So... Let's talk about bots, and how to fix them properly...

[zytukin]

If I may think out of the box a little...

If you don't like bots, why not avoid using them?
There is no right or wrong way to play. Thus no need to change them.

I have over 2000 hours in the game, 700 in my current factory. I have never used logistics bots.

[Deadly-Bagel]

If you boil the problem down, you get a different conclusion. Belts and trains are more fun than bots, why? There are more puzzles to solve, and more tools to solve them with. Logistics bots are boring because we only have two tools to influence them; Roboports and Chests. Not much to do about Chests, especially after adding the Buffer Chest, however I think something could be done about Roboports.

Currently a Roboport is an all-in-one tool for storing bots, charging bots, extending Construction and Logistics range, and holding spare Repair Packs. The last role is fulfilled by storage and/or buffer chests I think, but the other roles are unique to the Roboport. This means that by adding range, you're also adding storage and charge stations. This makes it pretty difficult to do anything clever.

What if Roboports were dismantled into their components? Bob's Mods already has this so the majority of the functionality is already there, could just do with a few tweaks such as the ability to specify what robots are allowed to enter a robot storage. To be clear, this would involve removing the Roboport. A greater barrier to entry perhaps, but then Factorio is hardly a simple game.

So what kind of puzzles have we introduced? Well, you can now plan your logistic network in more detail. What's the point of having Construction Robots scattered all around your base, or all collected in your solar farm? Pff, dedicated chests near a central storage makes much more sense, maybe a few extra around your perimeter for repairs. Likewise, Logistics Robots taking Repair Packs to your perimeter, why do they bother hanging around? There's never anything to pick up from there, so they could be denied storage in those bot chests. You would also need to more carefully consider your charging points so bots don't need to travel so far out of their way.

Yeah, this has been thrown around in suggestions, but the problem is this is actually a buff which is the opposite of what everyone seems to want. You would therefore need to introduce a nerf as well. I wouldn't recommend hitting stack size or speed, as these can be overcome with "moar robotz" so the simple solution is charge. Halve the charge that Logistics Robots can hold and you're in for some fun problems. You now need to consider not only the range of the network, but the range of the robots themselves. Needing to charge more frequently means more detours, but the actual charging process would be faster.

To be clear I don't really have a problem with robots myself, I only ever use them to carry products to a central storage (as the alternative would be either having them spread all over the factory, or a bi-directional bus) and maybe the occasional lazy hack like carrying Iron Ore around for Concrete. And of course to stock me with belts, inserters, etc and take away old belts and wood.

[bobucles]

There are three major factors that determine how well a bot compares against belts:

Code: Select all

Cargo x (recharge per map tile) / (energy used per tile of item movement) => bot power

Cargo capacity is the biggest multiplier of a bot network and changing it will have a non trivial effect on the power of bots. Decreasing the charge rate means more roboport spam, further reducing the efficiency of the network. The energy used per tile of item movement is not easy to solve. Bots become more energy efficient as they move faster and the shape of the network determines how much total distance a bot really needs in order to haul an item. If bots don't detour to recharge and move items entirely in one direction then this efficiency is 50% at the most.

Reducing battery size will decrease route efficiency as bots have to detour more frequently. It's a minor change that hurts low tech bots more than anything. Increasing the charge rate WILL make the bot network better. Reducing idle drain will make low tech bots better but not really affect high tech bots. Increasing energy drain per tile will hurt all bots. Increasing cargo WILL dramatically increase the network.

[MeduSalem]

Now, there are 3 types of transportation in Factorio: belts, rails, and bots. The main issue that many people see is that for several reasons, belts, rails, and bots are way out of balance. They were designed in a way where they should work together and complement each other, but instead bot's basically take over once you get them. Unfortunately, due to how well bots scale, no amount of nerfing can really fix that... But, while this is definitely a problem, it isn't the problem... You see in other circumstances, this wouldn't be a huge problem, if something is a better fit for most situations, use it, no big deal. The balance may be a bit off, but it won't really get in the way of having fun. Unfortunately, the issue with bots is that there are far fewer solutions available to solve problems with once you get them. (Need more iron smelting? Place down a blueprint with a bunch of robotic smelters! Too many bots to charge? Add more roboports! Bot network to slow? Add more bots or research upgrades that affect all of them! Tired of adding bots by hand? Automate it so you never have to think about it again! And that's about it...) The bots have an absurdly small number of problems and solutions to those problems, and, since most of the fun of Factorio comes from solving problems, there isn't nearly as much fun to be had once you get them... And that becomes a huge problem, because they aren't the first method of transport you get, they are the last, and they overshadow everything else for most situations.

While I could also present some "ideas" (which I also presented over the years already)... I won't present any solutions to the Bot vs Belt vs Train problem.

