Friday Facts #375 - Quality

[morsk]

Will mods be able to change the requirements for quality tiers, or is it always the module system?

I think it would be interesting if there were a unique catalyst used at each upgrade level. Maybe too complex for vanilla, but it would be nice to try accessing "quality" in different ways.

[S_Jacob]

I am hoping that the new Space-driven gameplay is different enough that this quality system will fit into the new play style. It does seem that having a small space factory and factories spread across multiple planets mean that they are a bit more one-off (this one produces A components using B resources while another planet produces C components using D resources), which moves away from using blueprinting for mid- to late-game expansion and focuses on designing new, more complex factory lines. Therefore, having a variety of machines of different qualities might be useful or at least won't harm gameplay like it would in the base game.

I am also glad that it is off by default in the base game as I agree with everyone else that it does not fit the current play style:

1) Quality Upgrades Already Exist - we already have multiple tiers of buildings, researchable bonuses and beacons and modules. This just adds another layer of complexity rather than expanding the existing methods (are we really going to have to figure out Tier 1 Quality 4 vs Tier 3 Quality 2?)

2) Unnecessary Item Complexity - 5 levels of each intermediate material is a nightmare to manage

3) Scrapping Is a Lazy Mechanic - You do not need to scrap anything in the current base game other than a couple of stacks of stone furnaces. I'm totally open to having processes that create by-products but then create a way to refine/use those in other processes. That would be an interesting engineering challenge.

4) Random Chance Doesn't Fit the Game Feel - Getting a better version of almost any item in the game requires additional resources, increased manufacturing complexity and planning, which is the fun of Factorio. The only chance-based item is uranium processing, which would actually be really annoying if we couldn't plug it directly into Kovarex Enrichment.

I am still very excited about the expansion and I know Wube is super responsive to community feedback, so hopefully we see improvements to this along the way!

[exterathor]

I like the idea.

In space exploration it is common to make sushi belt, but not in vanilla.
The recycler will introduce sushi in vanilla.

For megabasing, you will want to build your factory only with legendary items for UPS.
The mall will be a more important part of the factory.
Also, it will be mines dedicated to make high quality ore.

For just launching a rocket, I am not sure quality will be very important.
You probaly will want to have some high quality productivity modules to put in the rocket silo but that is all.
For fighting the bitters, having an high quality tank can be interesting too.

Notice, having high quality factories is a good way to reduce pollution, but efficiency modules are already good at it and simpler to use.

I think quality is a feature for experienced players who want to megabase or are able to make complex optimizations.

For trains and chests i think it will be an option "stack mixed qualities" enabled by default.

Another interesting question, how quality on fluid barrels is managed? It will be a good idea to make them usefull in vanilla.

[bman212121]

I like it.

The best part about it is that it breaks the "solved" equation where everyone just copies everyone else's blueprints and spam down the same stuff a hundred times. Thing is eventually everyone will get there, but this will certainly add a lot more strategy to the mid game as to how to navigate through it. Factorio at it's core is an RTS, and certainly a lot of RTS games don't play out the same way every time.

For the UPS mega basers, think of it as another stage of the game. Build up the mega base, then go through the RNG period where you mass produce all of your top tier equipment. Then once you have sufficient resources stored you can decommission the RNG base all together since you'll have all of the static items leveled up. I don't think there was a mention of it doing anything which would actually benefit science packs, so even if you did feed varying levels of quality into items it's largely irrelevant to your SPM.

The interesting thing will be, as with all RNG you can take a lot of the randomness out at scale. In the classic flip a coin scenario, the longer you flip the coin the closer it will get to 50 / 50 assuming it's not biased. Now flip 100 coins at a time, statistically it's going to get closer to 50 / 50 much faster because the number of iterations increases greatly at each flip. I suspect that if you were building 1,000 machines to generate random parts, the overall randomness will average out and run to run variance will be mostly eliminated. So instead of basic math you better brush off your statistics book and account for standard deviation when building your blueprints.

