Friday Facts #224 - Bots versus belts

[jcranmer]

Bot throughput is not infinite. Roboports can only charge 4 bots at a time and the more bots per tile you have the more congested they become. You can easily have hundreds of bots queued for charging or going ridiculously far to find a less used one. Since each bot in flight still counts as filling an order, it can quickly halt a factory that uses too many for too much.

If you design your roboports optimally, you support about 100 bots per roboport, and each line of roboports is worth approximately 10-20 blue belts depending on research level. Yes, overloading the network crashes throughput dramatically (robots normally only charge 75% of their energy, so letting them die completely means you need to overprovision by a 3rd just to keep even with the backlog). But if you follow the basic rules of mass roboport spam and keeping robot networks disconnected, the throughput capacity is pretty much infinite.

[Fahrradkette]

200 years ago, people had to travel by foot or in beast-driven carts because the didn't yet research neither cars nor planes.
For people back then, it was an appropriate way to move. While today, using cars and planes suits us better.

So if somebody who is posting his megabase says that bots are great, it mainly applies to his goal, which is basically to show off the amazing performance of the game.
In the other hand, speedrunners don't go for bots since they have a different goal.

I believe that nerving bots simply because one can do amazing stuff isn't the right approach.
Please don't go down that path you started by nerfing barrels but not buffing/improving pipes.

PS: don't worry too much about balance since mods can handle balancing quite well.

[jockeril]

Anyone player who argues that OPTIONAL content should be REMOVED from the game simply because they do not enjoy using it is, in my opinion, being selfish.

Bots do not impose their selfish goals upon you, bots live to serve. If the player doesn't want bots, they won't appear out of nowhere and ruin the player's game. Removing bots from the game would be a huge mistake because they do not pose a threat whatsoever to players who do not like to use them, while providing all other players with satisfaction and enjoyment.

I will forever be of the opinion that bots > belts, and removing them from the game would certainly destroy my large desire to play the game. Please do not remove the bots.

Who said anything about removing the bots ? (I'm assuming you mean the logistics and construction bots) they are an integral part of the game progression - when a factory outlives a belt system you move to a logistics network and bots become the tool for ferrying items around,

Removing bots would mean removing the logistics network system and that really ruins the fun of the late game. I doubt the devs are considering a change like that

[Jaedar]

I have admittedly not read all the pages of this thread, so apologies if this idea has been suggested and refuted before.

What if bots were limited by collision? Today, you can have 200 bots occupying the same space, interacting with the same chest at the same time. This makes bots true teleportation. But what if bots had the same collision logic as biters (but on an "air" layer, so still no collision with base or player). That might entirely kill performance, but what if bots were limited to interact with chests one at a time, and each interaction took some time. This would make bots a lot less useful in mass teleportation. The player could of course counter this by building more chests, using inserters to merge them and such, but this is not too far away from using some kind of belt anyway.

I like logbots, but I think it would be completely fine if their niche was low throughput logistics.

[cartmen180]

Put a cap on the amount of bots and make it an infinite research.

[vampiricdust]

Make belts better?

Belts will always need to be layed from start to finish and will always need surface area to be placed on, that is intrinsic to its mechanic. There is is no adjustment you could make to belts that would make it competitive with a mechanic that has no limitations at all.

Why are they competing? Bots have limitations, they just are not as obvious or easily hit because of their intrinsic traits. The real issue comes down to what is more fun? Being forced to keep using belts that are incredibly limiting and tedious or letting players earn their progression into powerful bots? Sure, you can brute force bots and not care how terribly inefficemt spamming ports and bots, but you can do that with belts without doing anything interesting.

Bots are late game tech, expensive, and take some effort to get going. Belts you just blop down, thousands and thousands of belts. It's mind numbingly boring. Bots are a change of pace that radically accelerate construction and ease excessive complexity that only exists to be complex. Why rip out a whole portion of the game play because it does something better than belts?

[vedrit]

My problem with belts is that it's so painfully painful to filter them, regardless of system saturation.
Simply using filter inserters isn't enough. If The belt/chest the inserter is dropping onto is full, then the items don't get grabbed. This will block an inserter down the line either momentarily or perpetually (until the player comes and picks up items)
Additionally, there's only so many items that can be transport along a single row of tiles on belts. And then there's pathing. Yes bots "teleport" items...if teleportion takes a significant amount of time. Belts have to be spaghettified if going through a dense factory. Bots fly in a straight line (except to stop to charge)
So, sure. You can call bots OP. But they're also necessary. A large, efficient factory simply cannot run fluidly without them.


If you want to change whether players opt for belts or for bots, nefing bots is NOT the way to go.
Improve belts. Make the belts more versatile.

[Kingdud]

It makes sense that as your technological level increases, so does your ability to do complex things more simply. Automation Machine 3s take up the same space as Automation Machine 1s, but manage to build *far* faster, and more complex, items. It makes sense that logistics bots end up being a great way to have very dense factory layouts centered on logistics chests.