Instead I will point something else out... which also the Devs should consider reading before anyone wants to reach out for a Nerf hammer or introducing complex additions/changes to existing features:

• Laser Turret vs Gun Turret vs Flame Thrower Turret problem
  
  Basically the same thing. Once Laser Turrets are a thing most people don't give a shit about Gun Turrets or Flame Thrower Turrets because they mean more effort. Laser Turrets don't require anything except electricity, which is bascically for free and can be conveniently provided to any location on the map.
  
  Solutions have been presented, but people either didn't want to acknowledge that there still is a balancing issue between them, or they like the lazy shortcut presented by Laser Turrets, or they can't agree on from community members presented solutions.
• Solar Power vs Steam Power vs Nuclear Power problem
  
  Basically the same thing. Once Solar Power is a thing most people don't give a shit about Steam Power or Nuclear Power because they mean more effort. Most people prefer Solar Power because it requires no upkeep and is more UPS friendly. Steam Power is even in a worse spot because even if people don't go for Solar Power they go for Nuclear Power instead and leave Steam Power to rot in hell.
  
  Again... solutions have been presented, but people either didn't want to acknowledge that there still is a balancing issue between them, or they like the lazy shortcut presented by Solar Power, or they can't agree on from community members presented solutions.

So now to the point why that is important:

There are several concepts in the game that become obsolete during later stages of the game for various reasons. If the devs didn't want that then they shouldn't have introduced a concept that is already overpowered on the drawing board in the first place and which has a huge potential to supercede what is already existing.

I've been playing computer/video games for almost 30 years ever since NES came out and if there is one thing I have noticed then that trying to be cutting edge with "asymmetric gameplay" is usually the source of all imbalances because of how one of the presented alternatives usually dominates the rest and it becomes clearer and clearer the further one is into the game.

I've helped shaping several games the past 15 years with giving ideas and advice while they were still in development... if there is one advice that fits the problems above... then never bring a feature that has almost no disadvantages as an alternative to features that have several disadvantages, as well as never bring a feature that has many advantages as an alternative to features that have only a few advantages. It just renders balancing an absolute nightmare.

People often say that they dislike rock-paper-scissors style games, but if you don't want to risk major imbalances like the above then that's the time-proven way. Tie all alternatives to suffer from a similar, if not the same, source of problems.... and make numbers comparable so they can be efficiently tweaked.



To conclude about the problem I want to say that from my perspective it is too late for changing anything radically about all those concepts because now that the people have been playing Factorio for several years already (It's what I call the Alpha-access-curse) they are accustomed to the way things are and if someone dares only so much as touching one of these delicate topics they are in for a major shitstorm coming from at least half the community, as could be seen in the 2 FFF threads.

So my advice is... just leave it be... and make it better in a Factorio sequel in a couple of years.

[bobucles]

Most people prefer Solar Power [over nuclear power] because it requires no upkeep and is more UPS friendly.

This is serious?

UPS isn't an average player's problem. Circuit network isn't an average player's problem. Rocket #300 and beyond isn't an average player's problem. All the "problems" that players keep citing are examples of putting hundreds of hours into Factorio and reaching the utter breaking points of the game engine. They are "problems" in the exact same way that stacking a thousand cabbages in Skyrim or playing 5 touhou levels at once is a problem. Players have gone so far beyond the scope of the "Regular game" that the mega base is the normal way to play!

There's nothing wrong with pushing Factorio even further beyond the ordinary limits. But keep in mind that the real meat of getting into the game is the rocket. Everything beyond that is gravy on the cake.

[MeduSalem]

Most people prefer Solar Power [over nuclear power] because it requires no upkeep and is more UPS friendly.

This is serious?

UPS isn't an average player's problem. Circuit network isn't an average player's problem. Rocket #300 and beyond isn't an average player's problem. All the "problems" that players keep citing are examples of putting hundreds of hours into Factorio and reaching the utter breaking points of the game engine. They are "problems" in the exact same way that stacking a thousand cabbages in Skyrim or playing 5 touhou levels at once is a problem. Players have gone so far beyond the scope of the "Regular game" that the mega base is the normal way to play!

There's nothing wrong with pushing Factorio even further beyond the ordinary limits. But keep in mind that the real meat of getting into the game is the rocket. Everything beyond that is gravy on the cake.

Yes this is serious... because even the devs spoke about the UPS problems when it comes to Bots vs Belts.

The devs also see it as one of the valid arguments for why they wanted to nerf the bots. To me that makes it obvious that they reached a limit in optimizing Belt mechanics and hence can't get it to work nearly as UPS friendly as the Bots.

Not that UPC concerns would be the only balancing issue about Bots or Solar Power or pretty much anything in the game... and I didn't say that UPS is the most important thing to me, but to some people it is.



