Friday Facts #305 - The Oil Changes

[Koub]

Ready to implement in 0.17.61 a mod

Fixed that for you

[zenos14]

Fixed that for you

I would still rather see it in the main game one day, after all it's such a nice solution

[jodokus31]

2) Make con bots not need lubricant somehow
3) Make Petroleum=>Solid Fuel far more efficient (I hate to suggest this, as it pretty much makes light oil worthless, but I dislike not having Heavy/Light oil to turn into fuel more so I think it'd go good with the .17.60 changes)

Hmm, I already feared those kind of proposals. Kind of logical, but.... I'm scared.

[zenos14]

2) Make con bots not need lubricant somehow
3) Make Petroleum=>Solid Fuel far more efficient (I hate to suggest this, as it pretty much makes light oil worthless, but I dislike not having Heavy/Light oil to turn into fuel more so I think it'd go good with the .17.60 changes)

Hmm, I already feared those kind of proposals. Kind of logical, but.... I'm scared.

Yeah, it's not a solution I like particularly
But it lets the "simplified" oil change stay and it (and 2) gets rid the two main issues that makes it a painful change even if it does make heavy/light oil less useful and oil processing itself unsatisfying

[lacika2000]

Ready to implement in 0.17.61 a mod

Fixed that for you

Koub: do you (i.e., Wube) monitor achievements in modded games, to get statistics similar to those vanilla games on Steam?

[crambaza]

2) Make con bots not need lubricant somehow
3) Make Petroleum=>Solid Fuel far more efficient (I hate to suggest this, as it pretty much makes light oil worthless, but I dislike not having Heavy/Light oil to turn into fuel more so I think it'd go good with the .17.60 changes)

Hmm, I already feared those kind of proposals. Kind of logical, but.... I'm scared.

Combine a Regular Engine, and some belts and gears, to make a tank tread construction bot, like Wall-E. They move slow, but don't need electricity. They get enough energy from the Burner Roboport for the one task, abstracted by the Burner Roboport needing to run off of Solid Fuel, made from Pet Gas.

Like a smelter that loses a bit of stored energy for every plate it makes. The Burner Roboport loses a bit of stored energy for every tank treaded construction robot that it sends out.

[xnmo]

Because I finished writing my post 2 minutes after Koub's moderation notice:

Many if not all the suggestions were considered and several of them I even implemented. I already tried to explain the downsides of each of them

And many tried to explain the downsides of Wube's chosen change.

In short, the changes make oil less annoying to set up in the basic form. I hate hearing that this is only for newcomers, I think it’s just generally better, not having to worry about as many things.

So completely gut the green science tech tree and further abstracting Factorio from believable oil processing purely for "convenience's sake," at the cost allowing the option of earlier bots should people choose to pursue them - a textbook example of "dumbing down" a game.

In case of no changes, the basic oil processing was annoying yet completely unsustainable, deadlocking all the time. This lead to just spamming tanks which is a really weak solution, and forced the player to rush advanced oil processing, thus rushing blue science. However with all the things in green the player is generally in a good spot.

I always hear people talking about "rushing advanced oil," "you gotta rush oil" etc. etc. over and over, as if this ought to imply some kind of negative connotation. But then I look at the Factorio tech tree and my immediate thought is "what would I even want besides advanced oil?" It's the best and most immediately useable tech at that point in the game. Everything else - level 2 modules, electric furnaces, even military 3 requires more research and infrastructure and a much longer return on investment than advanced oil. So what if someone wants to "rush" green science for landfill and military 2 instead of automation tech? He can, but it's suboptimal. Someone can research substations and replace every single one of his medium poles with them before researching advanced oil, but it's suboptimal. What's the issue here?

If you want less basic oil deadlocking (which was already entirely manageable, though improvements be made), fix that by shifting recipes around like heavy oil for sulfur or earlier cracking, not gutting a core gameplay mechanic!



But what's the point of rehashing all this over and over, Wube is tired of listening to criticism at this point. It's a waste of time unless they decide to reopen this topic up for discussion with their community in a later FFF.