The single most important thing to remember though: The same people who build bot bases by and large don't care about your 'purism' argument or 'what's suppose to be'. They care about min-maxing. These are the same people who build complex underground belt setups to fully compress ore from smelters. The same people who bothered setting up barreling and barreling to transfer fluids. They'll do whatever is the 'best' way, because being 'the best' is what they care about. "Oh! so I can do anything?" No. Take a lesson from WoW to learn a CLEAR message that you can not do anything. The one thing you must not do is irritate these people by making things overly hard.

Right now, belts are a huge frustration because getting a belt fully compressed is so difficult. Belt spaghetti is made worse by blueprints requiring lots of 4-input materials. You either end up doing split belts, and hoping the blueprint doesn't require so much of something that you need full belts of things, or you end up having very space inefficient layouts where you've got 4 belts surrounding a factory, and a fifth belt leaving. It just gets SUPER ugly for later recipes, especially when trying to build them at scale.

So...do whatever you like, but don't irritate your most passionate players (the min-maxers). They are the lifeblood of any game, because they push the content creators to make videos, write articles/guides, and modders to mod. Piss off the min-maxers and your little EA game won't ever grow after it leaves EA.

[lottery248]

in a nutshell if you are anti-logistic, consuming a certain amount of energy is already a penalty to use the flying robots, no matter how this can be addressed easily by such as adding more solar panel, because belts do not even use energy. despite for rail, they use generally little.

actually if you want to nerf/remove the logistic robots or construction one, make this decision, if that will be implemented, to be optional when having a new map instead of leaving annoyance to the players who mainly use the logistic robots. although the current logistic robot is way too powerful, as far as i know in the post, only few anti-logistics have had used the logistic robots at least once.

for the people prefer belts, compared to bots, belts do not use power at all, and regardless of how long you use, yet bot do. ie, IMO here by power consumption, bots are no more powerful than belts. and something even worse in case you are going to migrate major amount of stuff with robots, you will normally bottleneck the power consumption unless you get a more powerful electricity production or accumulator to address them. in case you are going to solve this matter by nuclear power, then the waste of energy is everywhere as nuclear reactor won't pause the power production as on full satisfaction thus waste after that.

for the train transportation, i would believe that cargo wagon is way too small to insert stuff which i would suggest at least 2 times more storage as of now, as its size compared to steel chest, you get 8 slots less than steel chest, unless there are higher maximum speed like 480KPH i assume, it is nerfing the cargo.

you are often putting some high density bases in (minor/major) disaster, which belt transportation would not all the way fit the entire supply. assuming the base is full of certain production machines, and you could barely place more belts to split the material to be inserted for new product. that is why the flying robots are useful to solve them (in this case).

i understand that this kind of nerf is good to improve the level of challenge, but establishing them as compulsory in the future is not acceptable when the actual technology could achieve something like transporting a container with a few controller and bots. this game's timeline is kinda like hundreds of years far from now, even a drone, with enough capacity limit, nowadays can transport stuff. for an example: Amazon (if i remembered correctly).

[Nick-Nack]

• add cooldown for grabbing/dropping item (like landing / take off) and allowing only x bots accessing the building (chest)

This! If you limit interaction to one bot per second, then bot networks need more space without removing any functionality. Instead of scaling up by creating more bots (which doesn't take any space), you need to scale up by placing more requesting chests.

[Caine]

Removing bots would mean removing the logistics network system and that really ruins the fun of the late game. I doubt the devs are considering a change like that

They sort of are, check the Friday facts. Though he explicitly states he will not remove them, Twinsen seems to regret adding them and is wondering what the game would be without.

[ouch]

I think this can be solved without any balance changes, but with UI changes.

You know how you can create train tracks by selecting the start point and then select the end point? Why not do that with belts? If I could simply tell the game I want a belt to start here and end here and let it weave a path in an efficient way similar to the way train tracks do. That would take a lot of the tedium away from using them. Bonus points for it even deciding when to build underground belts when it would save belts.

I've played this game quite awhile before it came on steam and I still don't understand why I'm still placing belts one tile at a time years later. Especially given that the belts are pretty much this game's mascot or something.

And if for some reason you can't do that then why not let us strait up build underground or elevated belts on stilts? That would also take some of the tedium away.

If you add a timer preventing multiple bots accessing a chest then what's going to happen is that people are going to build chest grids with low item quantities set in each of them to make up for it. It's just going to make things a mess, cause performance issues, and even worse, those specifically trying to avoid the tedium won't have a valid solution to that and stop playing altogether.

[BiBaBeluBaB]

i think log bots are important late game

id like to have faster belts. maybe add power consumtion to high tier belt. also i like the idea of having con bots early, that also resuply the player. log bots could need space sience.

Another idea would be to add consumable ressource cost to log bots. something like using up the nuclear fuel cell. e.g. batteries that break after some amount of flytime, so that robots need to go to a repair station, where u deconstruct the old batterie and build in a new one. for the old ones u could introduce a recycling process, where u get maybe some amount or iron and or coal back. like 10 to 1 ratio or sth.

and or nerf solar.

[lottery248]

and or nerf solar.

this suggestion will often lead the late game to be vulnerable in regard of sustainability (and practically poor resources of such) if all the oil fields found fail to supply the electric network completely, as coal has less energy than oils or something and is theoretically limited.