I also think that bots are way too overpowered in general. I mean they fly unrestricted above every obstacle... no amount of energy they require, no amount of nerfs to the cargo size, etc is ever going to counteract the convenience of not having to deal with belt spaghetti. Even if they'd only carry 1 item each and require 5 times the energy most people wouldn't go for belts.

Any solution to the bot problem would have to involve some kind of concept where bots are restricted in their movement, maybe even being forced to the ground altogether. That would bring the puzzle element to the logistics network.



But that said balancing issues post-rocket-launch are worth their own discussion.

While I agree that currently there is not much point in making it balanced in lategame after launching the rocket, the devs are right when they say that as long as everything is so unbalanced in mega bases and with being forced to certain play styles to go even that far etc that they don't have much room for expansion there if everything but specific playstyles become obsolete.

[Aeternus]

Solar does have drawbacks:
+ No upkeep/no fuel requirements
+ Can be put anywhere
+ No pollution
- Relatively high initial investment.
- Requires a wide area
- Requires batteries to catch the nighttime woes
Solar parks require extremely large areas set aside, especially for megafactories. Feasable only if that area is available. Low power generation per tile, but if there's room, there's no reason not to.
The "Free energy anywhere" also makes it useful for stand-alone mining outposts far behind biter lines.

Nuclear:
+ Extremely high power generation capacity in a small area
+ Low pollution (in fuel production)
- Very high initial investment
- Slow to startup
- Requires some combinator magic to not be wasteful
- Most efficient once you have a very high draw due to adjacency bonuses.
- Water can become an issue
Nuclear plants are typically built in rows with tanks for steam storage and combinator magic to make sure the reactors aren't running continuously. I see them as an endgame power solution to replace chemical steam powered plants. Even a small nuclear plant can easily reach up to a gigawatt or two. Problem is that the research and building of these things cost a very high amount of resources. Once it's going however, if you dedicate all your uranium to fuel, even minor uranium fields can power your factory without any issue.
Supplying remote bases with power is a possibility with nuclear power - send steam in tankers to remote turbines.

Steam/Chemical boilers:
+ Low investment cost
+ Low tech, can be built quickly and early. Most basic designs with burner inserters and yellow belts remain viable at endgame.
- High pollution
- Requires fueling.
- Requires water, pattern tends to be wide. Big area needed.
Chemical fuel can remain an option but it requires a lot of room. At endgame, a single compressed blue belt of rocket fuel can supply 9GJ/s, equating to 4.5GW of electrical power generation - which ultimately comes from oil, an inexhaustable fuel source. It requires a wide area on a coastline however because of the water requirements. A single offshore pump can supply a row of 20 boilers, each boiler supporting 2 steam engines exactly.
If there is a local fuel source a small power plant can be built to power things on-site. Generally, not recommended.

As for turrets:
Gun turrets:
+ Obtained early in the game
+ Lots of stacking research bonuses
+ Ammo options improve as technology advances
+ High rate of fire
- Behemoth biters very resistant to these
- Low hitpoints. WAY too low hitpoints. Why are these more fragile then the lasers? Where is the "Reinforced gun turret" with double barrels and double the HP?
- Ammo consumption. And too high for my tastes.
- Low range
Basic defenses and... gun wise an advanced version of these is kinda missing from this game. I've tinkered with these, since they can cut down on power requirements for remote bases at the cost of additional logistics, but the biters seem to be able to take these puppies out too easily. They are just too weak.

Laser turrets:
+ High health. Tanky defense
+ Good range. Spitters won't outrange these.
+ No logistics needed. Just power.
- Requires some infrastructure plus oil chain to start production.
- Expensive turret
- High power draw while firing, can be problematic for unbuffered grids. Can be solved with a small accumulator cluster. Without batteries your fire rate will suffer.
- Completely useless during a blackout or power grid interrupt.
My personal go-to for all things defensy. I prefer having power headaches (can always use a power latch to cut power to production if defenses require it) then ammo woes (which takes valuable resources better spent on production of more laser turrets). If the gun turrets were as sturdy and range-y as these, they'd be better balanced and more useful for low powered remote mining bases.

Flame turrets:
+ AOE attack
- Limited field of fire
- Inaccurate
- Friendly fire hazard
- Requires liquid ammo
I don't consider these to be in competition with the other 2 turrets. It's a completely different beast - shooting an inaccurate AOE attack that will burn whatever wanders through. They are good when used at spots where you are frequently getting hit by groups but their inaccuracy makes them useless as standalone turrets. Supplying them is also annoying - with barrels now reduced in capacity. They also can burn other structures so watch where you place them. Walls are immune to fire though.


