Friday Facts #376 - Research and Technology

[Boreas]

Omg, the idea of triggers is HORRIBLE!!!
Why would I want to craft random amounts of res to unlock random tech? That doesn't make any sense and it is highly annoying.
Speedrunners will def have problems with it, as many other players who just put some techs on auto research from the get go. Now it will be impossible to do that, cause i need to "explore first". What if I don't like to explore? I like to use my RADARS and never go outside of my cozy base.

Maybe you can add a toggle for that, but that's laughable: you take away one toggle to put another in place.

[Z01C]

I can only design the game around what I enjoy, and I just don't enjoy having to keep track of a huge number of unique recipes, which all work technically basically the same... Easy to learn, hard to master.

Love it! So glad this is the the general approach by the devs. I'm somewhat surprised by the bigger mods that add tons of new samey recipes that only serve to clutter the crafting window for no real gameplay benefit.

Trigger technologies

Love it! As a scientist myself, I can't think of a single game where research does not feel weird and unnatural. Having to touch the material with which you're meant to make discoveries about is a great addition. Still having "science packs" is fine as a source of general consumables that one goes through in a real lab. It would be neat if labs needed, in addition to science packs, access to the materials they are meant to be researching such as a supply of oil to each lab that needs to research oil-based science or uranium-238 for nuclear-based science or even u-235 for nukes, but I couldn't imagine how this would be implemented to be fun for most other science.

[Kronus_Aero]

The techs to research productivity for specific recipes are not something I'm really interested in, making a my factory run more efficiently by queuing up a bunch of research and leaving the game running while I bugger off isn't especially appealing compared to alternatives that force me to put in actual design effort. Also Just saying, Things like RCUs already benefit disproportionately strongly from productivity modules compared to other late game resource hogs (like electric furnaces for purple science). In particular there is currently a big deficit between Iron and Copper with prod mods (you need 30% more iron than copper for infinite research, with legendary prod mods that gap grows to 70% !) and I feel not properly balanced these kinds of techs could make that deficit worse.

I'm open to the idea, but I need some convincing, really I'd prefer unlocking a new steel recipe that's more efficient, but requires more steps to process or additional resources (in small quantities, so easy to sustain, but still complicates logistics). You achieve the same goal of reducing the resource burden in the late game, but in a way that actually presents a design challenge. As it stands this approach feels lazy, not from a developer perspective, but from a player perspective.

[Zaekyr]

I had to think about this expansion idea for a while. Now though, I realize I do not like it. The space age travelling from planet to planet is fine, but the quality of items seems silly to me. Instead of quality of items why not add a wear value to items so that they would need regular repair from being used and then change construction bots so they could not repair and add new maintenance bots specifically for repair at a much higher tier. Could also expand repair packs so that different types would be needed for different repairs.

[blazespinnaker]

Is it just me, or did kovarex bring some of the RNG h8 on himself with that graphic

[blazespinnaker]

Well they wrote themselves what the point is.

I probably should have clarified, it's an improvement, without a doubt! It's all great. I am literally quivering with anticipation. Ok, not literally, but you know. Take my $$$$$

I post because factorio is awesome. If it wasn't, I really wouldn't bother. Life is too short.

Omg, the idea of triggers is HORRIBLE!!!
Why would I want to craft random amounts of res to unlock random tech? That doesn't make any sense and it is highly annoying.
Speedrunners will def have problems with it, as many other player who just put some techs on auto research from the get go. Now it will be impossible to do that, cause i need to "explore first". What if I don't like to explore? I like to use my RADARS and never go outside my cosy base extra?

We're all just so excited, folks.

[Anachrony]

Why would I want to craft random amounts of res to unlock random tech?

Well then maybe the requirements shouldn't be designed "randomly" then, but with intention. That's like saying that recipe ingredients are random. No, they're not. If one specific requirement doesn't make sense, that would make it a problem with that requirement, not the whole concept of the system.

That doesn't make any sense

This is probably the single worst criticism. Because it absolutely does make sense, more than the status quo does. Researching a bunch of stuff to use some material that you've never even seen yet is backwards video game logic, which is understandable for game purposes but you can't justify that it makes more sense than judicious use of the trigger system.

Speedrunners will def have problems with it, as many other player who just put some techs on auto research from the get go.

So the game should never change, so people can keep repeating the exact same gameplay loop they've memorized? How does that make sense for a 2.0 expansion? Lots of things are going to change, and speedrunners are going to have to relearn tons of stuff if they want to upgrade. It's very common for speedrunners to keep playing old games because they find it interesting to beat that specific configuration, so some speedrunners may choose to keep competing for 1.1 times. But you can't expect to keep 2.0 frozen with no new features so that it won't confuse people who dislike change.

