Friday Facts #416 - Fluids 2.0

[Sopenas]

The point is very similar as it is with belts vs. trains. In normal gameplay belts could provide more than enough throughput realistically versus trains. When you scale, you want modular setup and trains do that. You absolutely do not want to put pipes into every possible direction and you also need to solve certain things like balancing production of fluid too which requires pumps and tanks with circuits.

Whenever someone asks "whats the point with trains", it makes me instantly think that they do not use trains that much to begin with...

I will just add pipes in my railway tracks section blueprint . These railway tracks are going across all megabase. So wherever my trains goes i will have all fluids. and trains will just transport other materials. i see no issue here with pipes goind in every direction. They will be neatly placed along train lines and that's it

And i do not like it because it removes need of transporting fluids at all

[justAstoro]

The point is very similar as it is with belts vs. trains. In normal gameplay belts could provide more than enough throughput realistically versus trains. When you scale, you want modular setup and trains do that. You absolutely do not want to put pipes into every possible direction and you also need to solve certain things like balancing production of fluid too which requires pumps and tanks with circuits.

Whenever someone asks "whats the point with trains", it makes me instantly think that they do not use trains that much to begin with...

I will just add pipes in my railway tracks section blueprint . These railway tracks are going across all megabase. So wherever my trains goes i will have all fluids. and trains will just transport other materials. i see no issue here with pipes goind in every direction. They will be neatly placed along train lines and that's it

And i do not like it because it removes need of transporting fluids at all

Exactly. Besides quality and new belts. Trains almost useless.

[functional]

The point is very similar as it is with belts vs. trains. In normal gameplay belts could provide more than enough throughput realistically versus trains. When you scale, you want modular setup and trains do that. You absolutely do not want to put pipes into every possible direction and you also need to solve certain things like balancing production of fluid too which requires pumps and tanks with circuits.

Whenever someone asks "whats the point with trains", it makes me instantly think that they do not use trains that much to begin with...

I will just add pipes in my railway tracks section blueprint . These railway tracks are going across all megabase. So wherever my trains goes i will have all fluids. and trains will just transport other materials. i see no issue here with pipes goind in every direction. They will be neatly placed along train lines and that's it

And i do not like it because it removes need of transporting fluids at all

Slight problem here; you could already do this in 1.0. You can easily tile railway tracks with pipes and pumps, especially in vanilla as you only have to deal with few liquids. Not to mention that expansion will involve more liquids; in something like A&B you would deal with too many pipes to realistically be able to do this in the first place.

Oh and yeah, you can also do this with belts already too. In fact it's very easy to do normal vanilla run all the way to the rocket launch with just a belt bus without ever having to use trains. Yet it's rarely argued that cargo (or fluid...) trains are useless.

[Alphasoldier]

Haven't logged into my forum account in forever, but with this latest FFF (and I've read every single one of them, genuinely love the insight they give time and again) I've felt the need to also make my voice heard.

I think that neutering this function into another single pool connected resource is a bad idea.

Fluid mechanics in this game, regardless if it's been inconsistent and sometimes confusing, has always been a challenging and fun experience to deal with. It's an aspect of Factorio that sets it apart from every other resource management, which is a good thing!

I fully understand that with the insane new production boosts that the old system does not function well enough anymore. However, the current solution is nothing short of fully removing a large feature of this game.
Instead, there's two things I've considered, both of which have to do both of your possible solutions.
- The first was about increasing pipe quantities, so why not add a new type of pipe with this exact functionality? Could make them 1x3 or 2x2 (align them with the rail grid), and because they're bigger, they'd also have a larger tile throughput. Maybe even make them airborne like the rails. Though I realize that canonically heavy airborne pipes would not make much sense, perhaps underground instead?
- Second would be a slightly different version of the current solution, but instead of just having one container system with instant transfers to anything connected, calculate the distance fluid needs to go from a source to a requester with the speed depending on pipe saturation as a type of "pressure system". So that instead of having one system, you have a system for each of the individual connections (or even junctions). I know that if you have a lot of connections this could likely exponentially slow things down, but I imagine it optimizes quantity so much better.
- Finally, I just thought of another. An addition to the current system, as an "improved" pressure system, which would be a hybrid of everything that's been discussed, so maybe not the most optimal solution, but I imagine it could be possible be if programmed right and work better mechanically and intuitively.
That is to say, you use the current system until pipes are fully saturated, and when the entire "container" of connected pipes are, the throughput can be maximized well beyond the current transfer limit, making the throughput instantaneous. A push from one side immediately shoving it into the requester. At this point the puzzle/challenge would be making sure that the output into the pipes is always higher or equal to the required source to keep saturation maxed. This way you can also use pumps as one-ways and containers as buffer or a hybrid source/target.