On the bots vs belts vs trains though... I remain of the opinion that the fact that an infinite amount of bots can instantaneously transfer any amount of cargo between their chests and themselves is the issue. That kind of nonblocking throughput versus the limitations of even the stack inserter... makes it no contest. That bots need energy to move stuff... like you said, electricity is practically free.
Belts are great for short to medium distances where you need to bulk move lots of the same cargo. Trains are ideal for high volume over great distances. It'd be great if the bots would be a flexible short range solution - low volume over medium distances. That bots beat stack inserters and the fastest belts for short range cargo transfers (like train loading/unloading stations) is something that... bothers me. Bots have not defined their role in the game. They need the drawback that tells players "hey, don't use these for this purpose."

And yea, people like building megabases. It may not have been the initial intent of this game, but Factorio is DEFINATELY a game that invites you to go big, bigger, biggest! I think once this game goes stable and launches to the big crowd, you'll see a lot more folk wanting to do that. Make it a logistical SimCity - and the maps of Factorio are a LOOOOT bigger then SimCity's ever were.

[MeduSalem]

[...]

Honestly... I have discussed this to death already in the "Solar Power is less of a no brainer"-thread, so I won't go too much into detail about it anymore.

People have been using the "High initial investments" and "space requirements" fallacy as an excuse to justify the lack of real downsides for years now.

One can see from the fact that people still prefer to use Solar Power and Laser Turrets excessively even though it is suuuuuch a hassle and downside that the investments are sooooo high and it requires sooooo much space and require sooooo much energy that they are effectively self-deceptive about it.

In reality the Initial Investments and Space Requirements and Energy Requirements are NOT a drawback at all within the large scope of a factory, especially the longer you play. At least not one that can be compared to alternatives that require on-going investments in form of resources that eventually out-match the initial setup costs, which is especially true for the turrets, but also for energy production.


No, people use Solar Power and Laser Turrets out of convenience... it is the lazy man's shortcut and that brings us back to Logistic Bots.

[Aeternus]

Well, we'll have to agree to disagree on power I guess. Me, I prefer nuclear power when it becomes available, and use solid fuel powered steam engines until then. Mainly because if you have the structures, you can plop down 600MW of generating capacity in seconds. Not really gonna happen with Solar I fear
And the space issue does become problematic at disconnected outposts. The bigger they are, the more you need to defend.

Turretwise I'm fully agreeing with you. Laser beats gun in almost every aspect, so once you get lasers, why bother with gun turrets. Could take a page from Bobs mods - the sniper versions of the gun turrets would be an awesome addition. Greater range, very low fire rate (so low ammo consumption). But mods are that, mods... not the core game.

[MeduSalem]

Well, we'll have to agree to disagree on power I guess. Me, I prefer nuclear power when it becomes available, and use solid fuel powered steam engines until then. Mainly because if you have the structures, you can plop down 600MW of generating capacity in seconds. Not really gonna happen with Solar I fear
And the space issue does become problematic at disconnected outposts. The bigger they are, the more you need to defend.

Turretwise I'm fully agreeing with you. Laser beats gun in almost every aspect, so once you get lasers, why bother with gun turrets. Could take a page from Bobs mods - the sniper versions of the gun turrets would be an awesome addition. Greater range, very low fire rate (so low ammo consumption). But mods are that, mods... not the core game.

I use combinations of all three turrets in my recent plays.

Defense

Defense.PNG (516.02 KiB) Viewed 7712 times

It is chunk-aligned and repeatable. So basically 2 Laser turrets per Gun turret, and 2 Gun turrets per Flame thrower... and a filler Laser Turret between.
I developed that pattern in that fashion because I wanted it to be chunk-aligned and also because the Laser turrets are much weaker than Gun Turrets and Gun Turrets weaker then Flame Throwers, but also because including the Wall in front of the turrets the Flame thrower's cone overlap slightly this way almost all the way up to the wall, and also so that the Gun turret coverage overlaps so that even if a Biter comes exactly between 2 gun turrets both of them can start firing at the biter already at the range of ~17-18 tiles.

But that said, I think the Laser Turrets shouldn't just require plain electricity. I made several times the suggestion that they should changed to require also ammunition. Like in form of new batteries once the old ones are worn out due to constant recharging... having to deal with used batteries much like used Nuclear Fuel Cells... Or having to deal with Plasma Cells or whatever.


Power wise I am also going for Nuclear Power to cover the most part of my energy demands... just because I find Solar Power boring. I only stamp down Solar/Accu stuff if I happen to have a free spot in my otherwise modular grid layout of the base I am not using otherwise yet, but it isn't excessive.

