Factorio Forums

Hey everyone. Haven't seen this posted anywhere, sorry if I'm late to the party.

I love this game. It tickles all the right spots in my brain and is one of the few games I can still devote hours to and not regret it.

There is, however, one problem that bugs me every time I play. I can accept the hand-wavery around how everything works (how do the belts get power? How can I reach things so far away? Don't worry about it, it's just a game), but the way science works makes zero sense to me. Why does adding the same gear and plate over and over result in new knowledge? How does adding a battery and steel and a filter inserter 1000x times result in both flying robots and power armor? Why do I need filter inserters and batteries to improve my gun turret damage beyond a certain point? Why do I have all this advanced technology but I apparently can't use it without sacrificing thousands of circuits to the science gods? Does this civilization I belong to not have computers and data storage?

The game is fun, and the system provides a great way to balance gameplay, but it just doesn't make any sense. I end up resenting the process because I feel like I have to fight the science pack system - instead of making things I want, I have to devote tons of space and resources to just science pack manufacture.

I had an idea last night that I'd love to see in a future release. It would be a MASSIVE departure from the way things work currently, so I know a lot of people would oppose it on those grounds. But I think it would make the game more fun and more logical, while retaining the same or more difficulty and allowing the devs to tune difficulty and tech balancing more easily.

What I'm envisioning is throwing out all the science labs and assorted apparatus, including the science packs. Instead, we have a new building - think assembly machine sized, with the same tier progression. Call it a research lab, you could use the same sprite and color it differently for all I care, although in my head the labs connect together - more on that later.

Instead of assembling and delivering science packs, you'd deliver raw materials - the kicker is that you have to deliver the components that make the thing you're researching. The "lore" would be that you need to prototype the new invention until you get it to work. This would also solve the 'knowledge gap' - the idea would be that you already know about the theories behind all this stuff but you need to figure out how to manufacture it on this planet.

So, to start the research process in a new game, you build a Prototyping Lab. Instead of science packs, to unlock logistics you have to deliver it copper plates and gear wheels - skip the red pack manufacture. The plant would have an internal inventory you could dump the items in until you get mass production going. Later you could belt the items over and use inserters to get them in.

Skip ahead a bit. Now you want to unlock steam engines. Instead of being able to make it right away, maybe the devs want to require a bit more teching up (this is just an example). So you now have to manufacture pipes - let's say 300 pipes plus 300 iron plates. You could make them manually and drop them in, or belt them, but you've already used belts for your previous researches and you're out of room. So you plunk down another one, and they connect together. Now you have room for twelve item inputs - three per side, times two labs.

Don't have room to add any more labs stacked together? Or maybe you want to research advanced oil processing and you need to hook up oil directly to the lab? Well, first you'll need a tier 2 Lab that has the pipe input (just like assembling machines 2 - this is also how you increase research speed, the labs would be faster at each tier instead of requiring more research - because they're basically just factories it makes sense that you could build better ones later on). But your oil processing is really remote from your lab area and piping all that up across your factory is virtually impossible. So you plunk down two or three labs near your oil field. But how do they connect? You use a new kind of cable (let's say 4x copper cable + 1 copper plate yields purple cable, as an example) and you can string these 'data cables' across power lines to connect them. Now the labs are effectively the same unit connected by 'ethernet', and you've also introduced the player to cable-laying and network gameplay features which gives you a nice segue into the circuit network stuff.

Uh oh, gun turrets are not effective enough. Time to make a factory area to mass produce gun turrets and firearm magazines and pipe them into the lab for prototyping. The more you can produce, the faster you can unlock better damage and shooting speed.

Time for laser turrets! Better be prepared to belt in your circuits, steel, and batteries! No more unlocking a technology and then realizing you can't use it yet because you are nowhere near prepared to make the components for it. (Seriously, you can research laser turrets super early on but you can't really make them until you are basically ready to make science pack 3 en masse!)

I think this would be a much more natural-feeling way to do science research. I think this could be easily adjusted to retain the same basic difficulty and tiering we have now by balancing the requirements for each tech. It would integrate much more naturally with the existing factory systems, because you're just siphoning off some of the stuff you're already making to do research on it, instead of adding whole new assembly areas for useless items.

