefit is relatively small. And you need to consider all 5 outputs even before you unlock all quality tiers if you expect your builds to continue to work after you unlock them.

It's not exponential. All of your green circuit assemblers of different qualities don't need their own dedicated red circuit assemblers for output -- they all feed into the same group of red circuit assemblers.

So you only multiply by 5 once, not once for every step in the production chain.

Interesting point! Do you think there's ever a case where quality does beat productivity for science, maybe in super late-game setups or niche builds?

[quineotio]

Just to add that another problem with the whole-factory design is the speed module incompatibility. Like for military science all 3 intermediates can use quality, but actually putting quality there requires hundreds of assembly plants, on top of the messy output to sort. It's just not worth the tiny gain.

That's a good point. I think they should drop the speed module quality penalty. Quality modules already conflict with productivity, so having them conflict also with speed is very restrictive.

[Panzerknacker]

The randomness factor of quality was one of the system's core properties as pointed out by Kovarex, the lead dev. I don't see how this will be removed from the system. I do think tho that big improvements can be made in terms of mainly balance.

[quineotio]

The randomness factor of quality was one of the system's core properties as pointed out by Kovarex, the lead dev. I don't see how this will be removed from the system. I do think tho that big improvements can be made in terms of mainly balance.

A reverse system might work - where you have recipes specifically for higher quality, but they spit out a percentage of lower quality items. At least then you could set the cap on what quality would emerge. This would be useful on certain things like power poles, where uncommon and rare are useful, but epic and legendary not so much. And it would make quality science more viable, because you could aim for e.g. uncommon science and only have to worry about two quality levels. It would also allow you to use productivity modules.

Doing it this way would basically invalidate the quality module though.

[mmmPI]

So it's only something you'd do if you wanted a specific challenge, and personally I think it's too tedious to bother with.

That's a good definition of quality, "something you'd do if you want a specific challenge". I think by definition the very late-game thing are going to require much more investment. Especially for one thing that's optionnal and designed to potentially generate countless setups with high volume of material beyond what's necessary to win the game. It's the one thing there for players that like this kind of thing, i think if you don't like those it's better to not over-rely on the optionnal mechanic and use it less, rather than advocating for it to be removed or changed from the game.

For the general proposition -1, quality isn't gambling, it sound like someone who haven't found a way to do something claiming the thing is too hard or impossible. ( skill issue ).