Friday Facts #224 - Bots versus belts

[Killcreek2]

At the moment:

Trains are in a great spot, #1 choice for long distance. High initial cost, but cheap to expand / increase throughput.

Bots are in a great spot, #1 choice for short distance. High power requirements & difficult to optimise, but easy to increase throughput.

Belts should be the #1 choice for middle distance, but simply cannot scale up as well in endgame at the moment, even though they are cheap, due to space requirements & throughput hard limits. <- This is the problem that needs addressing, not to nerf bots to make them "level" with belts.

So: buff the belts! Give them some love!
- Improve the throughput of the existing 3 YRB tiers, OR add faster tiers / stacked belts for endgame into vanilla.
- Compression ought be straightforward & intuitive, but still require some player effort. ie; via sideloading &/or splitter merging. I'm personally against inserters compressing belt directly, but; given the underpowered state of belts vs the other 2 transport methods, this may actually be beneficial & encourage belt use.
- Re-consider some of the "Smart" or "Filter" splitter suggestions: these could enhance / replace the clunky & large solutions for belt / lane balancing & mixed item feeding currently required.
- Maybe add a "Belt speed increase" infinite research topic. Though this would require inserter changes too ...


I like belts, AND bots, AND trains. Have used them all extensively in my factories. "Only Belts" became tedious to expand after a certain point due to space vs throughput issues. "Only Bots" resulted in an uniform & [imo] dull factory with no style, that I got bored with after a single rocket launch.
The most fun & challenging factories I have built, involved using all three transport options to their fullest!

TL;DR: Buffs to one underpowered area of the game are better than applying nerfs to its "rival".
Options are good, rail-roading players towards one option over another is just bad ~ players can choose for themselves which features they do or do not wish to play with: this is something Factorio already does right. (b'-')b

[ixnorp]

I tend to use bots in the medium to late game. I'd describe myself as a complexity-loving revamp-mod-using wanna-be Megabaser who never actually finishes any megabases. I also don't really enjoy using bots as much as I enjoy using belts, but I use them anyways. Why? Here are some of my thought processes:

• Aesthetics. Belts just look cooler and feel more in the spirit of Factorio. I like seeing things on belts and they provide an easy visual indicator about what is going right and what is going wrong.
• My late game factories tend to devolve into stamped blueprints of requester chests -> factory -> provider chest. This is ugly, kinda dumb, but undeniably more efficient in terms of space and effort spent to properly distribute materials between assemblers and the like.
• Belt balancing is a pain in the ass and I hate doing it. I feel like the game should do it for me. I know that blueprint books exist for all sorts of balancers but I don't like aspects of Factorio that I can't personally figure out fairly quickly in my head. When I look at things like a 12 lane belt balancer, to me I just see a monstrosity that represents all of the annoying parts of Factorio that force me to pause what I'm doing and break out a spreadsheet. Perhaps more information in the UI could help solve this problem? The kinds of things I want to know are mainly "how many items went past this section of belt in the past N seconds?" and I don't think I should have to make some complicated circuits device to figure it out. It's great that some people can make that, but to me it flies in the face of the beautiful simplicity of playing factorio where you can just try something and if it works, great. If not, try again. I can tinker with it like legos. That's the sort of feeling I seek from Factorio.
• Belt compression is not an interesting problem. I feel like this is the sort of tedious thing that the game should do for me while I worry about the grand design of things. I like the removal of hacky-feeling "insert onto splitters" for full compression sort of things. Factorio is about the making of stuff. If I wanted to spend a lot of time concerned about the details of exactly how stuff gets from Point A to Point B I'd go play Transport Tycoon or something.
• Fluids are still a mystery to me. I've been playing since 0.9 or so and fluids have always just felt weird to work with. There's little to no explanation of how pipe pressure works and why my offshore pump that claims to provide a zillion units of liquid per second can't supply the factory 1000 units of pipe away. I know it can't but I don't know why and I can't math it out on a spreadsheet even if I sit down and try. So for fluid applications I tend to just barrel everything and use bots to ship it around because I can rely on their throughput to be a known value or at least buffer enough that I don't worry about it.
• I feel a lot of social pressure from the rest of the community any time I read the forums or reddit where everyone is telling people with spaghetti bases that they're dumb, bad people who are playing Factorio wrong. I even stopped watching some of the YouTubers because they insist on calling anything they don't like "cheaty" and generally being extremely condescending.

