"Eras" of Factorio

[Lallante]

Most people's play of Factorio, especially as a noob, can be categorised in a number of different phases.

In the earliest and shortest phase, the earlygame manual period, the player is largely limited to manually fed, coal powered drills and furnaces with wooden chests. Some players spend a lot of time on this stage, but most rush through it ASAP.

After the first few red science packs are manually crafted and boilers, electricity and steam engines are researched, the player enters the "Coal power electicity phase". Later milestones, which could be cast as phases of the game, include the automation of red and then green flasks, trains, the transition to circuit networks, the transition to solar power and accumulators, the automation of production of blue flasks, the transition to logistic networks, the "robot age" and the point at which one goes on the offensive against biters.

At the moment, research options are overwhelming and often by midgame you are researching tech so fast you have researches stuff you wont ready to use for several hours if at all. At present the main phases of reseach, by contrast with the phases of play, are simply the colour of flasks you have automated.

I would propose making these phases more explicit "Eras" or major milestones - akin to Age of Empires "Ages", and create certain productivity conditions required to progress to the next Era. This could also be done non-linearly (you could skip the rail era or the robot era say). Reaching a new "Era" would unlock the research associated with that Era, which would otherwise be hidden from view (except for an "Era requirements summary" icon).

The "gate" between milestones/Eras would be a level of automated production of specific material sustained for an extended period. For example to progress out of the "manual era" one would need to produce iron plate and copper plate at a rate the equivalent of 4 furnaces each (or 4 and 2) for a period of 5 minutes. To progress to the robot era the goal might be a certain number of electric engines produced per minute over a period of say 20 minutes. This would encourage purposive gameplay, maximising throughput and creating modular factories - the hallmarks of good players. One might need to dismantle and rebuild earlier parts of the factory to meet new throughput thresholds.

This also addresses a common complaint about factorio - there arent enough intermediate goals unless you set them yourself, other than "automate production of everything". This is especially true when it comes to efficiency and throughput - the game doesnt particularly encourage building massive throughput, and its certainly not needed to research faster than you build out, or even to complete the game. Achieving particular, very high, thresholds of throughput for certain items could unlock specialist research for example. There is no need for these unlocks to be linear, nor even to be constrained solely to tech tree unlocking as rewards - other factory wide bonuses could be applied.

This balancing suggestion would work best with the planned tech tree overhaul AND some kind of linear diminishing return in research speed when using multiple labs (see excellent thread on this point here: viewtopic.php?f=16&t=18118)

[ssilk]

This is eventually helpful: A definition of the "eras" by the grand master:
viewtopic.php?f=6&t=2122&p=15749#p15749

[Pappus]

The OP is definitely right a simple setup of e.g. 3 red/green factories running automated (and that is an easy task even for a beginner) let you research things so fast that it simply becomes annoying. What is the point of clicking again and again and again on e.g. laser dmg, turret dmg, laser turret dmg etc etc. Making it a singular bigger research would be a good start and increase the time you spend in the initial stages.

[kaZ]

The charme of factorio that it is not limiting a vast variety of possible playstyles. If someone wants to launch a rocket just using burner miner, why not?! Just imagine LEGO would fix you in building a dollhouse, over and over again . The downside is, what you also mentioned, that it get's "boring and annoyoing" very fast. Just look for some mods like Marathon to enhance your gameplay.
As always the Fun part is not reaching the goal, it's the Way to it (Building Infrastructure, outposts...).and that's why we play . If there was more content for a goal that One would be boring too soon.
You could also visit daniel34's modded server (no belts, no rebuild, marathon). We where just able to install the robonetwork, After playing almost 2 weeks and we're not new to the Game
With the right settings on ressources mp will be very challenging. One of the best Situations to mention: only Laser def, continous attacks ofmany biters, no Access to new fuel and only 3k Coal left . I also bet you will see things you never thougt of. For me the most recent ideaothers came up with using Cars in a different Way like a 2x2 chest or hopping to another Island by putting the Car on the other side

[Hexicube]

As long as there's a button to forcibly unlock that era, I don't see an issue.

I'd also rather not have progress directly tied to production rate, as that's 0 when your factory is backed up. Instead it should probably consider the number of placed machines that meet requirements, such as mining drills on iron ore and furnaces containing iron ore, which would imply a potential production rate of X (in this case, whichever one is lower).

[MalcolmCooks]

The OP is definitely right a simple setup of e.g. 3 red/green factories running automated (and that is an easy task even for a beginner) let you research things so fast that it simply becomes annoying. What is the point of clicking again and again and again on e.g. laser dmg, turret dmg, laser turret dmg etc etc. Making it a singular bigger research would be a good start and increase the time you spend in the initial stages.