On the other hand speed runners might want to play the luck lottery (Think SMB3 hand grabs) at small scale because it might actually be worth it to invest a bit of resources to get that luck to make the speedrun faster. (But in doing so takes time and resources to set up, meaning there will be lot more possible strategies available) It could add an early game scenario which would allow time saving, but then at scale I suspect most of the RNG won't change the overall time of the run because you end up making so many items it averages out. It just means there are so many strategies for early game you'd need to figure out the risk / reward, and different teams might take different approaches to it.

[S_Jacob]

On a separate note from above, love that the FFF debates are back! I still remember huge discussion chains especially around 0.11, .12 and .13. For the most part the debates in this community have been open and insightful and it was great that Wube often incorporated our feedback. Excited to have you guys creating something new again!

[ospeovedos]

I'm quite concerned that the inserters can't pick up different quality items at the same time. In the full animation, you can see the inserter on the far left on the bottom row of recyclers occasionally pick up 2 normal quality circuits, then pause since all of the next circuits on the belt are uncommon.

This will be quite bad for a well-optimized setup. A single higher-quality item going down the belt will cause the inserters that pull from it to slow down since they can only carry the same quality in a stack, essentially acting like a non-compressed belt setup... a high-quality item will slow the production.

I don't particularly mind the theory behind this change, but I expect better items to be... better. If they cause the production to slow down compared to not having them, it's not a good idea.

[TheBuzzSaw]

People clearly didn't read the whole thing. This entire system is optional. There's no one with a gun to your head saying you have to use it. Just don't unlock the quality modules and you play the game as if they were never there. Problem solved.

The names need changing 100%, a few suggested name sets are great but do not release this with the generic gacha tier names.

Be real here. The game will be balanced around their existence. It's basically telling players that the only way to go to the next level is to grapple with a huge RNG system. From there, players will now obsess over slowly replacing all their equipment with max quality variants. It will also be EVEN HARDER to reason about crafting speeds/ratios.

This is not a feature that can just sit there quietly in the background. It will overshadow everything we do.

Personally, I would welcome a world gen option that disables quality as a concept from the outset.

[melind]

On the other hand speed runners might want to play the luck lottery (Think SMB3 hand grabs) at small scale because it might actually be worth it to invest a bit of resources to get that luck to make the speedrun faster. (But in doing so takes time and resources to set up, meaning there will be lot more possible strategies available) It could add an early game scenario which would allow time saving, but then at scale I suspect most of the RNG won't change the overall time of the run because you end up making so many items it averages out. It just means there are so many strategies for early game you'd need to figure out the risk / reward, and different teams might take different approaches to it.

I don't think its going to shake out like this. Quality part way through the speedrun is really no different from prod modules, so I doubt it will get used at all. If these modules were available in the speedrun today, they would be used in exactly 1 spot which is crafting the silo which is the worst kind of RNG since its:
1. hugely impactful RNG
2. right at the end of the run
3. completely uncontrollable

It would only make runs worse since you'd end up with like a 10 minute time difference in runs based purely on luck crafting the silo.

[JKv1]

People clearly didn't read the whole thing. This entire system is optional. There's no one with a gun to your head saying you have to use it. Just don't unlock the quality modules and you play the game as if they were never there. Problem solved.

The names need changing 100%, a few suggested name sets are great but do not release this with the generic gacha tier names.

Unlikely to be treated as optional. Would YOU run your factory 100% slower knowing that there is a way to make it go brrr? In my case, I dont have any moral misgivings about cheating a system that wishes to pad out gameplay. No. As I did with the Kovarex enrichment process that produces infinite amounts of U238 so that my nuclear plants can run forever with no form of input control whatsoever, I'll mod my way to 100% legendary components DAY1. I won't see hide nor hair of greens or blues or purples.
Matter of fact, if you bother with those instead of gunning for legendaries immediately, YOU are not only wasting your fictional (ingame) AND actual (your CPU's clock cycles) resources or time, your actually spending watt-hours on no progression AT ALL. Speed Modules Lv3 on Beacons, Prod Lv3 in Assemblers. All of them will be Legendary.

[Rebmes]

:O

This is great! Who'd have thought?

[bman212121]

On the other hand speed runners might want to play the luck lottery (Think SMB3 hand grabs) at small scale because it might actually be worth it to invest a bit of resources to get that luck to make the speedrun faster. (But in doing so takes time and resources to set up, meaning there will be lot more possible strategies available) It could add an early game scenario which would allow time saving, but then at scale I suspect most of the RNG won't change the overall time of the run because you end up making so many items it averages out. It just means there are so many strategies for early game you'd need to figure out the risk / reward, and different teams might take different approaches to it.