These are just some of my personal thought processes and not an attempt to force anyone else to play Factorio in my preferred way.

(edit: fixing a couple of spelling and grammar errors)

[cvkurt]

Stacked belts! bring them on... I'd think adding more 3Dish stuff to game would be great.

I don't see nerfing bots help us have more fun in general or specifically with belts. I think belt enhancements would help make the belts more fun.

[Cleany]

Here is what is wrong with this debate:

I have put so much work into the smart/programmable Make-Everything part of my factory...

There are thousands of players like this who have put hundreds of hours into their factories. It stimulates a core part of the brains creational mechanism, it is "their" factory, it runs like a machine. It is a creation that they love.

And then you get comments like this:

Bots take away the fun of building and managing your factory. As soon as you have them available (and most players go straight towards it) together with the ability to do blueprints, that's it. The game becomes a copy-paste process.

That is insultingly dismissive of millions of hours of playtime.

Bots quite objectively also enable thousands of players to put hundreds of hours into their creations.

or

You are nothing more than a lazy pleb. This s a LOGISTICAL thinking game. Belts are a logistical delivery subsystem that you have to learn to use properly.

Now fair enough this was a criticism of a guy who was talking about pasting blueprints made by other people of belt balancers, but the point is that many of those who are saying that bots "ruin the game" are doing so in a matter that belittles thousands of players for whom, it clearly does not.

Logistics robots enable players to do things that belts and trains do not. Yes you can use them to replace belts entirely (I would never dream of it), but so what? How about leave them alone and stop acting as if they are inferior because they have had enough of the "challenge" of using belts to move things around in small areas already to be able to research the logistics system?

Perhaps the intent of belts should be to create a sense of misguided intellectual snobbery which you can use to be annoyed with people who think differently to you?

And trains are the most fun part of the game anyway.

[Hedning1390]

Yeah GL with belts when you play mods that have 8 item input recipes.

See what you are missing whey you can solve everything with bots. An assembler can access up to 48 lanes. 14 if you need to keep it simple and have straight belts which continues past the assembler.

Forcing players to play with belts means they have to find solutions to problems. It's like a whole new game.

[Username88]

The jump from no bots to bots seems very abrupt today maybe it would be possible to have a more interesting intermediate end game stage? The main reason bots seems so powerful is because they can fly and go anywhere, you could consider to make them land moving robots so they will be blocked by buildings and concrete could be used to speed up movement so roads have to be built to have a fast transportation net for robots still making larger bases simpler than with belts.

To give items a property of weight would also create some interesting possibilities of mix conveyor, robot and flying bots possibles since normally the earlier in the resource chain things should be heavy (ore / raw products and so on) and making everything slow down based on the weight of the items being transported. Conveyor does the real heavy lifting early in the chain while the further out you get the more viable flying bots are. (a form of this would at least favor belts in many cases since it seems very strange flying bots can transport any item at the same speed even thought A and B should have very different weight) power consumption for transporting with bots could also increase based on weight of items.

Combing these two would mean you end up with a bot network requiring land moving bots/conveyor for raw resources since they are heavy while while as you move up more with more advance components favor flying bots, inf end game research can then improve the flying bots toward/past what is implemented in the game today.

Nerfing bots alone seems like a bad idea but to give them more complex requirements than being able to fly anywhere seems like something which could add some interesting game play, improving belts in the inf research would also be nice.
Another way to mix belts and robots better late game might be that flying bots really cant fly further away from the roboport than the range of the port alone so all items must be delivered to the roboport by other means trains/belts/other and then the flying bots distribute it out to the chests in its range roboport acts as a warehouse.