[Aflixion]

In case the post gets removed from the FFF306 thread

Before I get started, I want to caveat this with the fact that this topic (the new introduction campaign) does deserve discussion, but maybe there were more urgent topics to cover this week?

Are we just going to pretend that nothing controversial at all happened in the last 2 weeks? The last time any FFF thread generated this much discussion this quickly was the proposed (and scrapped) bot nerfs in January/February 2018. The only other FFF threads to pass 20 pages of discussion did so after a few weeks. The last time any single release announcement thread generated this much discussion was 0.12.35, and it took a few weeks to get 15 pages whereas the 0.17.60 thread is already there after 3 days.

It's insulting that the only dev who made any substantial effort to discuss the changes and explain things was V, even though he doesn't fully agree with the changes. It feels like he was sent out as a sacrificial lamb because you knew the changes wouldn't be well-received. The only explanation of why a change was even needed in the first place was "We believe this contributes to make oil a bit less of a difficulty spike", and there was no dev response to the post detailing exactly what the problem was with blue science to begin with (it wasn't oil complexity/difficulty): viewtopic.php?f=38&t=73684&p=445760#p445760. You only addressed alternatives to your proposed changes and said only after the changes were released that "no change" was always an inferior Option A with no explanation of why it was inferior.

This leaves me to conclude that you wanted to make oil simpler to try to appeal to newer players. Players who would otherwise never even try the game because it's not the type of game they want to play. All you did with that was alienate your current playerbase, myself included, because you gave us the impression that our feedback about why a change wasn't needed was ignored.

Even though it was not my idea, I was very significantly responsible for the changes and they would not have happened without my agreement, therefore you can fully blame me for them... and sacrifice me to the dark gods if that is what you desire.

Honestly, there isn’t that much to write further. The proposed solution from last FFF got implemented and the reasons are the same as stated many times. We already presented almost the same thing twice in two weeks, I don’t believe yet another discussion would yield much different results.

Many if not all the suggestions were considered and several of them I even implemented. I already tried to explain the downsides of each of them, and why the no-changes option as well. I’m far from civilization on a phone now so can’t link nor am I going to make a megalong post, but I think the answers are there, though probably lost in the sheer amount of posts. I’d be happy to answer later when I get to a computer if there are further questions.
In short, the changes make oil less annoying to set up in the basic form. I hate hearing that this is only for newcomers, I think it’s just generally better, not having to worry about as many things.
In case of no changes, the basic oil processing was annoying yet completely unsustainable, deadlocking all the time. This lead to just spamming tanks which is a really weak solution, and forced the player to rush advanced oil processing, thus rushing blue science. However with all the things in green the player is generally in a good spot.

Also, I honestly cannot spend another full weekend fully reading the forums and spending several hours writing replies.

BTW your opinions were absolutely not ignored.

I just went through all of your posts on the oil changes from FFF 304 and 305 before 0.17.60 was released. The ONLY mention of not doing any changes at all was this sentence: "Not doing any changes at all of course still has the old problems, like not being able to have self-sufficient long-term BOP and unnecessary tedium/confusion."

BOP is not supposed to be self-sustaining long-term. It's supposed to be an inefficient stepping stone to the more efficient AOP recipe. Deadlocking is supposed to be a risk of staying with BOP too long, that's part of the puzzle. The long post by Antaios that I linked above addresses the tedium/confusion part, and you still have not responded to it.

See why I say it feels like you ignored us? Just saying you didn't ignore us doesn't change the fact that we got absolutely no feedback from you on why scrapping the changes was unacceptable, why the "problems" you described were even issues at all for the starter, inefficient tech.

I do genuinely hope you enjoy your time off. Personally, I'm just going to uninstall the game and wait to see if you're even going to reconsider this change before I think about playing again. I used to recommend Factorio to every one of my friends because of how awesome you guys were as developers, but I honestly can't say I'll continue to do so after the experience of the last 2 weeks.

[Koub]

2) Make con bots not need lubricant somehow

I'd rather see a low tier ground based con bot limited enough so that regular conbots would be a significant upgrade, but that would help kickstart the scaling needed to automate blue science, so that we don't go from mandatory BOP=> AOP to mandatory BOP=> Bots.
Well I'm aware that it's not exactly mandatory, but ... mandatory-ish ? .

