Friday Facts #409 - Diminishing beacons

[thriem]

Rarely ever fiddled with Beacons.
Overall, i am excited for the addon though not for everything alike. Especially the quality levels seems to become a nightmare - for every min/max gamer. And I see it coming, the recycle loops for non-legendary items, just to give it another try.
And at the same time, any ratio resulting may be thrown overboard (?), which may ruin my setups, where each n wagon of a train feeds a 1/n of a production line ( 4 wagons, 4 production lines, no need for balancers ) - and ideally, the end result would be, that x input trains yield y output trains. Where I can now end up, with a train not leaving because it is missing a part, and a supply train not coming because a production lane is on halt (fixable, but the imbalance still persists).

And not having used beacons much before (mostly because i play with sparse resources), they get now a buff worth trying. Though, production modules in beacons sound quite tempting.

[Illiander42]

We can test it already. Some madlad back-ported quality to 1.1.

Wait, really? Is this a mod I can download? I'd be curious to try it.
Edit: Is it this one? https://mods.factorio.com/mod/janky-quality

That's the one. Don't run it alongside anything that adds items though, or you'll run into factorio's recipie limit.

[GoldenH]

I just don't think beacons are interesting. I also don't think this will change builds with the beacons. You'll either do 8 or 12 just like now and in the same array. What would be interesting if there was say a max % per quality level and that ratio changed over time, say if at no quality 8 was best then at +2 quality 12 was best and then max quality 5 was best.

Or maybe if its effect only stacked on diagonals.

or if it only had 1 range but it could be put above buildings.

This is just more complicated.

[malecord]

Umm, is it only me or nothing actually changed with beacons? Ok, you have diminishing returns, but more beacons is still more gain. So extreme end game is still going to be assemblers surrounded by 16 uglifiers, or am I missing something here?

It looks to me that this change is mostly benefits space platforms because there space will always matter, even in endgame. But on actual factories, where space is infinite, I'm unsure of the effects.

[FasterJump]

So if I add a beacon with speed modules, production speed will increase. And if I add a beacon with efficiency modules, production speed will decrease, because the speed beacon will be less powerful? I'm okay with it actually, even if it makes computations just a bit more complicated.

[Alice3173]

At seven pages, I'm not reading the entire thread to see if anyone else has suggested this already but I think a more interesting alternative would be research that adds more beacon slots with diminishing beacon range instead. If every module inserted into a beacon after the first decreased the range of the beacon, then players would be motivated to come up with different layouts for different situations.

[me_me_me_rawr]

Any buffs to Advanced Oil to offset the 25% nerf to 16 beacons builds?

I would bet between quality modules and quality refineries that won't be a problem.

[Elocutiona]

Going further with this idea, this something could for instance be cooling, maybe even with a heat pipe connection (since heat pipes are so underutilized now). You need to remove heat from the beacon via heat pipes and the benefit of doing so could improve beacon performance. You could either have beacon temperature impact effectiveness continuously or perhaps just in a few performance tiers. Quality scaling can be included by having cooling and quality impact different beacon stats, for instance cooling could improve transmission while quality might improve beacon power draw, are of effect or even just the amount of cooling required (higher quality beacons need less cooling since they use less power).

Having beacons require cooling would also open up end-game scaling for other support production chains too, the heat extracted from the beacons has to go somewhere after all. There could be different cooling systems at different tech levels like simple passive radiators or regular forced-air cooling, large evaporative coolers which need plentiful access to water, refrigeration loops which can bring beacons to sub-ambient temperatures but need to be filled with some special refrigeration agent and so on. Since we're going to visit different planets, the different environments would even influence which cooling solutions are more efficient and in turn each planet could impact how effective or viable big beaconed builds are, after all if cooling is more difficult then perhaps you don't want to have them on certain planets, so you might change your logistics decisions on what to process locally too.

This sounds brilliant and absolutely needs to be a mod if it isn't already, even if it's just the simplest possible case of the beacon shutting down above a certain temperature.
For me, this would make beacons an actually interesting endeavour, as the current drawback of power draw actually becomes a limiting factor, not just an annoyance you solve with more solar panels. If 12 beacons around a machine needs so much surrounding cooling that the beacon:machine ratio approaches 12:1, that would be a legitimate example of what I would call diminishing returns. You can boost a machine to absurd levels using quality modules, but doing so would be such a comically inefficient use of space and resources that it'd have questionable viability even in the far late-game. Unlike gigantic solar panel arrays, you can't just put cooling arrays far away where their space requirement isn't a problem. If a max quality 12 beacon building needs its own city block because the cooling is the size of an 8 reactor setup, that would affect logistics as well, with trains and bots spending dramatically longer simply getting from place to place.