[Mobius1]

Cons:
Belts:
- Waste resources by buffering them on itself
- Slow, very very slow to move items from point A to point B
- Inserters takes longer time loading/unloading items from/into it than with chests
- Any kind of belt setup takes (distinct number of recipes) times more space, which hurts FPS on big factories.
- Belts are UPS unfriendly, big factories belt-based are unplayable, production per second challenges aren't viable with belts.

Bots:
- Hard to start assembling them because belts
- Requires energy management for big factories with nuclear power hurting UPS or solar fields hurting space management
- Requires micro based logistic setups to avoid UPS hurting on a large-scale logistic network
- Requires circuit networking logic to prevent overflow or underflow of resources and to prevent waste
- Requires large usage of train network to keep the production high

I think I saw someone writing that bots takes away the logistic difficulty of the game? Wut?

Pros:
Belts:
- Cheap n' easy to make
- Can move items anywhere regardless of distance without power management or circuitry logic
- Can replace trains since its easier to add more lanes than loading/unloading on them twice for the production chain

Bots:
- Anyone can use them on any base scale
- Enable building of smaller factories with good productivity
- Can automatically supply the player with resources
- Can build things
- Can repair vehicles
- Are fast on transporting items
- Have 0 impact on inserter loading/unloading speed
- Don't waste resources by buffering them on itself
- FPS and UPS friendly, enabling to build big bases.

Now, to finish the comparison:
Players will inevitably use bots even tho they are more expensive and harder to craft due to the simple fact they provide more benefits on the long term than belts.
ANY factory built with bots produces less pollution since you don't require as many assemblers to get the same productivity, are smaller since the designs don't need 1 belt for each input/output and allows building big factories with the same FPS/UPS as a medium belt-based factory.
So, if the dev team hates the "necessary evil" the fix to that is to make belts at least as viable to use as bots, adding a cheap belt movement speed research could make the game more belt-based than bot-based. Adding more features to belts that makes them viable to be used on large scale bases can make them more important than using bots.

But untill we have better belt system, bots will always be preferred.

[leoch]

The real reason logistic bots are over-powered is that lack of collision modelling means that hundreds or thousands per minute can visit a single chest. Maybe you can find a way of limiting bot density without killing UPS but that's probably not easy.

Actually, I think you should go further and let bots build things like the villagers in Age of Empires or similar games, carrying resources and involving actual build-time, for anything bigger than inserters and belts. (Yes, if logistic bots are currently over-powered, construction bots are insanely over-powered.)

Also, belts need improvements. Loaders which produce fully-compressed belts from simple containers are ridiculously useful; they should be added somehow, along with bigger chests (2x2 or 3x3 or even bigger like the warehouses mod). Feel free to nerf chest capacity while you're at it if you like, it's hugely OP with current stack sizes and relative to train wagons.

[rantoc]

Want feedback, ok here is what i think about the bots.

Are the logistical bots OP? Yes
Are the construction bots OP? No

What could be done to keep them but at the same time balance them?
* Could add a landing time and only one bot is allowed to land / start per position making the rest hover while waiting making bulk transportation bad with bot's while still allowing them to do "special deliveries". Making them more complimentary than primary source of transportation.
* Could add a list of allowed goods that the bots is allowed to carry and the rest is simply is to bulky to be carried at all or slow them down significantly - Full speed for lighter items like say magazines and chips but slow as hell on bulkier things like ores or refined sheets/steel would just like the above allow the bots to co-exist with the conveyors but not replace them
* Add say 3 weight classes to the goods, light, medium and heavy and then allow max 1 of heavy and it slows them down to a crawl, medium weight say 1 but at ok speed and light max 2 at max speed

I always enjoyed options and with the above it wont remove the logistical bots from the game but place them where they can be useful but not as main means of in base transportation of goods

[Caine]

So: buff the belts! Give them some love!
- Improve the throughput of the existing 3 YRB tiers, OR add faster tiers / stacked belts for endgame into vanilla.

