Friday Facts #225 - Bots versus belts (part 2)

[Ifalna]

To be honest: I am very far from a megabase design.
I play marathon and right from the get go I am constantly confronted with the inadequate nature of belts and inserters.
To be blunt: these units simply can't keep up and even mechanics like "inserter needs time to grab/deposit stuff" is enough to make assemblers not work at full capacity w/o interruptions.

I have yet to find a satisfying and compact way of feeding the output of 5 copper Assemblers into one green chip factory (beaconed of course, vanilla is no problem at all). No matter what I do, at some point it will stall. Even on a blue express belt, the second copper wire factory experiences delays in depositing the items. That is with only module one. Don't want to imagine module 3.

At some point (yes I know it's OCD to want factories to pump w/o interruptions) stuff just gets way too frustrating and you go "screw this, let the bots handle it".
And suddenly, that hungry SoB of a green factory purrs like a kitten.

Belts and inserters are fundamentally flawed in this game. They introduce delays and weird behavior, don't even get me started on all the awkward loops we have to build in order to get items to a lane we want. Dear lord, we can launch ships into space and build portable fusion reactors / laser defense systems but we are too incompetent to progam an inserter (that can grab goods from both lanes!) to reverse the movements and deposit on a lane of our choosing?

Give me a break.

Instead of mourning about bots, how about you fix belts behavior and throughput first?

[uint]

There is zero reasoning to say they break the game, they just change the way you think about your base.

That's what I was trying to say, just put in other words. The game about belts and the game about bots are, in a way, two different games. So Factorio as a whole, being, by the very-late-game, about bots (as they provide higher throughput, more convenience and lower CPU load than belts) is mainly a matter of the game-design decision. Some would like it, some wouldn't. Personally I don't like the uselessness of belt-like mechanics for higher throughputs, and consequently don't like being forced to use bots as a replacement for belts. By the extension, I don't really like how useful bots are for higher throughputs (it just doesn't feel right to have both high flexibility and high throughputs in one transportation methods). But as these are my personal preferences, they don't really matter for the development of Factorio As for what would be more "fun" for the "playerbase as a whole", I have no idea.

We might as well argue the artillery train breaks the game because it does more dps than the pistol and is easier because it fires automatically, removing the challenge of clearing biter bases. So let's nerf shells to match the dps of the pistol.

It's not the matter of matching the dps, but the matter of every different mechanics having its niche and its task, which it carries better than everything else. Military is not really a good example, as it still feels like a placeholder for the most parts. But as an illustration, you have different weapon mechanics: bullets-based weapons, shotguns, fire, nukes, artillery and so on (some, like grenades and rockets, are excessively useless as for now). What for me would feel like a good game design, is all these mechanics being relevant at any stage of the game, and upgrades being provided within each mechanic (like SMG is a straight upgrade for a pistol, and uranium bullets are a straight upgrade for regular bullets). Now if by the introduction of the artillery train you have found that it performs any task better than the SMG, it would have been a somewhat game-breaking option. But you can't shoot a biter in the face with artillery, so for now SMG is safe.

Returning to bots and belts, I think that even in the lategame, belt-like logistics need to be better than bots in at least something. Then, you could decide on what to use in a particular case, and that would be more fun than having no real alternatives to the bots as of now.

[Pascali]

Thats the fun of the game. If belts or bots could do anything, therew wouldn´t be a reason for clicking on factorio on the steamlist.

I have a suggestion:
Implement an ultimate superbot in the game. You need only 1 of this. He can beam all things to anywhere in less then 1 second. And he can build the perfect setups automatically. You just have to start a game, and go out play football. After 4 hours coming back to you pc and watch the rocket fly into space. That´s fun! Buy a second copy of factorio and play football with 2 guys! After 2 hours come back to your pcs and watch the rocket fly into space!

[vampiricdust]

@uint

For me, belts and bots have their niches. Belts are cheap high capcity throughput that works without power and entirely dedicated to the path you set. Ideal for anything that needs to be moved in mass farther than a bot's flight distance without charging mid delivery. If a bot needs to charge at all, it destorys throughput and requires higher supply settings to bridge the delay.

Bots are really good with burst moving items from train stations to feed belts. Personally, I find beacons to be overpowered as they counter the production module drawbacks and people abuse this worse than bots and they force people into using bots. Of course this ties into choices in Factorio dont really matter.

Belts have no inherent logistic ability at all and using combinators to route goods is a clunky frustrating mess less you have EE experience. Things like filter splitters are a good step towards making belt logistics compete with bots.