Benefits:

1) Solves the logic gap between what we're doing to get research, and the results
2) Allows far more flexible difficulty tuning
3) Encourages the player to build out their factory in step with their research needs
4) Ties game progression to factory scale much more closely and logically
5) Does away with the science manufacturing requirements in favor of requiring factories for each new tech you want to unlock
6) No more useless items - no more filter inserter mass production until/unless you are ready to make them for yourself
7) Could provide a structured in-game tutorial/hint system - instead of having to figure out each item, the lab could just tell you what you need to provide to unlock the research you want
8) Could easily retain existing tech tree GUI and structure for consistency, just change the requirements for the individual items

Drawbacks:
1) Major change in how the game works
2) Would be a massive amount of work to implement
3) Could make the game too hard too quickly, as you would need to make a LOT more assembly areas than you do now
4) Could provide far too much strain on resource production if not balanced correctly

What do you guys think? Sorry I didn't have time to make a sample set of images of what I'm talking about, but I think it's pretty straightforward. Maybe this is too ambitious and would have to wait until Factorio 2 or something, but I'd love to see some future release switch to a system like this.

It sounds like a good idea that's terrible to play. Look at how many actual technologies there are. Sure, some of them would have the same input, but most don't. There's a lot of one-off materials and intermediaries once you get a little ways into the game. Getting each automated sounds like a royal pain. If you need a lot of prototype labs, it'll make sense to keep em in one spot. Setting up a production line for each sounds frustrating, and involving of lots of player time, without awards from automation. (At least as long as you're using belts. once you have robots I can see it being easier). If it doesn't take much and you only need one or two labs that you can move as your production needs move, then it's a similar problem. Still using lots of player time needlessly.

As well, gas products are a major hurdle. You can only ever have one input for it I'd think, which is a problem when you need to switch out different gases. Can only imagine the headache there.

I wouldn't mind it though for a few select researches. Say, weapon research if it continues into infinity. The idea of putting an endless amount of laser turrets into a machine for it to test and figure out how to make it slightly better amuses me.

Actually, the result would be that you build those prototype labs, and replace them with assemblers as soon as your research is done. And build "new" labs at your next production area for the next thing you want to make.

So the actual result will be that you don't have research production at all anymore, you just have a factory that needs "bootstrapping" on assemblers before they produce output.

Simple example, instead of a steel factory, you first build your prototype labs instead of the forges, then when your research is done, you replace them with forges.

I actually really like this idea. The way the current game is set up I set up my base for science then once I have unlocked everything I tear down and rebuild the way I really want it. My only question is can we prototype multiple things at once? I am not talking about red belts and red splitters but more like laser turrets and pump jacks. Also there is no reason that assembly machines need to replace the prototyping machines. Just have the assembly machines eat the resources until it has worked out all the kinks. Once it is done it can simply output the product.

I think this might be a much better way to do science. The current system generalizes the components to allow for mass production and easier balancing(as the devs only have to tweak a few numbers). Your system would require the player to be more flexible as everything has to be automated(even the stuff that currently nobody automates (immediately)), but also the devs as they have to balance with way more components than just a handful of science packs.

The one big downside I can see with the system is that the current technology tree would take much longer to get through as currently trivial research(given enough labs and production) would to significant player time to setup the automation for. This is not an unsolvable problem, but would take more development time to solve.

It sounds like something that might be fun as a one-off gameplay mode or mod, but having to manually reconfigure research production for every single tech or item unlock would get very tedious very quickly as a standard game mode.

Also, there's a dedicated suggestions forum for this kind of topic. Just throwing that out there, in case you missed it.

Nah, this is in no respect more "natural" than the current science system.

The same explanation you gave may be applied to current labs: you as a character do know how much your civilization have discovered, but is highly unlikely, that you'll be able to recreate the complex mechanisms by memory alone. Schematics, blueprint and manufacture instructions are both important and somewhat heavyweight, so before the embarkation the engineer was given a memory stick with all the knowledge of the world compressed by future version of gzip. So the 'science' you do, is actually running unzip algorithm. The encoded data may be different, but the method of extraction is the same, hence the standardization of consumables. The more complicated technologies pack more information and are zipped with other flavours of compression method, hence the science pack tiers.

And a machine that solves the designing problems for varied types of input products gives an impression of a more complex and intelligent automaton than a current laboratory.

Gameplay-wise this actually doesn't bring much: instead of branching some resources to science pack production you now simply funnel your whole factory output through the laboratory area. Devs did increase the variety of scientific packs for 0.15 and I believe there are mods that make laboratories require a gazillion more of science packs of different types, but completely eliminating standartized requirements in favor of individual requests, seems like going too far.