So what kind throughput numbers are we talking about? What would be decent for belts to be feasible?
Currently a fully braided belt can push 80 items per second.

[StoneLegion]

Mass Kill Bots!

[roman566]

First fluid, now bots, what next? Trains? They too move tons of stuff that could be moved via belts, so why not nerf or throw them away? Honestly, it seems that I will stop playing vanilla factorio at 0.15 and any upgrade would be with tons of mods whose only purpose would be un-nerfing everything.

[Sniperfuchs]

It's painful to see the non-argument of "well then just don't use them" over and over in this thread. A lot of people seem to completely miss the point of the discussion and don't even realize it. The topic is not if you or whoever likes bots but if they are a good addition in terms of gamedesign, which they are most likely not.

A good analogy would be Super Mario Odyssey, as weird as that sounds. In that game, you have any freedom you can think of (similar to Factorio, just play how you want). However, that means that there are some very challenging Moons to collect, but also some stupidly easy ones which literally fly around in the world just waiting to get picked up. And now you guys come in and say "well just don't pick up the easy ones and only get the difficult ones" (basically the "don't use bots" argument).
Do you not understand how this is terrible design and immediately pushes players in the wrong direction? The game suggests collecting as many Moons as possible so what kind of gamer would just ignore a reward literally in front of them because it's too easy? A VERY tiny fraction who like the challenge, that's it. Same goes for Factorio, the game pushes you to build bigger and better, yet bots are the only option at some point. The argument that you can play Factorio the way you want is just wrong, sorry for being so blunt. You wanna build a megabase? Sure, here is a bot smelter. Oh you would like a megabase with belts? Too bad, better play with 20 UPS.

The only valid point is that, as was said, it's late-ish in the tech tree. It makes sense that old things get replaced by better new stuff, but if that better stuff makes the game less interesting and challenging, it's questionable if it was a good addition in the first plae. Something so defining for Factorio such as belts being completely replaced by something so easy to use just doesn't feel right in my opinion.

Don't get me wrong, I love both bots and belts in their own regards and can see the upsides and downsides of both, but this is kind of a huge design flaw that is basically "necessary" because of megabases, at least until belts get a huge buff or we would get another solution to manage logisics that's not just "plop down this chest and a roboport, done".
And last but not least: I'm purely arguing from a gamedesign point of view, which means that while bots might be bad for the game, they might not be bad for the players. If the goal is to make the core audience of Factorio happy, everything should stay the way it is. If the goal of the developers is to make a well-rounded game that gives new players the right hints (aka bots not being the 100% superior option for everything like it is now), then things should definitely change one way or another. There is an argument for both sides.

[Mimp]

The big issue between bots and belts can be summed up in 3 words:

Throughput.
Throughput.
Throughput.

Bots can move the items. Belts can't. That alone is enough to put belts in the grave.

Bots win hands down when it comes to the difficulty of setting up high throughput networks. When a belt network grows the number of twists and turns and underground systems increase with it. Doubling the number of belts is much more than twice as difficult and the challenge grows in some kind of exponential way. I'm talking about players spending hours or WEEKS trying to set up 40+ belt lines to feed their base. Bot difficulty on the other hand stays very simple. Add more bots. Cram in more roboports. Easy. The hardest aspect of a bot network is not with the bots but rather with setting up the train network to get everything flowing. Even then it is much easier to understand a large, modular train network than it is to understand hundreds of belts crossing 10's of thousands of tiles attempting to feed a factory.

This is why I like mods that add loaders, express stack inserters, and higher tiers of belts. They all improve throughput and make belts great again. Item compression is probably required to get belts to be able to beat bots for raw throughput though.

[Engimage]

Reading and thinking even more I do see packaging the best solution possible.
Stacked (packaged) items can be only handled by stack inserters 1 at a time and can't be transported by bots.
This would repurpose stack inserter instead of grabbing piles of items which is rediculous.
Problem solved.