I also use chemical Fuel every now and then if I have to burn off something. Like for example wood or wooden poles.
Or to get my refineries unstuck as I haven't found a safe way to prevent the strange case of running out of Heavy Oil/Lubricant while being full of Light Oil / Petroleum, or the case of running out of both Heavy/Light when full on Petroleum in some ridiculously weird edge situations. That there is no way to convert them back is crap sometimes. Yeah I only crack as necessary, but still it happened to me that I ran out of Heavy Oil a while ago and I don't even know why it happened as I couldn't reproduce the cause. So I included the possibility to burn off whatever I happen to have too much of if it's the only way to get the refineries unstuck.
Also as a failsafe backup in case I run out of Uranium.



But back on topic...

On the bots vs belts vs trains though... I remain of the opinion that the fact that an infinite amount of bots can instantaneously transfer any amount of cargo between their chests and themselves is the issue. That kind of nonblocking throughput versus the limitations of even the stack inserter... makes it no contest. That bots need energy to move stuff... like you said, electricity is practically free.

I had another thought about that... For me it is basically beaconized setups that are the source of the problems, especially for belts.

1. One of the problem is the inserters themselves. Since a stack inserter can move 13 items in one swing from a machine to a chest due to inventory-to-inventory-bonus... it automatically beats the belts, as it just takes too long for an inserter to put stuff on a belt or to pick up from a belt. It becomes more clearly though in highly speed-beaconized setups only, which often also don't allow for more than one or max 2 inserters per side due to space restrictions.
2. Also there is the problem that Inserters can only put stuff on the far side of the belt, which makes belts awkward to use in confined space (such as between beacons) as one side of the belt is unused most of the time, halving their effectiveness. Using splitters for belt-rebalancing in already such confined spaces is often not possible or very, very, very ugly, like all belt-rebalancer bullshit contraptions which I never use exactly for that reason that they are ugly.
3. Then there is also the problem of recipes which require multiple resources... can't fit enough belts in the space between the machine and beacon. Sushi belt is throughput limited (and prone to errors) and Belt braiding is like a hack already anyway.
4. Some recipes are problematic in general due to how much resources they eat and their crafting times. Even if the contraption fits within the beacons... you often can't feed more than a few assemblers before express belts become the throughput bottleneck.

I think above things should be sorted out first for belts... and then see if people still refuse to use belts. Because above reasons are why I am using bots over belts since I couldn't get around those limitations using belts, at least not in a convenient / satisfying way.


Though I say if the bot access to a logistic chest would be limited in a way that it matches with the speed of how long it takes for an inserter to unload onto a belt/load from a belt then it might actually become somewhat balanced.

But the question is... do we really want that? ... And if so... Why exactly?

But if the throughput isn't the problem at all... then why even change anything?

I think that the bigger question for the belts vs bots topic is... the role that belts should have and the role that bots should have. And before we can discuss on what to do... their roles should be clearly defined.

[Aeternus]

Yeah I only crack as necessary, but still it happened to me that I ran out of Heavy Oil a while ago and I don't even know why it happened as I couldn't reproduce the cause.

If it was right after one or more big expansions that require blue belts... that would be why. Bulk producing blue belts can require so much lube that all your heavy oil gets converted. Logically this can be the only cause in vanilla - there's nothing else that requires massive amounts of heavy oil. I had that same problem in my factory early on because I didn't keep a reserve of heavy oil, and only a single storage tank of lube. ~2000 blue belts later there was a bit of a problem. Kicked my refineries back into normal refining for a while to solve it... then expanded the storage for heavy oil afterwards. Haven't had a problem since.

As for speed beaconized setups... I really don't get those. You add a beacon that speeds up adjacent structures but the energy requirements, resource consumption and output all increase. The same effect can be obtained by simply duplicating that production chain. I rarely use beacons as a result of this, instead setting up a serialized production line, preferably with Productivity modules wherever possible. This slows the line down (solving the "cant get stuff to/from there fast enough" problem) while increasing output and decreasing material cost. It does require a very large amount of power, but that's what nuclear plants are for

Beaconizing with speed... useful in rare cases imho, when something can't be parallelized and has to work faster. Rocket launch facility comes to mind. Most of the speed beacon setups are at most space savers, but going over the top with these can bottleneck at the item transfers.
Beaconizing with efficiency... only effective if you can save the energy the beacon costs, which is 450kw continuous. Good luck with that...
Beaconizing with productivity isn't possible sadly.

[Deadly-Bagel]

As for speed beaconized setups... I really don't get those. You add a beacon that speeds up adjacent structures but the energy requirements, resource consumption and output all increase. The same effect can be obtained by simply duplicating that production chain.