I'd see those to be early green science, and could even become a component to regular conbots (like for example : basic bot + flying frame => flying conbot).

When automating science, blue is still a big step in inftrastructure needs. It's not hard per se, but the scaling is tedious when built by hand, withour the QoL of con bots (copy paste, move, ...).

To illustrate, in order to craft 72 science packs per minute :

Red : 72 copper & 144 iron / minute (216 total ore)
Green : 108 copper & 396 iron / minute (copper x1.5, iron x2.75 vs red) ( 504 total ore, x2.3 vs red)
Blue : 540 copper & 864 iron / minute (plus a little petroleum gas and coal) (1404 total ore, x1.95 vs red+green, the (now simplified) oil setup, and steel setup).

They would allow to somehow automate the laying of the infrastructure for that, including the eventual first train tracks, the first defensive wall of the base, and so on, before rushing blue science to get bots.
I feel this would eliminate the 0.17.60 change's main drawback.

[jodokus31]

2) Make con bots not need lubricant somehow
3) Make Petroleum=>Solid Fuel far more efficient (I hate to suggest this, as it pretty much makes light oil worthless, but I dislike not having Heavy/Light oil to turn into fuel more so I think it'd go good with the .17.60 changes)

Hmm, I already feared those kind of proposals. Kind of logical, but.... I'm scared.

Yeah, it's not a solution I like particularly
But it lets the "simplified" oil change stay and it (and 2) gets rid the two main issues that makes it a painful change even if it does make heavy/light oil less useful and oil processing itself unsatisfying

What if engines, lubricant, con-bots etc. are on a different research branch. Just red+green. With an additional oldBOP (or similar) research as root. the newBOP stays where it is now.
The new player can ignore oldBOP and do the stuff, which he wants with red+green and even blue science.
But if he wants bots, he can choose this direction and struggle with multiple outputs.

It just has to be clear, that its more advanced, but not strictly behind blue science.
We could even invent a new science pack: "robot science"

[lacika2000]

2) Make con bots not need lubricant somehow

I'd rather see a low tier ground based con bot limited enough so that regular conbots would be a significant upgrade, but that would help kickstart the scaling needed to automate blue science, so that we don't go from mandatory BOP=> AOP to mandatory BOP=> Bots.
Well I'm aware that it's not exactly mandatory, but ... mandatory-ish ? .

I'd see those to be early green science, and could even become a component to regular conbots (like for example : basic bot + flying frame => flying conbot).

When automating science, blue is still a big step in inftrastructure needs. It's not hard per se, but the scaling is tedious when built by hand, withour the QoL of con bots (copy paste, move, ...).

To illustrate, in order to craft 72 science packs per minute :

Red : 72 copper & 144 iron / minute (216 total ore)
Green : 108 copper & 396 iron / minute (copper x1.5, iron x2.75 vs red) ( 504 total ore, x2.3 vs red)
Blue : 540 copper & 864 iron / minute (plus a little petroleum gas and coal) (1404 total ore, x1.95 vs red+green, the (now simplified) oil setup, and steel setup).

They would allow to somehow automate the laying of the infrastructure for that, including the eventual first train tracks, the first defensive wall of the base, and so on, before rushing blue science to get bots.
I feel this would eliminate the 0.17.60 change's main drawback.

This exactly. At blue science suddenly everything expands, the mines, the smelters, the base perimeter, but there is no automation available to help with the repetitive tasks... as mentioned earlier, the oil “wall” is a symptom, not the cause.

Having low tech con bots available (green science based) would be really a game changer for vanilla games.

[zenos14]

I'd rather see a low tier ground based con bot limited enough so that regular conbots would be a significant upgrade, but that would help kickstart the scaling needed to automate blue science, so that we don't go from mandatory BOP=> AOP to mandatory BOP=> Bots.
Well I'm aware that it's not exactly mandatory, but ... mandatory-ish ? .