[Mestiff]

Factorio seems to be a game which cares about realism for some extent. For me it always seemed super unrealistic that bots are able to fly through each other. So why not limit the amount of bots in a certain area at the same time or add pseudo collision detection so that bots need to fly around each other - maybe even with some safety distance?
It would nerf bots, yes, but not because of balancing, but rather realism.

Doesn't bother me on the realistic parts.... Bots flies, so they can be at diffrent heights.

[ske]

Buy a second copy of factorio and play football with 2 guys! After 2 hours come back to you pcs and watch the rocket fly into space!

This sounds perfect! Fresh air, friends, success; how could anyone be against such an innovation?

[Mestiff]

While limiting bot access to chests might sound good when comapred to limited access chest have with belts, it will create even worse monstrocities like these below (just ideas, multiply it by 100x when smart people math them out):




Once again - there is no issue with bots being too strong. There is issue with belts being too weak. In this case, instead of nerfing bot access to chests, find a way to buff belt access to them (psst, pssssst... Loaders...).

If you think it's monstruous why would you build it when there is another viable solution? The point is precisely that if you want a better throughput and still use bots you'll have to use more space otherwise you'll have to use belts.

Moreover, those images describe useless situations and set-ups.

On the first image, only the first chest on each file will take whatever item the train take here and even if bots take only 10 items/s away from the chest it will empty it quickly enough. If you manage to have full trains non-stop without rest for fully upgraded inserters and really need a second chest, the bots will not be the best solutions and you'll have to think about making a belt set-up (it's the topic's purpose, making belts the best solution over bots in some really high throughtput situations.)

On the second image, i really doubt you'll need this much chests to compensate the fast assembling machine (even with fast modules beacons which is a really ugly monstrocity) for most of the recipes. For the recipes who need a really high throughtput, then again the belt will become the best solution, which is the purpose of changing how the bots are actually working.

[British_Petroleum]

Idea: bots take time to load and unload, and only 1 bot can access a chest at a time.

[British_Petroleum]

Idea: bots take time to load and unload, and only 1 bot can access a chest at a time.

Yesssss! Also seeing a bot carrying 3 oil refineries is a joke (lol)

[Pascali]

Or make bots only work at day - so you need belts for a working system, but can do bot-fun at day.

[Ifalna]

Also seeing a bot carrying 3 oil refineries is a joke (lol)

More so than our guy carrying locomotives, multiple kilometers of concrete plates, a brick wall, assemblers, oil pumps and fluid storage tanks in his pant pockets?


BRB, planting inflatable train on tracks....

Really, the problem is: belts top out at 40 items/s.
If you need more, you're essentially screwed, since we don't even have fast/stack long arms for a second lane.
Then it means duplicating the entire setup and wasting tons of real estate.

I do agree to some extent that the actual problem aren't the bots. It's the module towers. These aggressive setups of prod/speed don't work with belts.

[Lubricus]

I do agree to some extent that the actual problem aren't the bots. It's the module towers. These aggressive setups of prod/speed don't work with belts.

From my perspective the beacons makes the belts a more fun challenge without needing to build a ridiculous amount of assemblers. Without the beacons it would be tedious to build all the assemblers and probably kill the UPS before you get to production rate that is fun to try to handle by the belts.
In a similar way I build a big town style factory with trains that had only two wagons to create a fun trains chaos without having to build an extreme amount of production.

[kalkross]

How about late game portal undergrounds, that you can link and teleport belt throughput.

[melzedan]

I thought a while about nerfing bots. But however you nerf their throughput it's so easy to compensate; "Throw more bots!"

But what if you make bots randomly drop stuff?

Even a 5% fail rate on a delivery would be a major wrench in perfectly timed setup. Such a nerf would make bot networks harder to predict and have an increased resource drain compared to belts. It would make bots a little more interesting than just another way to do the same thing as belts, but better.
Another twist could be to induce a harder penalty the longer distance the bot travels. Or a careful bot that's extremely limited but won't drop anything (for that precious satellite perhaps?)

https://mods.factorio.com/mod/Bot_Servicing

I like the thought of your mod. But since the items just gets automatically picked up by another bot, again all you have to do is just add more bots to the network to compensate. My thought was to trash the items carried thus making bot networks require more resources.