I do understand that you can't implement belt speed research due to belt physics of moving items by fixed amount of slots per tick

Introducing just faster belts will not help at all. You need to increase throughput drastically making belts best for transporting items in bulk.

[Zavian]

Yeah GL with belts when you play mods that have 8 item input recipes.

You can easily fit 11 separate belts feeding an assembler, plus one output belt, That is enough for the most complex recipes I've seen. (Admittedly routing all those belts, and using 8 or more beacons/assembler would be very challenging. But solving logistical challenges/optimising things is part of the appeal of factorio, at least for me).

[TroZ]

I don't mind the bots as you get them later in the game.

If you want to buff belt usage, I have a few suggestions.
Make the fast and express belts a bit cheaper. Maybe only 3 gear wheels for fast and 6 for for express, with only 10 lubricant. It takes a while to make express belts, if you don't set up a bunch of assemblers, that sometimes it's just so much easier to use bots that belts don't make sense. Having the belts be cheaper would make them favored in some more situations.
Buff splitters by having a few selectable modes, which will make them more useful. By default, they should work as they currently do. However, when you click on them there should be options for balancing as they do now, or favoring a particular side, for both input and output. A side that is favored will have output go to it first, or input taken from it first. This would allow 'priority splitter' behavior for main buses without using circuit network wires, and also allow a splitter with one output but two inputs to favor one side over another ( i.e. pass through coal to boilers until the coal runs out, at which point wood from the other input would be passed through). An option to also balance lanes could also be useful in some situations.

Also, you have some achievements for using bots, why not some for using belts? Perhaps having placed 10,000 belts or having a single item travel a certain distance by belts (5K?).

[Mobius1]

This is why I like mods that add loaders, express stack inserters, and higher tiers of belts. They all improve throughput and make belts great again. Item compression is probably required to get belts to be able to beat bots for raw throughput though.

Bot research speed beats belts on the long term, doesn't matter how great your factory is, bots will ALWAYS move more items per second than belts given you have enough bot research speed.
They will always be better than belts because this
I think I'll go with the crew and say "KILL THE BELTS" that way they won't have to bother optimising the bottleneck of factorio, there won't be any buffering issue on the game, assemblies can start accepting 276 inputs and can start outputting 12 items at a time, belts can never get that far.
So on and so forth, bots enables higher complexity features.

[GenBOOM]

Stacked belts! bring them on... I'd think adding more 3Dish stuff to game would be great.

I don't see nerfing bots help us have more fun in general or specifically with belts. I think belt enhancements would help make the belts more fun.

I really hope you are not the only other person that agrees.

[Sylos]

Things that bots can do:

• Construct blueprints
• Repair
• Deliver items to player / Take items out of trash slot
• Deliver small quantities of items to chests
• Deliver large quantities of items to chests

In my opinion, getting blueprints constructed and items delivered to the player need to be much easier to get to than they currently are. It's neither terribly fun to constantly have to gather items from your factory, nor is it particularly exciting to hand-build the same thing again and again.
Automatically repairing should be an option at some point, but I don't think it needs to be available early on. It makes it more interesting, if biters can actually breach your walls with enough work.

Delivering of items to chests can IMHO even be a little harder to get to.

My suggestion:
Make the Personal Roboport easy to research and create a new type of low-tier robot, which can launch from the Personal Roboport. This will be mostly like the "Flying Robots" from the post, except that I would limit them to the player's vicinity, so they build blueprints near the player, they repair things in range of the player and grab things from chests around the player (or put there from the trash slots). They won't be able to deliver to the player from all over the factory, just when the player is close to a storage/provider chest. This gives an incentive to research actual logistics bots, even if you don't want them to replace belts and maybe also gives incentive to bring your supplies closer together into some storage/restock area. More advanced construction bots should still be available to repair and replace destroyed buildings without the player needing to be nearby.

Or maybe don't separate Construction Robots from these personal Flying Robots, but make non-personal Roboports harder to get to. Problem with that, though, would then be that Construction Robots as well as Logictics Robots can deliver to the player.