I'd see those to be early green science, and could even become a component to regular conbots (like for example : basic bot + flying frame => flying conbot).

I would like this a LOT, I've always felt Factorio needed more bot... not tiers, but subtypes, doesn't need many, but one or two kinds of primitive bot (or Roboport) you can upgrade in functionality as you go up in tech would be nice, and I love the idea of having a basic one you add a flying frame to later on to get the current conbots.

When automating science, blue is still a big step in inftrastructure needs. It's not hard per se, but the scaling is tedious when built by hand, withour the QoL of con bots (copy paste, move, ...).

To illustrate, in order to craft 72 science packs per minute :

Red : 72 copper & 144 iron / minute (216 total ore)
Green : 108 copper & 396 iron / minute (copper x1.5, iron x2.75 vs red) ( 504 total ore, x2.3 vs red)
Blue : 540 copper & 864 iron / minute (plus a little petroleum gas and coal) (1404 total ore, x1.95 vs red+green, the (now simplified) oil setup, and steel setup).

Yeah, even after all the time I've played, blue's always the largest jump in growth between tiers it feels like

They would allow to somehow automate the laying of the infrastructure for that, including the eventual first train tracks, the first defensive wall of the base, and so on, before rushing blue science to get bots.
I feel this would eliminate the 0.17.60 change's main drawback.

Well, I personally would still have a few small things to complain about, but yeah, they'd be mostly "Old man yells clouds" territory than any actually complaint grumble, grumble, less stuff to make into solid fuel, grumble grumble
But seriously, adding a primitive, preblue con bot on would be great and solve a good chunk of the tedium of moving past red/green science, as well as give new players an early bit of designing and automated construction to play around with

....Though I'm now having visions of swarms of ill placed roboports leading to ground bots getting squished by trains, I like it

What if engines, lubricant, con-bots etc. are on a different research branch. Just red+green. With an additional oldBOP (or similar) research as root. the newBOP stays where it is now.
The new player can ignore oldBOP and do the stuff, which he wants with red+green and even blue science.
But if he wants bots, he can choose this direction and struggle with multiple outputs.

It just has to be clear, that its more advanced, but not strictly behind blue science.
We could even invent a new science pack: "robot science"

I'm less sure of this though, would there be repeats of tech or would it just have the old stuff?
If the former it feels kinda redundant tech-wise imo, it could work though if it's a branch off of the newBOP or something close to it, but yeah, redundancy and all that, I feel it might lead to newBOP being skipped on further playthroughs or being seen as a filler tech which I don't like personally, but it could work, would just need quite a bit of balancing
If it's the latter, it still might be mandatory as you'd need lubricant and engines for other things which

I do like the idea of robot science though, feel it would need more options for bots and roboport ussage to justify. But it could also be used to imply more advanced assemblers use a bit of an AI or something? I don't know, neat idea, just not sure what would be on it

[mcdjfp]

As much as I like the idea of early ground based bots, I shudder to think of the pathfinding calculations that would result.

[Koub]

As much as I like the idea of early ground based bots, I shudder to think of the pathfinding calculations that would result.

I have thought about it, but their use would mainly be to get through the end of green science, until blue science. Usually base sizes are reasonable at that moment of the map.

[Adamo]

As much as I like the idea of early ground based bots, I shudder to think of the pathfinding calculations that would result.

I wrote a mod, which is on the mod portal, that does path finding for various tasks with sheep, err, compilatrons. The unit API is actually very fast when used properly. There are some issues where things get stuck, but, I was able to do pathfinding on 10,000 units without using a terrible amount of CPU usage (the algorithm was designed to be very cautious about this, of course) -- with an average around 0.4-0.6 in the time info display, with spikes up to around 1, so not even using as much CPU as, for example, YARM while it's tracking a lot of large fields.

[crambaza]

As much as I like the idea of early ground based bots, I shudder to think of the pathfinding calculations that would result.

I have thought about it, but their use would mainly be to get through the end of green science, until blue science. Usually base sizes are reasonable at that moment of the map.

And tie each bot to a specific burner roboport, so they can't stray from it. There is no ultra long pathing, just short pathing. And of course, they need to ignore belt movement!