Also, the idea of having to create coolant loops sounds rad, anything involving belt loops or things feeding into themselves is right up my alley.

[bnrom]

Only tangentially related: are the any plans to have modules interact with weapons?

Given that now weapons on space-platforms are almost a kind of core resource building (in that one needs the weapons to break the asteroids and that the number of asteroids the platform can safely encounter is limited by its rate of damage - and set by orbit selection and speed) it feels that letting modules affect them would make sense.

Speed modules could give fire rate.
Productivity modules could give damage bonus.

Maybe one could even create new modules, e.g., range.

[XT-248]

I am on the fence with this small passion project.

On the one hand, it does change the calculus on how many beacons and what quality of modules to use per assembler to obtain the desired higher throughput per machine.

On the other hand, I do not believe it will change much, as I personally don't use beacons until very late in the game with a very large megabase that has sustainable high throughput, and even then, I only use them sparingly.


I think this will encourage changing from almost always 12 beacons to one machine to 8 or 4 or whatever the new meta will be. The same is also correspondingly accurate for 16 beacons for larger machines like Foundry.


The short implementation period has a few small silver linings: This mechanic is easy to tweak/retune or overhaul in the future if needed. The changes don't impact the large majority of my factory, and it will continue to operate the way it used to before.

[CXZman]

Hello!

I think the overall changes to beacons do sound like an improvement over what we've got in 1.1, but I have to admit I'm also a bit disappointed that all this amounts to is effectively a tweak on numbers rather than some more fundamental change to make beacons more interactive or in-depth somehow. Beacons go from passive power consumers with an AoE bonus to passive power consumers with an AoE bonus but different numbers on the bonus. I appreciate that the change in numbers will very likely result in a change in how we're going to build, but I would've liked the beacons themselves to become more interesting.

I know that sounds pointlessly vague, so just as a basic example, how about beacons require something in order to function beyond just power? Routing something to the beacons themselves would shake up the layouts and usage patterns for sure. Going further with this idea, this something could for instance be cooling, maybe even with a heat pipe connection (since heat pipes are so underutilized now). You need to remove heat from the beacon via heat pipes and the benefit of doing so could improve beacon performance. You could either have beacon temperature impact effectiveness continuously or perhaps just in a few performance tiers. Quality scaling can be included by having cooling and quality impact different beacon stats, for instance cooling could improve transmission while quality might improve beacon power draw, are of effect or even just the amount of cooling required (higher quality beacons need less cooling since they use less power).

Having beacons require cooling would also open up end-game scaling for other support production chains too, the heat extracted from the beacons has to go somewhere after all. There could be different cooling systems at different tech levels like simple passive radiators or regular forced-air cooling, large evaporative coolers which need plentiful access to water, refrigeration loops which can bring beacons to sub-ambient temperatures but need to be filled with some special refrigeration agent and so on. Since we're going to visit different planets, the different environments would even influence which cooling solutions are more efficient and in turn each planet could impact how effective or viable big beaconed builds are, after all if cooling is more difficult then perhaps you don't want to have them on certain planets, so you might change your logistics decisions on what to process locally too.

I just think the end-game scaling could really benefit from a bit more depth than just figuring out how to place the expensive AoE bonus buildings. I don't mean to imply that it's all bad or anything, train builds and direct insertion builds in general can get pretty tricky to design for beacons already (it's not all 8 or 12 beacon builds with belts or logi chests), but I think it would be very interesting if building big had a bit more of a design space to play in.

Great idea!

[Svip]

How about the quality adding more module slots to the beacons, and the diminishing returns is based on number of module slots affecting the assembler? That would allow the player to use fewer beacons at peak performance if the beacons were legendary quality.

[mmmPI]

Going further with this idea, this something could for instance be cooling, maybe even with a heat pipe connection (since heat pipes are so underutilized now).

Also, the idea of having to create coolant loops sounds rad, anything involving belt loops or things feeding into themselves is right up my alley.