[Manawind]

Had an idea for stacked belts... what if you could take up to 4 blue belts into a building that was 4x3 and then build sections of stacked belts that are 2x2 each (though you can't interact with the items while they are in these stacked belts).... and at the other end have an other 4x3 building that splits them back out again.

Cause I know I'd love to be able to Mux and Demux my belts for longer distances or compressing them in areas that they are passing through but aren't needed.

Here is a very crude drawing.

[hitzu]

The problem with bots is exactly that they're almost infinitely expandable. That's why they're so powerful in a way that belts cannot fight with them in the late game.

I would propose to nerf them by changing roboport mechanics.
First: make the roboport area ROUND and slightly bigger.
Secondly: make the MINIMUM DISTANCE the roboports can be placed from each other.

These changes would give more thinking about roboport placement patterns and restrict spamming them in dense recharging areas. This would naturally place robots into their main niche — providing crafting components in small quantities. I would also propose to reduce the amount of cargo size upgrades up to 2 (so 3 items at once maximum).

Additionally you can add T4 belts with x4 throughput. And add the mechanics to build them easily like the rail building mechanics.

[Caine]

Had an idea for stacked belts... what if you could take up to 4 blue belts into a building that was 4x3 and then build sections of stacked belts that are 2x2 each (though you can't interact with the items while they are in these stacked belts).... and at the other end have an other 4x3 building that splits them back out again.

Long range, high throughput transport is not really the biggest issue, we already have trains for that. Getting resources to factories in heavy beacon builds is much harder with belts than with logistics bots. Such a 4x3 building does not address this issue and thus does not provide an alternative to logistics bots.

[Jelmergu]

Bot bases are usually less complex and less interesting to look at, manage and expand.

I don't agree with that statement, I once made a save(for lazy bastard achievement) which uses bots to unload my trains. Every time a train got unloaded when I was near it, I wanted to watch the bot go to the active providers to get everything into storage. Seeing 500 bots or something all fly to the same location and then spread out is quite a fun thing to see in my opinion

[aveius]

After the fluid wagons, it looks like you guys are doubling down on kicking hornets nests, uh? x)

As a more casual player (not looking at ultra-optimizing mega bases, yet still spent a ton of time on the game), belts feel a lot more appropriate for high-volume transportation of simpler or raw/intermediate items (especially plates) that are consumed in high quantities. Bots come in for more complex recipes requiring more kind of items, but in smaller volume over time due to longer crafting times, and they do simplify what would be annoying low-throughput, convoluted-paths belt delivery.

Comparing belts vs. bots, I also came to this:

What if bots were limited by collision? Today, you can have 200 bots occupying the same space, interacting with the same chest at the same time. This makes bots true teleportation. But what if bots had the same collision logic as biters (but on an "air" layer, so still no collision with base or player). That might entirely kill performance, but what if bots were limited to interact with chests one at a time, (...)

Going even further, collisions between bots while traveling would throw a huge wrench in their efficiency in bases "abusing" them. Busy intersecting/overlapping paths would be severely penalized, while continuous beelines would likely be less efficient than... a belt.
Part of the reason belts & inserters are limited is that you have a single layer of them (sort of 2 using underground ones), while bots have virtually infinite airspace layers to move through. Doesn't quite feel right looking at it from this angle, even if Factorio takes liberties with realism.

It definitely would make for a very different game though, on top of the potential to hurt performance.
Going on a limb, but if air bots collision was to become a thing, maybe a late-late-game research could be some sort of "quantum mechanics" research allowing for bots to pass through each other, allowing for the current gameplay?

[vampiricdust]

Roboports cost 45 red circuits, 45 steel, 45 gears. That is 305 iron, 225 copper, and 45 plastic. To spam them is incredibly expensive, not to mention each one can go from ~50 kw to 4mw power consumption. In active logistic factory, roboports will utterly dominate power consumption and draw more power than most of your entire factory. Inefficient logistic bot base designs will waste tons of power which causes you to spend a great deal more on building up your power generation then you ordinally would have to have 96 solar panels for every roboport charging 4 bots at peek power output (if there was no night). So ignoring the accumulators, so each roborport requires another 3360 iron and 2640 copper in panels assuming no night.

Times all these resources times 100 to "spam" roboports and you've spent 743,000 resources to build 100 of them with just solar panels and roboports. That is ignoring accumulators, power lines, logistic chests, and the bots themselves. Not to mention you need 28,800 tiles of space for just the solar panels.

Same number of resources can make 14,427 blue belts and takes up.less than half of the space. Again, that is just for roboports and solar panels, completely ignoring the actual bots and chests needed to make this work. Each logistics bot costs as much as 3 belts each. The belts require no further power management and will not stop working in a blackout. Belts do not stop working to recharge nor do they need research to move 8 items at time for 13.3333 items per second. For bots that means 3 bots per second, with 5 item researched, have to move items to surpass that. So for every second of moving those items, 13 blue belts could replace them.

Belts are cheaper and better than bots until you need tens of thousands of blue belts. Bots are incredibly expensive to have and take more infrastructure to support.