This might be interesting mechanic for endgame infinite bonus researches, however: feed a whole load of items to a high-tier machine (really complex machine, with advanced circuitry and processors and stuff...) and if you're lucky, it will improve your bonus in some occupation. Feed it turrets and bullets and stuff, and it will either give a bonus damage to bullets, or maybe improve turret hp, or (unlickely) improve the mining speed multiplier; give it pickaxes and miners, and discover a bonus to resource extraction speed. Then again... this all could be done with some token items like packs as well...

@Adil
I think it makes far more sense that the engineer has all the knowledge(maybe through cybernetic implant?) than that feeding a system thousands of tons of iron and copper somehow powers a decompression algorithm.

I don't really understand where you are coming from when you say you think this system is going to result in players feeding resources into one big science area just like now(and thus not add much gameplay-wise). Why would people setup an entire production cycle for say filter inserters in their science area only to break it down and move it once the research is done? Why not just build the lab in the spot you want the eventual manufacturing to happen and replace it with an assembler when the research is done? I can understand this technique for research that only requires basic resources early on, but anything more advanced would just be a waste of time I'd say.

Vim Razz wrote:It sounds like something that might be fun as a one-off gameplay mode or mod, but having to manually reconfigure research production for every single tech or item unlock would get very tedious very quickly as a standard game mode.

Also, there's a dedicated suggestions forum for this kind of topic. Just throwing that out there, in case you missed it.

You're still in the mindset of the old system here. You don't have to reconfigure a science area as you're not going to have one for most researches. You just build your automation, like you already do currently after the research is done, and put a lab at the end. Then you can replace the lab with an assembler once research is done.

TripleOmega wrote:

Vim Razz wrote:It sounds like something that might be fun as a one-off gameplay mode or mod, but having to manually reconfigure research production for every single tech or item unlock would get very tedious very quickly as a standard game mode.

Also, there's a dedicated suggestions forum for this kind of topic. Just throwing that out there, in case you missed it.

You're still in the mindset of the old system here. You don't have to reconfigure a science area as you're not going to have one for most researches. You just build your automation, like you already do currently after the research is done, and put a lab at the end. Then you can replace the lab with an assembler once research is done.

If it is like what you said, I don't think it fits Factorio. At the end, you are just extending the gameplay by wasting resources to unlock the next technology. Factorio is about making an efficient factory with many production chains. With the current science system, you need to make additional production chains for the science packs, which is the main goal of the game.

To me, it isn't a problem that technologies can be unlocked by feeding science packs to the labs. There are many games that use the concept of "research points", "skill points" etc. and you earn these points by doing things that totally unrelated to the technology/skill you want to unlock. The science packs in Factorio are just different types of research points.

The thing that bugs me is how the packs are produced. Turning a copper plate + iron wheel into red liquid is fine. But turning an inserter + transport belt into green liquid? hm....But it is just a minor issue. It can be solved by changing the science pack icons to be more solid, e.g. knowledge cubes. Or it is just another example of game vs realism.

TripleOmega wrote:

Vim Razz wrote:It sounds like something that might be fun as a one-off gameplay mode or mod, but having to manually reconfigure research production for every single tech or item unlock would get very tedious very quickly as a standard game mode.

You're still in the mindset of the old system here. You don't have to reconfigure a science area as you're not going to have one for most researches. You just build your automation, like you already do currently after the research is done, and put a lab at the end. Then you can replace the lab with an assembler once research is done.

I didn't say "science area", I said "research production". That was intentional. Your assumptions about my "mindset" are a little off target.

Having to repeatedly interrupt other activities to replace the labs with assemblers and build out new research production in between every layer of tech development is exactly what I was talking about. It seems like it would demand a lot of your time and get old fast, since you're reducing your ability to automate the things and let them run through several layers and/or branches of tech development while focusing your attention elsewhere.

At any rate, I'd need to see a working prototype to be convinced that this might work as a standard game mode. There are a lot of practical little details that would need to be sorted out to make it play out in a fun and engaging fashion, and handwaving those away with "you just do this" or "you just do that" in the absence of a functional model is not compelling.

@Mooncat
This system does remove the production chains of the science packs, so I can see how you'd view that as a significant negative. On the other hand it would open up design space for future additions that are currently held back by the existence of the science pack system. Whether or not that is a worthwhile trade-off is up for debate.