Now it will be impossible to do that, cause i need to "explore first". What if I don't like to explore? I like to use my RADARS and never go outside my cosy base extra?

We don't know the details of exactly what all the checkpoints are going to be. But most players do have to leave their base to set up an oil field already, so the requirement of finding an oil patch to research oil processing isn't some drastic change.

The requirement to visit other planets, on the other hand, is a pretty drastic change, but that's kind of what you're signing on for if you choose to play Space Age. It will be an optional mod, so you can just not play it if you don't like moving.

[TOGoS]

> Quality

Good idea. I think it will make the game a little more interesting. I don't mind the idea of randomness at all, and like that it's something I could choose to ignore, but then later choose to mess with it because it's an interestingly different way to play the game.

The names are fine.

> Discovering new materials

"I think Factorio would be better if the player had more reason to go and explore than just making more mining outposts forever" -- TOGoS, 2019, paraphrased (https://factorio.com/blog/post/fff-309)

So this one gets a thumbs-up from me also.

> Research queue always on

Yeah, probably for the best. I keep trying to click the queue research button in this Krastorio game I've started and am always a little annoyed that it's not there.

> Productivity researches

I never get far enough into the game for this to matter (I launched a rocket once, back in 0.12, and I guess that was enough for me), so no opinion.

[hanli427]

Last week you said there was a +300% production cap on buildings, does this include the production bonus from the research then?
I guess it does, as it justifies diversifying the bonuses as well.

[Boreas]

Well then maybe the requirements shouldn't be designed "randomly" then, but with intention. That's like saying that recipe ingredients are random. No, they're not. If one specific requirement doesn't make sense, that would make it a problem with that requirement, not the whole concept of the system.

I don't see the intention behind "50 iron plates == steam engine". Iron is half of the resources you use in all crafts. I would approve this if you had a lot of metals in this game, like gold or lead. Also, some people are driven not by exploring, but by researching.
One specific req can absolutely mean that it is a concept problem. I've yet to hear a good point against this.

This is probably the single worst criticism. Because it absolutely does make sense, more than the status quo does. Researching a bunch of stuff to use some material that you've never even seen yet is backwards video game logic, which is understandable for game purposes but you can't justify that it makes more sense than judicious use of the trigger system.

If it is not broken -- don't change it!
Status quo makes more sense than adding confusing and clutterous mechanics. Factorio is not an rpg, it's a sandbox: you don't have to pretend your super-genius engineer doesn't know anything about oil. There is something fundamentally wrong with this "immersive mentality". I don't need or want this.

So the game should never change, so people can keep repeating the exact same gameplay loop they've memorized? How does that make sense for a 2.0 expansion? Lots of things are going to change, and speedrunners are going to have to relearn tons of stuff if they want to upgrade. It's very common for speedrunners to keep playing old games because they find it interesting to beat that specific configuration, so some speedrunners may choose to keep competing for 1.1 times. But you can't expect to keep 2.0 frozen with no new features so that it won't confuse people who dislike change.

I never said that.
Even if I meant that, there is nothing wrong with lack of change, especially when you don't have a problem to fix in the first place.
Your disrespect for speedrunners is also not welcome. You want some playerbase to struggle just on your whim.
Point is, there is no need to add extra steps that take time, provide no challenge and give no use to a majority of players.
At least not like this.

Oh, and by the way:
1) game has a tutorial for that purpose
2) players struggle with difficulty spike and not "not knowing what the h an oil spot is"

The requirement to visit other planets, on the other hand, is a pretty drastic change, but that's kind of what you're signing on for if you choose to play Space Age. It will be an optional mod, so you can just not play it if you don't like moving.

I like new content and I welcome change. Give me more content and places to settle. Don't give me a pacifier on a form of a research trigger. In its current form it's humiliating.

[thermomug]

Last week you said there was a +300% production cap on buildings, does this include the production bonus from the research then?
I guess it does, as it justifies diversifying the bonuses as well.

They added a productivity cap property to the building prototype I think (technical stuff). Meaning buildings now CAN have a cap, but for now this will only be relevant for recyclers as far as we know. And also only if mods are used.

Regarding the new FFF:

There are some really good counter-proposals for quality names, but if they are deliberately overlooked we can just edit the names in the translation files anyway...