I don't think its going to shake out like this. Quality part way through the speedrun is really no different from prod modules, so I doubt it will get used at all. If these modules were available in the speedrun today, they would be used in exactly 1 spot which is crafting the silo which is the worst kind of RNG since its:
1. hugely impactful RNG
2. right at the end of the run
3. completely uncontrollable

It would only make runs worse since you'd end up with like a 10 minute time difference in runs based purely on luck crafting the silo.

The best part is we won't really know until someone puts it to the test. If there is a time save somewhere you can bet it will be used. If there is no time save then like the production modules it will be ignored since it doesn't add any value to the run. If it doesn't add any value then it doesn't really change the run whether it exists or not.

[qwesz9090]

This addition is questionable but I am actually very interested in seeing where this leads.
Pros: Optional, No recipe bloat, It is a unique mechanic that could not be done with mods and can be expanded upon by mods,
It is therfore well thought out and ticks some important boxes a major update should have (according to me).

Cons: I prefer it if the devs focus on interesting options for "horizontal scaling" instead, ITEM BLOAT, sorting 5 different tiers sounds miserable, trying to get everything to tier 5 seems miserable.

And I think this is the core of what makes this feature so divisive. The chances of getting top tier stuff is low so it is supposed to be unreachable for most of the game and an optional challange. But people are already imagining themselves optimising the fun out of the game because they NEED the best tier of everything. It is like if WUBE released a 4 tier assembler that only unlocks if you finish seablock and everyone went mad thinking they must now finish a game of seablock before simply playing the game.

And I somewhat agree with them. Having different tiers on tiers of buildings can induce choice paralysis, there is a reason why people rush tier 3 assemblers and spam that for the rest of the game.

But overall I am hopeful for this feature. I think the thematic idea of having quality matter is great for factorio and the new logistic challanges of handeling different qualities is interesting. I actually dreamt up this concept myself once. ( I laugh at people claiming randomness is "unrealistic". Of course, I still agree with you that this certain implementation feels a bit off.) We don't really have all of the details yet, but the numbers just feel off to me. Stuff can't get worse in quality (?). 5 tiers of qualities is probably too much. The process of producing a high tier product feels wierd. I am afraid it will feel like trying to climb a quality ladder of randomness with lots of intermediates. It doesn't really fit with my idea of "quality" in a factory. Like, doing things fast should result in worse quality. The quality should be more normally distributed and I really don't like the idea of the quality "compounding" over long intermediate product chains.

In my idea of quality, there is no "Legendary" or "Perfect" quality. There is only "bad" or "good enough for its purpose".

If I would redesign the quality probabilties, I would make it 4 quality tiers and raw materials start at tier 3. Producing intermediates will very likely decrease the quality if quality modules of an equal tier are not used and even if max quality modules are used, it will basically have an expected quality change of +-0. A late game factory dedicated focusing on quality would therefore have a distribution around 15% tier 4, 50% tier 3, 30% tier 2 and 5% of tier 1 intermediates ready for use in final products. The quality of the final product is determined by the worst quality input, having sufficient tier quality modules and a dash of randomness, making it important to sort the qualities. Speaking of sorting, I think sorting qualities should be a larger part of the process. Maybe have larger "Quality assesser" building to sort them, as the splitter function kinda trivialises it.

Bonus thoughts:
I think the numbers should be toned down a bit. Halving all current bonuses from quality would make it feel more "optional".
The most fun part of this could be the logistics of distributing different qualities for different uses. I hope they can make low and medium quality stuff feel useful. For example the quality of a stone furnace, ammo or miner shouldn't really make a difference, while the quality of a space suit makes a bigger difference.
I want recipes that requires a certain level of quality! But I totally understand if this is only a modding thing and doesn't make an appearance in the dlc.
I praise the idea of making the recycling machine kinda shitty. It really decentivises the boring "recycle until good" strategy some people have mentioned, which incentivises interesting logistics.
I hope we can toggle seeing the quality icons, since they are really distracting. It would be cool if the quality of intermediates was compleatly hidden from the player and could only be measured using a "Quality assesser" building or something, untill the player unlocked some "Quality assessment vison goggles". Though I totally understand why that is not in the base dlc.