Why not just not research things when you don't need to? Nothing is forcing you to always have an active research. You'll also save power and resources which you can use to make other things. When I am in the green science stage I probably actually spend more time not researching than researching, because I am too busy building and exploring with radars to use all the new tech that's available.

[halcyonforever]

Most people's play of Factorio, especially as a noob, can be categorised in a number of different phases.

The "gate" between milestones/Eras would be a level of automated production of specific material sustained for an extended period. For example to progress out of the "manual era" one would need to produce iron plate and copper plate at a rate the equivalent of 4 furnaces each (or 4 and 2) for a period of 5 minutes. To progress to the robot era the goal might be a certain number of electric engines produced per minute over a period of say 20 minutes. This would encourage purpose gameplay, maximising throughput and creating modular factories - the hallmarks of good players. One might need to dismantle and rebuild earlier parts of the factory to meet new throughput thresholds.

I just want to point out that this method of "gating" was used before. Sim City(2013) used this to unlock next teir's of a building. Sometimes it was a very natural progression, but sometimes they were very out-of-whack. Example being some of the unlocks for oil production required 2 Million revenue in a day. The only way I found that possible was to build up massive stockpiles, wait for the market to peak then sell it in a day. While there is some truth to this tactic it also meant I had to nearly shut down my city because all profits were cut off while I waited for the tanks to fill then empty them and hope I earned enough.

Suffice to say it is one reason I don't play that game. Let's not make the same mistake.

[Khaim]

The OP is definitely right a simple setup of e.g. 3 red/green factories running automated (and that is an easy task even for a beginner) let you research things so fast that it simply becomes annoying. What is the point of clicking again and again and again on e.g. laser dmg, turret dmg, laser turret dmg etc etc. Making it a singular bigger research would be a good start and increase the time you spend in the initial stages.

A research queue partially solves this issue. There's a mod for this, but it's not very clean and I wish there was a queue in the core game.

That said, yes, I think many of the techs need to be merged together. For example there are something like 6 robot techs you need to research to unlock anything useful. If you only have a few techs then all you can do is stockpile intermediate products for later. This isn't really worth the complexity cost; it would be much better to simply have a single big tech that unlocks everything at once.

[Serenity]

Research coming way faster than you can really use it is definitely an issue. 5 Red, 6 Green and 10 Labs, you can just breeze through it. Same with blue science a bit later. And it's not quite apparent what a resource sink research is until you really think about it. There is too much of a temptation to research stuff because it's there and it (falsely) seems inexpensive. But you probably don't need much of it at the moment. In future games I definitely need to learn to pace myself and only research what I really need. I think I might also scale back the number of assemblers and labs (maybe 3R, 3G and maybe 5 or 6 Labs) and make it expendable

[bobucles]

I do think that research happens too quickly. The change in research mechanics is partially to blame, because you can run parallel research far more quickly. The research tech upgrade also makes a fast process even faster, when it doesn't need to be.

Ultimately research speed depends on how many resources a player chooses to divert to their research. Since you can't build things without first researching them, there is an incentive to get the research portion of your factory built first and foremost. This means you are blazing through research options while you build up the other portions of your base. If you get the research blasted out before the final products are finished, hey guess what. It means it's time to stop getting the other stuff done and build the next leg of the research chain!

[Aru]

You should be able to research at an arbitrary speed, that's in the theme of automation. The more resources and labs you have, the faster it gets done, and I like that. I think many of the complaints about research could be eased merely by allowing us a research queue, to line up tasks all at once, then picking new tasks is not so annoying. In addition, there should be some display near the research bar showing number of active and total powered research labs, and which science packs are missing from the idle labs. (Or rather, a button to open this display.)

And, research efficiency is a pretty much useless upgrade. If it were truly 'research efficiency' (more research per science pack), then it would be kind of mandatory early research. But really, it's just 'research speed', which I don't bother with unless there's nothing else left, because you can achieve the same effect by building more labs.

Perhaps, there should be a logarithmic curve on research speed? Getting more labs speeds research up, but not linearly. After all, there is a limit to how much you can parallelize real-life science. This would also make the research speed upgrade more meaningful, because it's cost/benefit ratio goes down as your income goes up, but it's effect is linear unlike adding more labs. As with quantity of labs, you would be encouraged to ramp up research speed upgrades gradually as your productivity advances. (The numbers would be balanced differently of course.)