[AndrewIRL]

38 pages of comments!

Why not throw my opinion into the ring?

I thought the oil jump was exactly that - a jump. The oil changes, while maybe imperfect (seemed fine to me) are definitely a big move in the right direction. When I read the FFF I was just nodding along, sounds good, yes, yep, ok... it wasn't until I went to comment on FFF306 that I found it was controversial.

Smoothing out the learning curve for beginners is an EXCELLENT idea. Big jumps in the learning curve is a significant factor in why beginners never make it to experts.

Disclaimer: I didn't read the whole thread, apologies to anyone who already wrote exactly this.

[jodokus31]

EDIT:
Just for the record. I played with the idea below and I think, it does not add much.
Actually, blue-science is like robot-science, just more.
The shortcut itself does not help enough and it adds again more clutter, which, I'm convinced, was the goal to remove.

What if engines, lubricant, con-bots etc. are on a different research branch. Just red+green. With an additional oldBOP (or similar) research as root. the newBOP stays where it is now.
The new player can ignore oldBOP and do the stuff, which he wants with red+green and even blue science.
But if he wants bots, he can choose this direction and struggle with multiple outputs.

It just has to be clear, that its more advanced, but not strictly behind blue science.
We could even invent a new science pack: "robot science"

I'm less sure of this though, would there be repeats of tech or would it just have the old stuff?
If the former it feels kinda redundant tech-wise imo, it could work though if it's a branch off of the newBOP or something close to it, but yeah, redundancy and all that, I feel it might lead to newBOP being skipped on further playthroughs or being seen as a filler tech which I don't like personally, but it could work, would just need quite a bit of balancing
If it's the latter, it still might be mandatory as you'd need lubricant and engines for other things which

I do like the idea of robot science though, feel it would need more options for bots and roboport ussage to justify. But it could also be used to imply more advanced assemblers use a bit of an AI or something? I don't know, neat idea, just not sure what would be on it

Robot science could be all things like lubricant, electric-engines, robotics, con bots, logi bots, logistic system, robot speed/capacity, personal roboports. Of course lots of changes, which are probably too much at this state.

The tech-tree could be like this:

Tech tree additional oil shortcut

oil science2.png (234.52 KiB) Viewed 6601 times

AOP is still there to get to this point, but you can have an optional shortcut.

[FuzzyOne]

...
Alternatively, it could be an "integrated oil processing" or something, where crude + water -> petrol. The recipe is thus ~3x slower than the cracking solution, because it's doing all of the processes at once.

This I think is the best approach. First, the changes in the FFFs I believe are designed to avoid the balancing act when you are just learning about oil. However, with just Crude oil in and Petroleum gas out, there is nothing being learnt about multiple pipes on one side of a plant. Requiring both Crude and Water for the BOP at least keeps the multiple pipes in the simplest way (input, not output & balancing) while evening out the learning curve. The other thing that could help is because unlike the various recipes in the chemical plant that have varying outputs, the refinery always has the same outputs, just Heavy oil and Light oil disabled for BOP. Thus the outputs for Heavy and Light should still be indicated, but with a "no" symbol through them (red circle with forward slash combined) and still disallow any mixing even though they are disabled.

This I believe gives the smoothest learning curve, prevents the "straight pipe along both sides" mistake and starts to teach about how to manage multiple fluids for one plant. It also makes sense that you are using water when the output is only Petroleum gas and it should also be noted that the refinery already has a built-in burner stack which is why with BOP it is so inefficient.

[Oktokolo]

I'd rather see a low tier ground based con bot limited enough so that regular conbots would be a significant upgrade, but that would help kickstart the scaling needed to automate blue science, so that we don't go from mandatory BOP=> AOP to mandatory BOP=> Bots.

There is a mod by Klonan implementing ground-based Construction Drones. Their behaviour is a bit (pseudo)random and they tend to go crazy a lot - so they are not production ready just yet.
But if the mod would work just fine, ground-based bots would be my favorite even after faster flying bots become available just because the ground-based ones do fit way better with Factorio's art style.