I am excited for this feature! For the variety and because most gripes I have seems to be moddable.

[Foto1988]

Awesome, but please make it toggleable with the start of the game.

[melind]

On the other hand speed runners might want to play the luck lottery (Think SMB3 hand grabs) at small scale because it might actually be worth it to invest a bit of resources to get that luck to make the speedrun faster. (But in doing so takes time and resources to set up, meaning there will be lot more possible strategies available) It could add an early game scenario which would allow time saving, but then at scale I suspect most of the RNG won't change the overall time of the run because you end up making so many items it averages out. It just means there are so many strategies for early game you'd need to figure out the risk / reward, and different teams might take different approaches to it.

I don't think its going to shake out like this. Quality part way through the speedrun is really no different from prod modules, so I doubt it will get used at all. If these modules were available in the speedrun today, they would be used in exactly 1 spot which is crafting the silo which is the worst kind of RNG since its:
1. hugely impactful RNG
2. right at the end of the run
3. completely uncontrollable

It would only make runs worse since you'd end up with like a 10 minute time difference in runs based purely on luck crafting the silo.

The best part is we won't really know until someone puts it to the test. If there is a time save somewhere you can bet it will be used. If there is no time save then like the production modules it will be ignored since it doesn't add any value to the run. If it doesn't add any value then it doesn't really change the run whether it exists or not.

The issue comes up if they do save time and do get used because they will almost certainly be used to create the silo. So what will happen is you play all the way to end of your run, then you have a 10% chance for an upgraded silo. If you win you take 10 minutes off your run, if you lose you take a time loss. Its not fun RNG.

[exterathor]

Interesting.

I already have ideas for mall design.

Every "Normal" item is just send for trash on a sushi belt.
Other items are kept in a chest.

You can feed your mall with specific input belts designed for quality.

This way you can build a factory only with at least uncomon items.

You will only need to make a trash sushi belt.
A trash management system using recyclers.
And enventualy make a a dedicated iron and coper belt designed for quality.

[Ohz]

I was excited until the critics were so hard

Can't wait to simply have a try myself ! After all we love the kovarex enrichment, and this quality system is optionnal anyway

[exterathor]

They didn' t said it will be speed bonus for the silo.

[TheIronRelic]

LMAO has someone on the factorio team been playing borderlands? The quality names and colors match that game almost exactly.

[Eode]

People clearly didn't read the whole thing. This entire system is optional. There's no one with a gun to your head saying you have to use it. Just don't unlock the quality modules and you play the game as if they were never there. Problem solved.

The names need changing 100%, a few suggested name sets are great but do not release this with the generic gacha tier names.

Be real here. The game will be balanced around their existence. It's basically telling players that the only way to go to the next level is to grapple with a huge RNG system. From there, players will now obsess over slowly replacing all their equipment with max quality variants. It will also be EVEN HARDER to reason about crafting speeds/ratios.

This is not a feature that can just sit there quietly in the background. It will overshadow everything we do.

Personally, I would welcome a world gen option that disables quality as a concept from the outset.

The focus on the RNG part feels silly to me. If you are going to produce 100.000+ green circuits then you can very accurately calculate your production with the exact odds you have. RNG won't impact your production much, the more you produce, the lower the effective randomness will be. That's just how it works. The only real gambles will be with expensive/one-off items like MK2 power armor.

[Néomorphos]

I have mixed feelings about the Quality.

On one hand, it looks like I will be able to optimize even more, crunch more numbers (hopefully having game-integrated tools for that btw) and reach higher factory perfection
On the other hand, I am a bit disappointed the late game renewal is relying so heavily on number crunching. Satisfactory expands your crafting palette with alternative recipes etc, which leads to a renewed experience of building a factory for the same thing.
So I am a bit skeptical about the "lots of number crunching" part with almost no other lever of variety

Also, +1 to the naming scheme problem. It will remain functionnaly the same but feel much better.

Lastly, not a fan of the icons. they are visually very similar, and don't convey much feel