I think the trigger mechanic is a very effective solution to the posed problem (locking material-related tech). It will feel very intuitive to "touch" a new material to unlock its processing techs. Some people fear that research will become a material grind ... I mean - that's what it is already. Once you get used to the small change, it's great I guess

More infinite researches are super! Research productivity has some very interesting implication for megabases and the super-endgame. Balancing the progression of say mining productivity and research productivity on the pursuit to a megabase is a fun new optimization aspect (I love new complexity !!!).
I would love to know how it scales with levels though? Speaking of research cost scaling... more expressiveness in the formulas would be a great new feature too! (for modders)

One critique I have is the other (steel, blue circuits, ..) productivity researches.
Mining productivity research has always felt wrong to me because it creates a potentially infinite amout of resources out of thin air. It should either
- scale worse than linear, which is ridiculously cheap IMO. So I usually add a small quadratic term to my research scaling via a tiny mod.
- approach a fixed upper limit asymptotically. This suggests that most of the actual valuable material is wasted (say 80%) without research. Then waste is reduced by researching better mining productivity. Not possible at the moment AFAIK
Further productivity researches only make this worse for me. There is just no physical way to create 100 Steel bars out of a single iron sheet. That's magic, but factorio is a engineering-type of game and not about magic.

Great news overall! Cannot wait for next week, as always

[Vulkandrache]

If the requirement is one, then you might as well have the blue circuit tech as a requirement itself.

That misses the point. It doesn't accomplish the same thing at all. The way things are now, you can research not only the things that use blue circuits, but the things that have those techs as prereqs, and the things that have those as prereqs, until you're researched every single tech that has the same research pack requirement, all without having actually used any of the new recipes yet. Having any requirement at all that you actually build something, no matter how small, breaks up the techs under a research pack into smaller segments, and requires you to actually start building some of it before you progress further down those paths.

That is exactly why i dropped Nullius after a few hours.
I want to research everything until the next sciencepack or until i arrive at something i really want to use.
Then i go and build the entire production chain back-to-front, then i plug the ore in.
Forcing me to stop whatever i was doing because i HAVE to set up production for something
i may or may not need 2 hours later is a great way to get me to stop playing.

[fusionfan]

I have seen the unlock on trigger in mods before (e.g. Seablock), certainly has a particular appeal. Will be interesting to see the overall change to early game. Although having to research for mining drills is not my favorite (looking at you, Space Exploration).

[KuuLightwing]

Agreed on recipe productivity research also.

I think having alternative production chain that end up producing more steel would be a more interesting way to introduce a similar concept. It's not unlike having basic oil processing that outputs petroleum gas, or more complex advanced oil processing with cracking that ends up producing more petroleum gas. It will also require to actually design a new production chain instead of making existing one just output more as is or just awkwardly idle more if you already designed it for desired capacity.

[thedoh]

Omg, the idea of triggers is HORRIBLE!!!
Why would I want to craft random amounts of res to unlock random tech? That doesn't make any sense and it is highly annoying.
Speedrunners will def have problems with it, as many other players who just put some techs on auto research from the get go. Now it will be impossible to do that, cause i need to "explore first". What if I don't like to explore? I like to use my RADARS and never go outside of my cozy base.

Maybe you can add a toggle for that, but that's laughable: you take away one toggle to put another in place.

Speedrunner here. I think the casual gameplay should come first. We speedrunners will manage. The feedback I've given about this FFF was from the casual gameplay point of view. I think the triggers are great for early techs to serve as an introduction for new players, guiding them to the "right path." They can greatly benefit new player experience, but overuse, as with middle game techs (like oil) seem bad, but for different reasons.

[Tertius]

I'm still not sold on the implementation of quality. Primarily, in what world is a 90% failure rate acceptable when manufacturing a product?

Try to see it the other way round. And a little differently. The ordinary factory produces proper items. They are all fine and can all be used. But 10% of these items are really good. They can do more, because for some reason they're exceptionally strong or precise. The quality modules are QA testers. They are able to identify such an exceptional item. They put a quality stamp on the item instead of just a "passed QA" stamp.

So we get an item that is better than its input, and we identify this item, because we built in some QA tester into the assembling machine.
If the input was normal, the result is a rare item (Tier 2), and if we use an even better and more thoroughly working QA tester, even beyond that.
This sorting and testing and QA stamping is going forward, quality tier for quality tier, up to legendary.

Quality modules are not enabling an assembling machine to work better in random circumstances. No, they're just a QA tester able identify the better items coming out of that machine because of the general slight randomness coming out of every machine. If we have a bunch of assembling machines of better quality, we know we can drive them a littler faster, we can feed them with a little less input, and we know they're able to handle this, so they actually work better. Without that quality stamp, the engineer builds an ordinary factory that drives them slower and feed more input, even if the machine is better quality - as long as the engineer doesn't see that stamp, he doesn't drive that machine to higher levels, so an unidentified supposedly legendary item will be driven in an ordinary way and never be known to actually be legendary.

If you see it this way, even the proposed naming ("uncommon", "rare", ...) makes sense.