I have a large blueprint string that I usually use for research. It was carefully designed, and works well. I know some people would consider that boring, but I enjoyed the designing, and I enjoy reaping the benefits of good design (and I like just looking at it).

[mooklepticon]

Perhaps, there should be a logarithmic curve on research speed? Getting more labs speeds research up, but not linearly. After all, there is a limit to how much you can parallelize real-life science. This would also make the research speed upgrade more meaningful, because it's cost/benefit ratio goes down as your income goes up, but it's effect is linear unlike adding more labs. As with quantity of labs, you would be encouraged to ramp up research speed upgrades gradually as your productivity advances. (The numbers would be balanced differently of course.)

I like this! I've purposely limited myself to 1 or 2 labs, to slow things down. I think a log curve could place a nice soft cap on research speed.

[ssilk]

I also like this idea. But it should be only a decent effect.
AND (!) it must then be possible to research some tech at the same time.
That makes sense then: You can research some tech at once, that will of course take longer, but you spare a lot of resources.

But it complicates research a lot, cause you need to handle, what lab should research what.
:/
I think this makes eventually more problems, than it solves.

[FMTK]

Sorry, but I can't see all that "gets boring at some point" anymore
If factorio bores you, it's not the game for you. It allows you to play it in countless ways, to be creative and to use your brain.
Need restrictions? Define your own.
Bored of midgame? Start building a new factory.
Wanna be creative? Build a "realistic" factory with paths and gates to everything.
Wanna be creative? Build a train world.
Wanna use your brain? Endless possibilities like (incomplete list)
- Minimize the ways for your character to reach depots of all produced items
- Create a factory with themed sections
- Solve problems (not solved yet or in another way) using complex cable networks
Still bored? Combine all the points in the list.
STILL bored? Get another game.

You can spend thousands of hours in factorio without getting bored if it is the right game for you. Be happy about improvements in upgoming versions (a better readable tech tree would be nice), but in the meantime embrace this fantastic game made by a pretty small team with all the great possibilities it gives you.

Edit: I'm at 1253 hours now and wasn't bored a second

[thereaverofdarkness]

And, research efficiency is a pretty much useless upgrade. If it were truly 'research efficiency' (more research per science pack), then it would be kind of mandatory early research. But really, it's just 'research speed', which I don't bother with unless there's nothing else left, because you can achieve the same effect by building more labs.

Perhaps, there should be a logarithmic curve on research speed? Getting more labs speeds research up, but not linearly. After all, there is a limit to how much you can parallelize real-life science. This would also make the research speed upgrade more meaningful, because it's cost/benefit ratio goes down as your income goes up, but it's effect is linear unlike adding more labs. As with quantity of labs, you would be encouraged to ramp up research speed upgrades gradually as your productivity advances. (The numbers would be balanced differently of course.)

I agree. I was thinking that not only should lab efficiency actually increase science pack efficiency, it should be an expensive and slow research option early on, which gradually becomes cheaper and faster as you clear out the rest of the tech tree. This way, if you find yourself late in the game suddenly wanting to speed up your research speed, it's not too late to research lab efficiency, but it's not overpowered to research it in the beginning. In the end, you just save it until you don't have something more pressing to research, and you want to speed everything up down the road.

[Aru]

... I agree. I was thinking that not only should lab efficiency actually increase science pack efficiency, it should be an expensive and slow research option early on, which gradually becomes cheaper and faster as you clear out the rest of the tech tree. This way, if you find yourself late in the game suddenly wanting to speed up your research speed, it's not too late to research lab efficiency, but it's not overpowered to research it in the beginning. In the end, you just save it until you don't have something more pressing to research, and you want to speed everything up down the road.

That's not quite what I meant, I meant keeping the lab speed research exactly as it is (except renaming it to something that isn't misleading, such as "research speed" instead of "research efficiency"), while changing the effect (on research speed) of running additional labs from linear to logarithmic. To truly have a 'research efficiency' upgrade, would be a bit too much of a no-brainer mandatory research (just as the current upgrade is useless), and also somewhat redundant with the far more interesting mechanic of productivity modules in labs (which is allowed).

This change to research labs, would naturally make the lab speed upgrade neither useless nor an immediate mandatory thing, but rather a progression you'd balance against increasing lab count. You'd gradually increase both in tandem as income goes up.

The idea, was that the lab speed research costs more the higher it gets (as it currently does). While adding more labs, which all cost the same amount, each one is effectively slower than the one before it. The first is increasing cost for flat benefit, the second is flat cost for diminishing benefit, so as your income increases you'd want to invest in both.