In exotic industries, beacons require coolants to function that need be brought from "insulated pipes", special fluid pipes, not heat pipes, but somewhat related ? ^^ there are tiers of them, and they have some overload mechanism.

It's interesting and refreshing to use, but my fear with such concept is that it's altering a lot the game ( positive point for overhaul mod a bit less for expansion)

It does somehow adress "how many tiles are occupied by beacons vs other thing in late game". But it somehow translate into "how many tiles are occupied by infrastructure to supply beacons". The ratio of which is important i found, if it's too much, then you "create a factory to supply beacons", in which you can't use beacons, if it's too little, then it feels trivial to just pipe the coolant everywhere.

It's a deeper mechanism that the one proposed, i think interesting and would want to try it, but i do not "regret" such deeper mechanism not being annouced in the base expansion, as i think there could be mods to expand on such things, this is how i interpreted the last part, where it's shown that the value are a table "easy to mod", not some hardcoded formulas that can't be accessed.

[PaszaVonPomiot]

I don't like beacons because they feel like cheating. I like to put one or two here and there and that's it.

[FunMaker]

For me the change does nothing for the better. Min-Maxed Builds will not be changed at all. We all know, that if there is a perfect or good layout, many will use it. No change in gameplay will change that there will be one build to rule them all. If they changed the mechanics in a way, that there are many factors that are possibly optimized by a layout, then there will be one perfect layout for every optimized factor.
The targeted change addtionally is nearly a no brainer. There might be some edge cases where the sweet spot will change where you can add an additional assembler or add another beacon and one over the other would be better, but overall the dull mechanic of beacons will not change (and quality does not change that because every min maxer will build a giant production line of top quality modules and beacons and will not care over decisions along the way).

I like the idea, that beacons must be cooled by a heat connection - this way the beacon problem is placed in the normal space-puzzling game. The downside of this will be UPS loss with beacons. One may think, that is a bad thing - i think this will be okay, because putting 12 or 8 beacons because it is UPS friends is a dull decission of it's own. This way the UPS race is open for other builds.

[emty]

I don't like beacons in either form.
Almost always there is a single best solution - put max productivity in the machine, pump as much speed as possible through the beacons.
There are three associated challenges:
- design challenge, since they consume space around machines
- they require a lot of resources
- they require a lot of power, but unless power distribution is made more complicated somehow, it just translates into more resources, because solar energy exists.

What do they provide in return?
They only make factory more UPS-efficient, since faster assembling machines spend less ticks per recipe execution.


But moving two steps back, I don't understand why assembling machines have to consume CPU during recipe execution at all?
I assume, that IF power doesn't fluctuate, so production time is constant THEN every recipe should "consume" CPU only on the first and the last production tick, when you need to physically add/subtract the ingredients from the internal storage.

So here is my idea. Why not implement the scaling for the endgame differently, by introducing machine-only bulk recipes?
For example, by default there is a recipe:
1 red chip = 6 seconds + 4 wires + 2 green chip + 2 plastic

It is possible to introduce bulk recipe:
10 red chips = 60 seconds + 40 wires + 20 green chips + 20 plastic
which should be much more UPS efficient, at least on the machine level.

But, it is also possible to not do a simple linearly scaling recipes, but do scaling with diminishing returns or different ratios, for example:

Code: Select all

// UPS is better, but x1.5 worse performance per machine
10 red chips = 90 seconds + 40 wires + 20 green chips + 20 plastic
// even better UPS, but x2.5 worse performance per machine
20 red chips = 300 seconds + 80 wires + 40 green chips + 40 plastic
// 10% productivity bonus
11 red chips = 60 seconds + 40 wires + 20 green chips + 20 plastic
// 10% better wire consumption, 20% worse plastic consumption
10 red chips = 60 seconds + 36 wires + 20 green chips + 24 plastic
// Improves speed, but requires extra ingredients for better performance
10 red chips = 40 seconds + 40 wires + 20 green chips + 20 plastic + 10 sulfuric acid, x1.5 energy consumption
// UPS is still better, but needs more power
10 red chips = 60 seconds + 40 wires + 20 green chips + 20 plastic, x2 energy consumption

These alternative recipes can be made available and upgraded randomly over time through infinite research. Then there is an option to update your designs over time to take advantage of these potential efficiency gains.
Yes, this will be very chaotic, lategame experience will be different for everybody due to randomness and it will be harder to calculate best ratios. But I think chaotic = more interesting.