I once wondered the same thing, actually when you use Productivity Modules the speed of the machine is cut by 15% additive. In tier 3 Assemblers you can have 4 modules so the machine is only running at 40% speed. To boot, each module adds 80% energy consumption so you're also using 420% power. Applying a single Speed Module Beacon to increase speed by 50% (two modules at half effect) also increases power consumption by 70%, so from the Productivity values you're getting 125% (0.4 + 0.5) extra speed for only 16.7% (4.2 + 0.7) additional power. On paper, the more Speed Beacons you add the more efficient the machine, in reality there is of course the cost of the beacons to consider but generally the more beacons, the better. In addition, a single beacon can easily boost 5 machines containing 4 modules each to 125%, so for the price of 2 modules you've gained the effect of 5 * 4 * 1.25 = 25 modules.

[MeduSalem]

As for speed beaconized setups... I really don't get those. You add a beacon that speeds up adjacent structures but the energy requirements, resource consumption and output all increase. The same effect can be obtained by simply duplicating that production chain.

I once wondered the same thing, actually when you use Productivity Modules the speed of the machine is cut by 15% additive. In tier 3 Assemblers you can have 4 modules so the machine is only running at 40% speed. To boot, each module adds 80% energy consumption so you're also using 420% power. Applying a single Speed Module Beacon to increase speed by 50% (two modules at half effect) also increases power consumption by 70%, so from the Productivity values you're getting 125% (0.4 + 0.5) extra speed for only 16.7% (4.2 + 0.7) additional power. On paper, the more Speed Beacons you add the more efficient the machine, in reality there is of course the cost of the beacons to consider but generally the more beacons, the better. In addition, a single beacon can easily boost 5 machines containing 4 modules each to 125%, so for the price of 2 modules you've gained the effect of 5 * 4 * 1.25 = 25 modules.

Exactly. Speed Beacons cut down on the power consumption of machines with Productivity Modules.

In theory you can put Beacons so that each machine is affected by 8 beacons and each beacon affects 8 machines... making it a convenient ratio of 1:1, which makes it easy to just add the Beacon consumption to the machine consumption.

More in-depth theory about the ratio

Okay, realistically speaking you never get a perfect 1:1 ratio due to machines and beacons on the edge of the setup not meeting the 8:8 ratio.

But if on the edges the ratios are 6:6 or 4:4 or 2:2 you can balance and min-max the edge effects a little bit. It is a little bit of toying around and even more in-depth calculation of what seems to be best, which I will leave out for now.

Also the influence of the edges becomes smaller the bigger the setup is. An efficiency above 90% is achievable with 40x40 beacons, 86% with 30x30, but even less will do when looking down at the following calculations below.

Also the more "square" it is, the better it is. By "square" I don't mean that the layout resembles a square but that you have as many rows as a row is long. So for example if you have 20 beacons in a row you should also have 20 rows of beacons, which actually results in a rectangular shape due to the assemblers between the rows, but it is more important that the Beacons are "square". Squares have the shortest circumfences of all rectangles in relation to their area, hence why you should aim for that to minimize edge effects of beacons.

The following numbers are directly taken from within the game:

Calculation for Assembler 3

Calculation for Speed Beacons:

Machine Power consumption:

Base Machine consumption = 210kW
Additional consumption including PM3s and Speed Beacons = +880%

Machine = 210kW * (1 + 8.8) = 2058kW
Beacon = 480kW

Total = 2058kW + 480kW = 2538kW

Machine speed:

Base Machine speed = 1.25
Additonal speed including PM3s and Speed Beacons = +340%
Productivity Bonus = +40%

Machine = 1.25 * (1 + 3.4) * (1 + 0.4) = 7.7

Power per Speed:

2538kW / 7.7 = 329.6

Calculation without Speed Beacons:

Machine Power consumption:

Base Machine consumption = 210kW
Additional consumption including PM3s = +320%

Machine = 210kW * (1 + 3.2) = 882kW

Total = 882kW

Machine speed:

Base Machine speed = 1.25
Additonal speed including PM3s = -60%
Productivity Bonus = +40%

Machine = 1.25 * (1 - 0.6) * (1 + 0.4) = 0.7

Power per Speed:

882kW / 0.7 = 1260

Comparison:

1260 > 329.6

Calculation for Chemical Plant

Calculation for Speed Beacons:

Machine Power Consumption:

Base Machine consumption = 210kW
Additional consumption including PM3s and Speed Beacons = +800%

Machine = 210kW * (1 + 8.0) = 1890kW
Beacon = 480kW

Total = 1890kW + 480kW = 2370kW

Machine speed:

Base Machine speed = 1.25
Additonal speed including PM3s and Speed Beacons = +355%
Productivity Bonus = +30%

Machine = 1.25 * (1 + 3.55) * (1 + 0.3) = 7.39

Power per Speed:

2370kW / 7.39 = 320.7

Calculation without Speed Beacons:

Machine Power consumption:

Base Machine consumption = 210kW
Additional consumption including PM3s = +240%

Machine = 210kW * (1 + 2.4) = 714kW

Total = 714kW

Machine Speed:

Base Machine speed = 1.25
Additonal speed including PM3s = -45%
Productivity Bonus = +30%

Machine = 1.25 * (1 - 0.45) * (1 + 0.3) = 0.89

Power per Speed:

714kW / 0.89 = 802.2

Comparison:

802.2 > 320.7

Calculation for Furnace

Calculation for Speed Beacons:

Machine Power consumption:

Base Machine consumption = 180kW
Additional consumption including PM3s and Speed Beacons = +720%

Machine = 180kW * (1 + 7.2) = 1476kW
Beacon = 480kW

Total = 1476kW + 480kW = 1956kW

Machine speed:

Base Machine speed = 2.0
Additonal speed including PM3s and Speed Beacons = +370%
Productivity Bonus = +20%

Machine = 2.0 * (1 + 3.7) * (1 + 0.2) = 11.28

Power per Speed:

1956kW / 11.28 = 173.4

Calculation without Speed Beacons:

Machine Power consumption:

Base Machine consumption = 180kW
Additional consumption including PM3s = +160%

Machine = 180kW * (1 + 1.6) = 468kW

Total = 468kW

Machine speed:

Base Machine speed = 2.0
Additonal speed including PM3s = -30%
Productivity Bonus = +20%

Machine = 2.0 * (1 - 0.3) * (1 + 0.2) = 1.68

Power per Speed:

468kW / 1.68 = 278.57

Comparison:

278.57 > 173.4

Boom. There you have it.

• Assembler 3s with PM3s and without Speed Beacons are 3.8 times the power hogs than a beaconized setup.
• Chemical Plants with PM3s and without Speed Beacons are 2.5 times the power hogs than a beaconized setup.
• Furnaces with PM3s and without Speed Beacons are 1.6 times the power hogs than a beaconized setup.

Apart from the fact that you need more space (though space is usually not a problem) and a lot more infrastructure altogether (which may become an UPS problem because more inserters to update, more machines to update, much bigger power plants to update, more of everything to update etc).

That is why I recommend to never use PM3s without also using Speed Beacons.

Since I also recommend using PM3s to save up on resources in the crafting cascade... PM3 with Speed Beacons have become basically mandatory in late game.



And this is the root of the problem for why belts can't keep up with bots. The problem is not the bots, but the Belts which just can't deal with Beaconized Setups properly due to all kinds of awkward restrictions and limitations I already mentioned in my previous post.

[Deadly-Bagel]

I like how you were wildly off topic until you tied it in with the last sentence xD

Using belts with Prod modules does get a little awkward but it's not too bad, 5 lots of input will produce 7 lots of output. 2 lots of input will product 2.8 lots of output. The trick is what you intend to do with said output I guess - bus factories become onerous when you start looking at 7 Electronic Circuit belts - that's 5 Iron belts and 7.5 Copper belts -just- for Electronic Circuits, to say nothing for what they're to be built into. Sure you could just slam down a load of Assemblers and logistic chests but where's the OCD organisation that defines a Factorio player? If you want a lazy approach just have a modular setup and ship around Circuits by train.

[MeduSalem]

I like how you were wildly off topic until you tied it in with the last sentence xD

Well, couldn't help it... for some stuff to be explained you have to speak verbosely to eventually get to the point.

But that said, the PM3/Speedbeacon stuff is the reason why I abandon belts eventually. It is my perspective on things... and why I lose the interest in fiddling with belts.

Using belts with Prod modules does get a little awkward but it's not too bad, 5 lots of input will produce 7 lots of output. 2 lots of input will product 2.8 lots of output. The trick is what you intend to do with said output I guess - bus factories become onerous when you start looking at 7 Electronic Circuit belts - that's 5 Iron belts and 7.5 Copper belts -just- for Electronic Circuits, to say nothing for what they're to be built into. Sure you could just slam down a load of Assemblers and logistic chests but where's the OCD organisation that defines a Factorio player? If you want a lazy approach just have a modular setup and ship around Circuits by train.

I agree, it depends largely on if you are going to use the output directly afterwards or if you intend to feed a bus with it.

Bus systems are so... urgh. Once you have laid out a certain bus width you can't expand on it anymore without having to move the entire factory around, which just gets tiresome. You have to design everything in advance and then that is what you have to deal with if you don't want to tear down the entire factory again. Bus systems are just boxed in, no matter what.

Also I thought about using Belts and bots in the opposite way they were "intended" (if there even is an intention in the first place)... Using belts for short range (<50 tiles) transport within a factory module and using bots to do the bulk transport between those modules to overcome the bus problems. But then again... why don't use bots for everything already, doesn't make much of a difference anymore at that point.