[metralha761]

I think the main issue is, you can have 1 thousand of robots inside same square, you just need to go around this by limiting the number of robots inside same square.
This will limit the amount of robots can be over chest and limit the input or output amount.
Side effect of this is, chaos when have hundred of robots when want to be drop or collecting from/to chest, and the others don't let him get outside. But as they get out of energy they need to refill, they get out to ports and will free some space to flow.
Maybe this will create some beautiful behaviors like a chaos leak.

[hitzu]

I thought a while about nerfing bots. But however you nerf their throughput it's so easy to compensate; "Throw more bots!"

But what if you make bots randomly drop stuff?

Even a 5% fail rate on a delivery would be a major wrench in perfectly timed setup. Such a nerf would make bot networks harder to predict and have an increased resource drain compared to belts. It would make bots a little more interesting than just another way to do the same thing as belts, but better.
Another twist could be to induce a harder penalty the longer distance the bot travels. Or a careful bot that's extremely limited but won't drop anything (for that precious satellite perhaps?)

https://mods.factorio.com/mod/Bot_Servicing

I like the thought of your mod. But since the items just gets automatically picked up by another bot, again all you have to do is just add more bots to the network to compensate. My thought was to trash the items carried thus making bot networks require more resources.

That would be against the main purpose of bots: carrying moderate-low amount of complicated and sometimes very expensive stuff.

[Philon]

Hi, first time poster here! I wanted to post in the big discussion thread after FFF#224 (viewtopic.php?t=56218) but that got closed before I read the FFF.

So I agree that bots take away a huge part of the complexity in the game and for me that's not desirable. I always want to find a way to do it with belts, or trains if necessary.. But using bots is really nice for "small and annoying" productions. Those times where you haven't worked out the logistic complexity of getting a large production going but still you need a steady supply of a product and you need it NOW This seems like a very fair use case for bots and I like to use them like this.

If bots had more realistic collision physics then I think this whole discussion would not be taking place. Imagine if bots could never overlap mid-air! They would keep their benefit of flying over belts and assembling machines to supply whatever they want wherever they want, but using bots would not scale up very high. They would be in each others way when crossing paths, there might be cues of robots waiting to reach a chest, there would be cues of bots waiting for recharging. With 100s of bots around the same area, you might even get "lock up" type situations where everything stops moving (imagine pictures of traffic jams in india ! Meanwhile the graphics of a factory would not look buggy with massive stacks of bots flying around, so visually this would even be a bugfix.

Additional changes that could nerf robots or improve belts:
- let robots crash when they run out of fuel. Crashed robots could either be lost or be recovered by other bots.
- have blue belts go at x4 speed (up from x3 currently) so upgrading to each new tier is a doubling of capacity and is easier to plan for
- make it easier to compress belts by ensuring that all gaps on a belt are big enough for an item to be inserted into it
- you are adding priority splitters! this one is massive!!!! thanks

I hope this post is in the correct area of the forum and that I haven't violated any forum rules Would appreciate any kind of feedback from the community. Thanks for reading!

tldr: stop robots from overlapping in the air

[TheRaph]

I have a suggestion:
Implement an ultimate superbot in the game. You need only 1 of this. He can beam all things to anywhere in less then 1 second. And he can build the perfect setups automatically. You just have to start a game, and go out play football. After 4 hours coming back to you pc and watch the rocket fly into space. That´s fun! Buy a second copy of factorio and play football with 2 guys! After 2 hours come back to your pcs and watch the rocket fly into space!

Like this, and if not implemented in vanilla game, someone should make a mod.
And if Wube will implement this, they should make a second game - ego-shooter which plays itself, and a selfplaying race game. (Which will results in more young people playing football, soccer and so on - increasing health and fitness ...)

[thatguy321]

To make bots more interesting/challenging, why not require them to have maintenance done? This would also make it more realistic as in real life a flying machine can not go about infinitely without requiring some sort of maintenance.

There could be a separate bot repair/maintenance facility. You would have to feed it with lubrication, circuits, gears, etc. and then using circuit network conditions have some sort of strategy to shuffle your bots in and out of there.

Bots performance (speed, ability to carry multiple items, etc.) would degrade over time until they eventually fail and crash into your factory damaging whatever they land on unless they are maintained.

I'm not sure how CPU intensive it would be to track the maintenance level of 1000s of bots though.

Thoughts?

[Philon]

Interesting idea!

not sure about requiring the user to use circuit conditions though. I think it would beed to be an automatic process though. Like bots starting to blink in yellow or red before they finally crash and burn. Also you would need some visual indicator of which bots are under-maintained and which areas in your base are missing repair stations. with all of that going on, I'm not sure if the feature might become too big of a deal gameplay wise to still make sense..