And beacons... Well, they can stay, but I would at least nerf them, so there is a hard limit of them that can affect one machine. And I think it should be closer to 6 for it to look sane, not 12. Maybe even one beacon boost one selected type of produced item not all, so you have no incentive to do direct insertion everywhere.

[Bloodred217]

This sounds brilliant and absolutely needs to be a mod if it isn't already, even if it's just the simplest possible case of the beacon shutting down above a certain temperature.
For me, this would make beacons an actually interesting endeavour, as the current drawback of power draw actually becomes a limiting factor, not just an annoyance you solve with more solar panels. If 12 beacons around a machine needs so much surrounding cooling that the beacon:machine ratio approaches 12:1, that would be a legitimate example of what I would call diminishing returns. You can boost a machine to absurd levels using quality modules, but doing so would be such a comically inefficient use of space and resources that it'd have questionable viability even in the far late-game. Unlike gigantic solar panel arrays, you can't just put cooling arrays far away where their space requirement isn't a problem. If a max quality 12 beacon building needs its own city block because the cooling is the size of an 8 reactor setup, that would affect logistics as well, with trains and bots spending dramatically longer simply getting from place to place.

Also, the idea of having to create coolant loops sounds rad, anything involving belt loops or things feeding into themselves is right up my alley.

Total shutdown would probably feel bad since it would make current-style beacon use impossible, but I think even just something simple like a few tiers of effectiveness would do great. Personally I really like "difficulty" in games that rewards you for engaging with it, so for beacons I would like to see them made more useful (more powerful) as a reward for being willing to deal with increased complexity. Like if they're uncooled they reach high temperatures and only run at 0.5 transmission effect (like they do in 1.1, so nothing is taken away), with moderate, fairly simple cooling setup and sticking below a certain temperature you get 1.0 transmission effect and with some later-game, more complicated sub-ambient cooling you get 2.0 effect or something. Obviously the numbers would need to be play-tested / tuned to work right. Like you mentioned, the cooling itself would need to be sized appropriately and I suppose it would need a lot of attention to make sure the cooling-related machines required don't end up being more UPS-costly than the UPS you save by using the cooled beacons too.

Great idea!

I'm glad to hear you like it!

It's interesting and refreshing to use, but my fear with such concept is that it's altering a lot the game ( positive point for overhaul mod a bit less for expansion)

Yeah, that's why I was thinking that standard / "uncooled" beacons could be allowed to operate at the worst performance tier which would be the same as current vanilla, while different tiers of cooling would let them operate at superior levels. You would still have the option for power-only beacons but you'd also be able to boost them if you want, rather than cooling being a requirement even for baseline operation.

It does somehow adress "how many tiles are occupied by beacons vs other thing in late game". But it somehow translate into "how many tiles are occupied by infrastructure to supply beacons". The ratio of which is important i found, if it's too much, then you "create a factory to supply beacons", in which you can't use beacons, if it's too little, then it feels trivial to just pipe the coolant everywhere.

That's actually part of the reason why I was thinking of heatpipes. Not only are they an underutilized mechanic, but they're not effective over long range, so you cannot just make a single big cooler somewhere and connect your entire factory with heatpipes to it, since it would be too far away to work effectively so it adds another wrinkle to the planning process compared to using fluid coolant directly. If heatpipes are used you'd have to have to have at minimum some sort of heat exchanger going from heatpipe to some coolant in close proximity to each beacon group, or just have dedicated coolers on the build itself. Heatpipes would also allow for really basic / earlier game setups like just connecting beacons to on-site radiators and could go all the way to more complex and involved loops, rather than having fluid coolant loops as the only option.

The downside of this will be UPS loss with beacons.

The cooling machinery would certainly use some UPS, yeah, but ultimately it's just a tweaking numbers game to come out ahead. Let me put it this way, if a 100-machine build with current vanilla beacons can be turned into a 10-machine build with cooled beacons, but you also need another 10 machines for cooling you're still going from 100 active machines (+inserters for each and so on) to 20 active machines total. Those cooling machines will use up some performance but going from 100 total to 20 total should still be faster for UPS in the end. The numbers could just be adjusted such that the extra headache of cooling still comes out ahead in terms of UPS, so it's worthwhile at end-game. Different environments on different planets could also make cooling more or less effective, so the UPS-optimal solution might depend on where you're building too.