I hope these explanations made sense, and while I imagine that you clever devs have already considered these solutions, perhaps the slightly differing details might make you think about it again in a new light.
I love Factorio, I have for years, I'm genuinely looking forward to playing the DLC, but please please please don't simplify and ruin the fun of the the fluid mechanics this way.

[mcdjfp]

Due to some previous issues this add-on was not an automatic buy. Unless reverted, it is going to take an extreme amount of work (such as other features which would most likely eat up any performance gains) to move this back into the buy category. In fact, this is worse than a not buy add-on, it also cuts me off from future updates to the base game as well.

A static scene is extremely performant, and very boring as well.

There are many games out there, what kept Factorio filling the factory/automation game slot really came down to three things.
1. A fail state (provided by the biters)
2. Random maps (which while each planet is still random the fixed progression paths are a minor negative here as well in the expansion)
3. The fluid system. Not perfect by a long shot (undergrounds performing better and build order dependance), but I found it enjoyable, despite its quirks or maybe because its quirks required thought.

Honestly at this point, playing 2.0 or later Factorio is going to require a massive change, or waiting until a modder fixes it (maybe with a system better than either).

[Terrahertz]

The point is very similar as it is with belts vs. trains. In normal gameplay belts could provide more than enough throughput realistically versus trains. When you scale, you want modular setup and trains do that. You absolutely do not want to put pipes into every possible direction and you also need to solve certain things like balancing production of fluid too which requires pumps and tanks with circuits.

Whenever someone asks "whats the point with trains", it makes me instantly think that they do not use trains that much to begin with...

I will just add pipes in my railway tracks section blueprint . These railway tracks are going across all megabase. So wherever my trains goes i will have all fluids. and trains will just transport other materials. i see no issue here with pipes goind in every direction. They will be neatly placed along train lines and that's it

And i do not like it because it removes need of transporting fluids at all

Exactly. Besides quality and new belts. Trains almost useless.

Everytime I read these "Trains are almost useless"-arguments, I wonder if they did the math.

Yes with stacking and Tiert 4 Belts we get an increase in throughput from 45 Items/s to 240 Items/s.
But this is still well below what even 1.1 Vanilla Trains will do:

• Let's take a train with 8 cargo-waggons transporting ore into the factory, this train will contain 16000 Units of Ore

• Let's assume the train has one way traffic for increased throughput so our tracks will cover a 4 tile wide line (if you want to be pedantic 6 because of signals)

• So we have 4 Belts doing 60 Stacks/s with Stacksize 4 that would be 960 Items/s or 1 of our Ore-Trains every ~17 seconds

• Now I don't know about you, but I can fit a lot more than 1 Train every 17 Seconds on this kind of railway

But wait there is more:

• With Trains you get multiplexing and demultiplexing of goods basically for free

• As Trains transport stacks (or slots because stacks is now something else) and not single items you can also increase throughput be doing a bit of on-site ore-processing

• Also belts are way more expensive, both in UPS and in game resources.

And we don't even know what Quality does to trains, maybe cargo-waggons will have more slots available, maybe locomotives have more power so you don't need as many of them making you trains shorter, enabling you to fit more on the same track.

Now some of you might disaggree, but then tell me what I'm missing, give me numbers not just opinions or feelings.

[functional]

Also belts are way more expensive, both in UPS and in game resources.

This isn't entirely true. Trains are pretty efficient but they aren't able to compete against belts. Trains involve pathfinding and a container (containers also involve UPS loss), belts just take something from one place to another and the algorithm of doing so is extremely efficient after certain changes related to belt topology were implemented. Also direct insertion trains are harder than belts, though entirely feasible and there are bases that do it.

Most UPS-efficient bases will always do direct insertion as much as possible and after that only belts. And most importantly; no buffers or splitters. Typical best performing bases are just modular blocks that you can plop in grids (which again makes the "I can just blueprint some pipes" argument against 2.0 pipes funny; you can just grid the completed game if you wanted to, lol).

Larger trains, faster trains etc will not change this, especially with faster belts.

That all being said, I don't personally play with 100% optimized UPS as I just love the aesthetic of having tons of trains going around. Theres plenty of us to whom Factorio is more of a train simulator with extra steps than a factory simulator. And trains still have plenty of benefits; having 20 pipes running around in B&A doesn't sound very fun (although I know YouTubers who have done that too...)

[Zlutz]

I like this change.
When you realize that whole countries have just a single pipline or two, and millions of people are served from this single pipline - for all intents and purposes, the stuff is "teleported" from source to destination.

[Mooninaut]

I'm extremely pleased that the old fluid handling is going away. Old fluids were basically "magic", very difficult to understand, even harder to fix when problems arose. I much prefer the new functional system that's "unrealistic" over the barely-functional old system that's "realistic-ish-but-not-really".