And actually I try to avoid using trains within the base at all costs... because trains are even more fucked up due to bottlenecks in the terminals or due to bottlenecks on intersections and you can do even less about these problems than you can do about belts or bots, hence why I only use trains to haul in resources from the outposts and dump them off into the storages at the edges of my factory.



If I have to make a prediction then I think the future for hardcore megabases lies within a design of vanatteveldt, using modular science outposts/modules that run independently from each other and only require raw resources to be fed:

.15 marathon 400 spm with modular ore2science plants
Compact/boosted ore to 1/s science

It avoids the belt bus problems and could probably also be done completely without bots if necessary.
And with my suggested improvement in the first thread it also largely avoids the train issues.
I think it allows the design to be scaled to any size.

I say that is the result of all the limitations and restrictions the game has.

[bobucles]

It is difficult to calculate the real value of the productivity bonus in terms of energy. The power of getting 40% more items gets stronger with higher tech recipes and will easily translate into additional energy savings. For example when a high tech science output gets multiplied, that represents a bonus of about 170 copper and 90 iron. The total energy saved includes the cost of mining, smelting and crafting the intermediates all the way up. So even if your assembler is cooking away 3MW of power you can still be saving energy in the long run.

In general the higher tech recipes will give a larger return from productivity modules, both in terms of resources saved and thus energy saved. It is likely that low tech recipes will be wasteful on energy, but it will still reduce total belts of resources you need to supply an entire base.

[MeduSalem]

It is difficult to calculate the real value of the productivity bonus in terms of energy. The power of getting 40% more items gets stronger with higher tech recipes and will easily translate into additional energy savings. For example when a blue circuit gets multiplied, that represents a bonus of about 170 copper and 90 iron. The total energy saved includes the cost of mining, smelting and crafting the intermediates all the way up. So even if your assembler is cooking away 3MW of power you can still be saving energy in the long run.

In general the higher tech recipes will give a larger return from productivity modules, both in terms of resources saved and thus energy saved.

Yeah, obviously one saves also energy through the crafting cascade when using PM3s.

But that would only matter in direct comparisons between Setups using PM3s and setups not using PM3s...

... while above setups both use PM3s and the comparison was was basically about setups using Speed beacons and setups not using Speed beacons. Which means energy saved in the crafting cascade is the same, it's just the speed beacons that make the difference above.

[Ghoulish]

There is 1 simple change from a mechanics perspective (not sure how difficult to implement): Add bot collision.

I doubt this would happen, as it would load up the cpu even more, which is something I would imagine the devs want to avoid like the plague. (edit) This is why they were considering changing the charging time of bots, as it would be a 'solution' which is already in the game, it'd just need a tweak to the timing.

[bobucles]

I doubt this would happen, as it would load up the cpu even more, which is something I would imagine the devs want to avoid like the plague.

This desu. Bot collision is a disaster in the making. Even though it will punish tight bot bases in a very effective way, the UPS cost of checking bot collision all the time is simply too high.

The discussion is drifting a bit off topic, but most of the discussion on bots v belts has already been beaten to death. In general there are three major types of suggestions:
- Nerf bots through simple stat changes to make belts seem better.
- Change bot mechanics to increase overall difficulty.
- Buff belts to increase competition.

Option 1 is an easy option that will definitely work. However it will also upset players that lose effective designs and potential RPMs. Option 3 is my favorite, and out of the many flavors of #3 my personal favorite is upgrading belt slots to hold small stacks of items. It's kind of like the bot cargo upgrade, except for belts! Option 2 has the most varied discussion, but I haven't really seen an idea I can get behind.

[oracleofepirus]

- Nerf bots through simple stat changes to make belts seem better.
- Change bot mechanics to increase overall difficulty.
- Buff belts to increase competition.

Regarding #2, add a minimum distance between roboports. Increase number of charging ports to compensate, as it's abysmal to begin with. Adding a minimum travel distance has a qualitative effect on all decisions regarding bots.

Also, there's no reason any changes must be limited to bots and belts. Everything in Factorio is connected to everything else, so a small change in an unrelated spot can theoretically balance bots and belts. I'd advocate making efficiency modules affect the energy usage of beacons themselves. In the current state, we might as well glue the speed modules and beacons together and call it a single entity. We could just make beacons into "Speed Beacon 1, 2, and 3". There's actually no reason for beacons to have slots either, as you only ever use 4 speed modules. Belts and bots obviously have different endgames, and beacons need to reflect that, not shove the two together. The endgame for belts could be "power usage so low, the assembler grid includes its own power source, and you don't need an electric grid," or something.