[Elocutiona]

Going further with this idea, this something could for instance be cooling, maybe even with a heat pipe connection (since heat pipes are so underutilized now).

Also, the idea of having to create coolant loops sounds rad, anything involving belt loops or things feeding into themselves is right up my alley.

In exotic industries, beacons require coolants to function that need be brought from "insulated pipes", special fluid pipes, not heat pipes, but somewhat related ? ^^ there are tiers of them, and they have some overload mechanism.

It's interesting and refreshing to use, but my fear with such concept is that it's altering a lot the game ( positive point for overhaul mod a bit less for expansion)

That sounds neat, I haven't dabbled too much in overhaul mods though. Perhaps I'll read up on the EI details and theory-craft what a stand-alone beacon cooling mod could look like. If I'm feeling particularly ambitious, I may even attempt some rudimentary modding myself.
I like the idea of beacons needing to be connected to heat exchangers with heat pipes, and those heat exchangers accepting coolant and outputting warmer coolant, which would have to be cooled back down through a new radiator building.
I also like the idea of having the speed of heat transfer from heat pipe to coolant, and coolant to air be governed by actual heat transfer equations, such that it isn't just a simple question of having enough cooling, but about how hot you're okay with your beacons running at equilibrium.

[OldFart]

IMHO, All these diminishing returns are overcomplicated and don't improve anything. 8-12 beacons build was good and remains good. It slightly changes the final output number, but nothing else gameplay-wise.
A good way to fix that and give more meaning to quality is to hard-cap the amount of received effects.
Let's say, the hard cap is 8 standard beacons (400% in 1.1). In the beginning, all beaconed builds will use standard-quality beacons in standard beacon strip configuration. Later, with higher-quality beacons available, players will need fewer higher-quality beacons per build, and direct insertion becomes more viable.
Also, higher-quality beacons definitely should have more range. That, when combined with a hard cap, makes even more build optimal for the late-game

[Mandromeda]

The main purpose of beacons is to allow massively increasing your production in the late game while being more than just a module or a faster machine. To make use of beacons you need to adjust your building layout for them.
[...]
Since there aren't many other significant alternatives to these two layouts, all in all using beacons seems restrictive on build variety and tends to feel monotonous and uninteresting, especially compared to how varied the earlier stages of the game are.

I agree with the desired outcomes as well as the changes to improve how they'll be achieved.

I have a dumb idea I don't think I've seen talked about yet. I think it would encourage players to:

• Want to use beacons;
• Want to modify and/or improve their factory layouts throughout a playthrough; and
• Offer an optimization challenge for the true min-maxers, offering potential complexity that only smart planning and calculations can fully optimize, with through-playthrough increasing layout variance and complexity!

1. Transmission frequency bandwidths: Have beacons start off with being able to transmit over one 'frequency' coverage layer; affecting...
2. Beacon Interference: Overlapping coverages may, in lieu of 'Overload', (i) reduce benefits; (ii) provide only N layer(s') worth of of benefits; (iii) 'overload' by canceling benefits, or worse; (iv) 'interfere' making things worse than if there were no beacons at all. Hmm, I wonder how lightning strikes could impact things...
3. Researchable new frequency bands, or 'channels' would allow overlaps with whatever consequence (from harmless, helpful, benefit-killing to downright hurtful), and where these consequences can be either diminishing, additive or whatever.
4. Just like the real life radio spectrum, base beacons would all default to transmitting over Medium Frequency (MF) with a fairly large coverage area. Other MF towers nearby would have, perhaps, zero overlap or interference effect.
5. We could then research higher frequencies (HF, VHF, UHF, SHF, EHF) with (i) increasing 'information' benefits, at (ii) decreasing coverage radius, (iii) increasing interference penalties, and (iv) higher power costs) or; Lower bands (i.e. LF, VLF) with (i) decreased 'information' benefits, (ii) increasing radius (LF, VLF would have a HUGE radius), and (iii) better overlap effects, or maybe just 'overlap detriment avoidance', at a (iv) MUCH lower ongoing power cost.

I apologize if this has all been asked or proposed before. Thanks for making such a wonderful game!