[Panzerknacker]

To all the people going nuts about the UPS loss or gain with whatever system, I think the old system really deserves some more love there. It's actually pretty brilliant in itself considering how fast it still is compared to how good the simulation plays out in the game. Genius whoever came up with that code.

[mcdjfp]

That is one of the reasons I am so disappointed.

[wretlaw120]

I like this change.
When you realize that whole countries have just a single pipline or two, and millions of people are served from this single pipline - for all intents and purposes, the stuff is "teleported" from source to destination.

and whenever I drive by an IRL refinery complex, I only ever see one pipe per fluid. Its true!

Nuclear reactors, as well, are known for only having one pipe for the whole thing.

[eloepp]

People are still discussing "fun vs. realism" topic while i have a strong feeling it was a deliberate misdirection to distract people from the real change that was made to the game.

agreed. "fun" vs "realism" is just a distraction.

Even if fluids 2.0 is more realistic It's not realistic in the most important way - In real life there are complexities to the discipline of being a Hydraulic engineer. Factorio doesn't need to be realistic and really isn't, the systems emulate real life systems in a superficial way and the game looks for ways to challenge players based off those systems. The only thing that matters is that there will be seemingly little challenge/problem solving with Fluids 2.0, which to me is a bit disappointing. And I have always described Factorio as "problem solving" to people who have never played it, so it seems strange one important system wont include much of that.

[DeaderDisabled]

Personally I think that the realism is a part of the fun. Not sure how hard it would be to implement but maybe have the fluid settings be toggleable? In game toggle would be preferred but if that isn't workable then maybe a way to select between old and new fluid systems before starting the game?

[functional]

so it seems strange one important system wont include much of that.

This is a deliberate lie. There is barely any problem solving with the old mechanic meanwhile the relevant problem solving still exists as is: how to deal with the fact that refineries output three products. Cracking and all that. One thing new system also does is it makes water trains slightly more appealing because now you don't need all the spaghetti to deal with the kind of throughput that nuclear reactors require; you just have one clear buffer network and one consumption network and trains can unload very, very fast into the buffer network.

There's also just the plain fact that with quality, you'd be able to feed only one machine essentially. New system has no limits.

[functional]

Personally I think that the realism is a part of the fun. Not sure how hard it would be to implement but maybe have the fluid settings be toggleable? In game toggle would be preferred but if that isn't workable then maybe a way to select between old and new fluid systems before starting the game?

What do you think is then "realism" in this case? Because if the answer is "old system"; then you'd be wrong. Pressure propagates without having pumps everywhere. Neither new nor old system is very realistic for lots of reasons, and one can hardly say that one is realistic over other. The old one might give an impression of being more realistic because the simulation is somewhat more elaborate due to possible "pulsing" of the liquid in pipes (which also is not realistic at all...)

I personally though wouldn't mind it being toggleable, but you're going to run into some real issues with throughput in expansion. + I think it means all buildings need extra code, but I'm sure some modder will do it.

[mmmPI]

Neither new nor old system is very realistic for lots of reasons, and one can hardly say that one is realistic over other.

Here is a quote from the FFF :

The new system is a fairly large step back in terms of the "realism" of the fluid simulation in Factorio.

onecanhardlylol.jpg (249.38 KiB) Viewed 1413 times

[functional]

Neither new nor old system is very realistic for lots of reasons, and one can hardly say that one is realistic over other.

Here is a quote from the FFF :

The new system is a fairly large step back in terms of the "realism" of the fluid simulation in Factorio.

onecanhardlylol.jpg

And? They're still wrong, because in real world fluid doesn't require millions of pumps to propagate, and teleporting mechanics already existed (that's what underground pipes do; they're a single entity with a single tank). It's not a step back from realism; in terms of realism, quite little changed. Fluid now teleports larger distances, but it also happens to coincidentally model slightly more accurately the fact that pressure actually propagates really, really fast and really, really far. More than a kilometer per second.

If we assume a single meter to be a single tile, that means well over 1000 tiles a second. If you put a single long pipe for 1000 tiles, you will get almost nothing at the other end unless there's a pump behind every pipe. In new system, you will get something there instantly.

In terms of realism, arguing either way is quite moot, as neither is super realistic, but latter is more on par with intuition.

[Builder_K]

I've always thought that the issue with fluids is simply one of visualization; as a player it's really hard to tell at a glance what's going on.

If the game had open "troughs" instead of closed pipes, so that the player could see the amount of fluid in each segment, I think the "old" mechanics would be quite fun to play with.


Perhaps reskin pumps to look like large Archimedes' Screws, and do some shader magic to visualize fluid "flowing in this direction".

Give it that "kids playing in a stream" feel, damming and un-damming the flow.

[Loewchen]

teleporting mechanics already existed (that's what underground pipes do; they're a single entity with a single tank)

No they are not. They are two entities with two fluid boxes with fluid propagation delay and dynamic level drop.