@Vim Razz
I don't understand in what way I misconstrued your mindset. This new system does not have any standard "research production" either as the components you make for research are the same as the components for final production. So any "research production" is really just regular production that temporarily gets fed into a lab before real construction can begin. There would be no steps added to reconfigure anything, you'd just setup your automated production as usual, but add a lab at the end first.

I understand your point about interrupting the flow of gameplay to further research, it is one of the things I also pointed out as requiring attention. Perhaps this system would require a tech-tree with less but more significant elements. On the other hand a wider tech tree that allows for multiple things to be researched at once could also work.

I still think this type of "prototyping" would not even need research labs. For example lets say you want fast inserters. Make an AM and feed it 100 yellow inserters 200 iron and 200 green circuits and wait 50 seconds. Boom fast inserter is unlocked and the machine starts putting out fast inserters. For things like bullet damage you would need to make 100, 200, and 500 regular bullets for levels 1, 2, and 3 Then 4, 5, 6, and 7 could come from AP bullets. Stack inserter bonus could come form making a certain volume of stack and filtered stack inserters.

This would require a tweaking of some recipies to remove dependency's and possibility a coal powered variant of the AM (we will call it AM 0). For example AM0 needs to be able to make AM1 which needs to be able to make AM 2 which needs to be able to make AM 3. You need access to AM 3 before you can make an oil processing plant.

I'm near to move this thread into Suggestions, cause exactly this has been suggested about two years ago.

I think it is an interesting suggestion (and I do not mean that in any negative way!).
I see one problem though
Either you add a lot of incremental items for your top tier items, or you would be able to prototype/research them directly at the start. And it would not make a lot of sense that e.g. prototyping requester chests also gives provider chests and so on. Only things like oilprocessing where all the results are produced together fit that.
So in total I think it might be interesting to playtest such a system I think it would require a large rebuild of practically every item and all recipies etc. so I don't really think it would be practical from a gamplay point of view.
And there are also some Items you rarely care about to massproduce (Who needs more than one or two portable fusion reactors for example?). In those cases it might mean you had to massproduce Items which are supposed to be scarce (I know Alien Artifacts are falling out of the game, but how many would you have to supply to your prototyping before you can build the Item once?).

And it is not like you supply your labs with plates and cogs to research. The abstraction is, that you do the prototyping and testing while producing smaler bits and pieces (the testtubes/beakers). Which is not that bad an abstraction actually. I meand you don't complain about the realism, that in astronomy they convert missiles, radars and satelites into knowledge, do you? And every larger research facility (at least in the practical siences) has at least access to a mechanical workshop. Not only for maintanance but also for building test setups.
Stm

On one hand it would make research less trivial, though experienced players are never wanting for things to research and the cost of science packs is going up. It would certainly change the feel of the game and I don't think it would be a bad change.

On the other hand, there are a bunch of problems. First of all how would you power this "prototyper" to prototype Steam Engines? They at least would need to be available from the start. My other big problem is things like Power Armour, you're basically only ever going to want to create just one of them so pretty much nobody bothers setting up a production for it. Same goes for the modular components. The cost of them is not trivial either so either you would need to be able to balance the prototype cost (so it doesn't need thousands of Processing Units) or adjust the cost of the power armour.

And this feels like a bit of a gimmick, I suspect rather than "better" gameplay what we'll get is "different" gameplay which sounds better to us but only because we're used to the game as it is. So for the massive scale development all we've accomplished is changing gameplay (with a mild-moderate improvement) and a higher barrier to entry for newcomers.

On the other hand I'm pretty sure this would be possible in a mod...

Suggestions that touch core game mechanics ike research cannot be aimed to improve veteran players' experience at the expense of newer players, especially in a game like factorio where so few players make it to end game. In that case this would be godawful, as most players can't even get the automation of blue science. This assumes you're playing lazy bastard style, and never hand craft. With this method players that would've otherwise just automate their research for oil refining now have to build assemblers to crank out refineries that they wouldn't ordinarily need.

And what about products that most people skip just to use as a pre-requisite research (most lower tier modular equipment)?
And how do you take into account previous researches with multiple products (will have to separate them?)?

What you're asking is for older players to nearly double their playtime, and newbies to quintuple their playtime, not to mention all the above issues. That's pretty unacceptable imo, from what I understand the new science system will even make "blue" science easier.

Not every game is don't starve, prototyping just doesn't work here If it makes you feel better, the new research system will at least have science separated by type. E.g., "military science" which takes some military gear, "production science" which takes assemblers and stuff, etc.

https://www.factorio.com/blog/post/fff-159