[draslin]

I like how your suggestion does not fix the main issue you have with the implementation

How so? I'm suggesting that at the bottom rung, you go to extra effort to double, triple, quadruple, etcetera "refine " the base raw materials until you have the quality material you need. This is analogous to real world refinement of some materials.

And this approach already has an accepted in game analog in the form of uranium enrichment.

This seems to make much more sense to me than approaching it from the opposite end of the production chain. One doesn't double or triple refine a car, nor does one produce a dozen cars in the hopes of getting a single car with the attributes you want.

Which is the root problem I have with their approach.

As near as I can discern, you don't even make high quality resources to build your high quality products, its just random. My approach has you produce high quality materials at the base level. Anything produced from ore or oil could be refined repeatedly with a random percentage successfully moving up the quality. Thereafter, any products made with those materials are produced as normal but quality would be determined by the input materials. Which are then used to build things further up the chain, but at no point would you build 10 spidertrons, throw away 9 of them, and keep the 1 you were hoping for. Nevermind having bad luck and building 100, throwing away 99, and keeping 1. It makes no sense for an engineer to roll dice in this way.

[draslin]

I'm still not sold on the implementation of quality. Primarily, in what world is a 90% failure rate acceptable when manufacturing a product?

Try to see it the other way round. And a little differently. The ordinary factory produces proper items. They are all fine and can all be used. But 10% of these items are really good. They can do more, because for some reason they're exceptionally strong or precise. The quality modules are QA testers. They are able to identify such an exceptional item. They put a quality stamp on the item instead of just a "passed QA" stamp.

So we get an item that is better than its input, and we identify this item, because we built in some QA tester into the assembling machine.
If the input was normal, the result is a rare item (Tier 2), and if we use an even better and more thoroughly working QA tester, even beyond that.
This sorting and testing and QA stamping is going forward, quality tier for quality tier, up to legendary.

Quality modules are not enabling an assembling machine to work better in random circumstances. No, they're just a QA tester able identify the better items coming out of that machine because of the general slight randomness coming out of every machine. If we have a bunch of assembling machines of better quality, we know we can drive them a littler faster, we can feed them with a little less input, and we know they're able to handle this, so they actually work better. Without that quality stamp, the engineer builds an ordinary factory that drives them slower and feed more input, even if the machine is better quality - as long as the engineer doesn't see that stamp, he doesn't drive that machine to higher levels, so an unidentified supposedly legendary item will be driven in an ordinary way and never be known to actually be legendary.

If you see it this way, even the proposed naming ("uncommon", "rare", ...) makes sense.

Forgive me, but the mental gymnastics you have to do for their approach to make sense seems to me, at best, delusional.

Quality assurance testers don't look for exceptional items, they look for substandard items. They reject items that have failed to make a passing grade. And they do this for the sole purpose of (and bear with me here) quality assurance. To ensure that only the items of the desired specifications make it out to the customer.

The whole point of manufacturing is to produce an item of exacting, consistent, and reproducible specifications with as little waste as humanely possible.

I'm happy to acknowledge people may want and even prefer a more "World of Warcraft" loot mechanic thrown into every level of their factories, but I for one do not and cannot reconcile that concept with the idea that my avatar in game is a competent engineer.

[draslin]

I'm still not sold on the implementation of quality. Primarily, in what world is a 90% failure rate acceptable when manufacturing a product?

There is no failure rate in quality/factorio. No DOA. so, it's 0% failure rate, not 90%.

As for variance in quality, getting it exactly right is nearly impossible.

Think about recall rates on vehicles.

What you generally have in manufacturing is everything that gets produced falls within some standard range of specifications. Getting something perfectly right, would be 'epic' (ugh) indeed.

We're not talking about thousands of an inch variances here, we're talking about significant advantages over lower quality items. If you want a car to run at 100 mph and it can only run at 90mph because of substandard manufacturing, that is a failure. Particularly if what you paid for was a car that could do 100 mph.

[draslin]

I'm still not sold on the implementation of quality. Primarily, in what world is a 90% failure rate acceptable when manufacturing a product?

I don't know if it applies equally to all product types, but the whole quality mechanic makes me think about microchip manufacturing. It's fairly common for microchips to have manufacturing defects, especially at the smallest, bleeding edge sizes. But single defect doesn't necessarily make the whole thing valueless. In many cases they can test what's working and what's not, disable what's not working, and sell it in a different product category depending on how much of it is working.

Fair enough, but that's a little different than producing a spidertron with whole rows of extra inventory and/or more weapons, faster foot speed, etc. You're talking about manufacturing defects associated with the process, in a semi-niche example and failure rates there are much lower than what's propose here. Moreover, they don't set out to make things with a goal that only 10% of the product can be delivered with the intended specifications. Nonetheless, your example does argue, at least in my approach, for some intermediate products to be refined along the lines I've proposed.