Friday Facts #224 - Bots versus belts

[MINIMAN10000]

With Speed Modules being the prodominate Beacon choice, this contributes to the "belt vs. bot" rift too. As it's making production machines faster, putting out more items that the belts just can't provide throughput compared to the research upgraded bots. Solution: Limit number of effect sources an entity can receive based on tier. If over the limit, then no further bonus received.

That doesn't stop you from infinitely scaling throughput by just building more machines. Speed modules only make things faster. Building more machines also makes things faster.

The thing that makes bots the best option is they have unlimited potential throughput teleporting items.

The thing that makes belts the worst option is they have limited throughput and no way to improve it.

[JohnyDL]

Bot Nerf + Fluid Nerf + Artillery? => better idea

Make certain things more limited than others and only in certain places. We've seen it with Artillery ammo already, they have a stack size of 1 but a train wagon with a huge super gun on it has more capacity than an otherwise empty cargo Wagon. Why not just change the max stack of certain items at certain times. Here's some examples:

Barreled Liquids + Bots or Trains is OP? well change their max stack size of full barrels to 1 per slot in trains and 1 per bot because they're too heavy/compressed to be hauled while in crates it's about volume not weight so a stack size of 10 might be fine.

Similarly there could be a 'Boxed' item set for items like barreled fluid set. An example might be to take a wooden crate and a stack of an iron effectively compressing it. Now 1333 iron plates can be moved by a regular transport belt each second as a late game addition to rival bots but their drawbacks might include only the more heavy duty stack inserters can move them and only 1 at a time, they're too heavy for bots outright. Yellow belts have no issue transporting Boxed items (thus solving to some degree the belts being inherently worse) though Red and Blue might require extra research. Boxed items have a stack size of 1 so making them worthless for increasing stack density in chests or trains, maybe for players they have a stack size of 10 but they can't be crafted or uncrafted by the player. (Oh and a bonus fewer items moving around on huge maps = more UPS)

Satellites are delicate things, they shouldn't be moved by belts (which might shake or damage them) or bots (which might make sudden slows or stops jarring delicate equipment) so only the slowest most precise of inserters (the yellow ones) can move them, one at a time, between the assembler(s), box(es) and the silo(s), not even the player can move them by hand.

Landfill doesn't stack high enough for some people, the so instead of a set size in player inventory first slot has a stack size of 100 but each extra one gets a 10% stack bonus over the last, 100, 110, 121, 133 etc. If you're carrying round a little land fill it's for emergencies if you're carrying round a dozen stacks of it or more then you plan to fill in a large area having the ability to carry more would be useful the more you want to carry the more you'll be able to carry would be a huge QOL, for me at least, over dozens of trips for more landfill

This might be a better solution all round than just cutting down game and it adds options for new puzzles like the satellite one I suggested.

[GenBOOM]

I like belts, but bots can achieve infinite item throughput over a path. Belts cannot.

I like belts so much that I sometimes put cars on belts to increase the belt's total throughput. If the cars would flow through splitters and underground belts, I could stop using bots.

What I really want is a palletizer--something to compress a full chest of items into an item that takes a single space on a belt.

Maybe I can mod that...

there is already a compression mod out there

What about indroducing diagonal belts?

I would underground corners if more often if diagonal belts existed.

I'd just like to point out that the slow moving, extreme capacity belt system already exists in game. Build a car, drop it on a conveyor, and use inserters to manage its inventory. Add throughput with more cars. Easy.

The item throughput of a "carveyor" is absolutely insane. A single belted car has 80 slots of inventory space, and the slowest belt can easily push a dozen around per minute. The main weakness is that inserters are too slow. Cars are a small target and there's a very low limit on how many inserters can manage their incredible inventory space. You really need the high speed capacity of a Loader to keep the items moving at any decent speed. A secondary concern is that it builds up a huge inventory in limbo. That's a big issue for a smaller base where every item matters, but an endgame RPM base doesn't really care if there's 2 rocket's worth of items in the pipeline. A huge inventory lag is a fair price to pay for moving 20 belts of items at once.

Now I know that you don't want to drop cars on your conveyor belts. First off you can't blueprint it, so that's a huge pain. Second is that you need circuit conditions to manage the car inventories properly, and that's way beyond the expectations of an ordinary player. But consider a different type system which is basically a mini rail with mine carts on it. Huge amounts of a single item fill up a cart, and it glides down to where it gets unloaded. There are many similar suggestions for some kind of "bucket conveyor" that can achieve extreme throughput for those parts of the base that need it. I know that ore and plates aren't difficult to manage on their own, but belts can't handle moving the sheer bulk that a late game base needs. A bucket conveyor system can solve this. It needs to move a LOT of items, but the MIX of items doesn't matter because it's a lot of one thing and the horizontal SPEED doesn't really matter because the player needs throughput.

haha thats awesome, I'll keep this in mind

The strength of belts is that they move a quantity of items from point A to point B.
The strength of bots is that they are incredibly flexible and responsive to demands.

I think factorio would be best if it emphasizes these strengths, such that:
Belts are virtually always the optimum answer for moving a known quantity of set materials from point A to B.
and Bots are virtually always the optimum answer to problems that ask for flexibility, compactness, or coverage for 'spikes' in demand. (eg, supplying turrets with ammo, since running ammo belts to turrets has always been tedious and too simple to be fun)

And then, for advanced players, there should richness within edge cases where it turns out belts with circuit logic are better than bots, etc.

To help the strength of belts, I think other commenter's suggestions on stacking, etc, seems promising.

woo, we have another one

hahaha



http://lapsedhistorian.com/get-blower-l ... ing-tubes/

this is actually a good tech alternative that requires steam power and solves the problem of ammo, repair packs, and other small orders.

+1, brilliant idea.

^.^ yeah I'm glad I found that article, its actually super interesting

[Zavian]

I'll admit that I haven't used a compression mod, but we need more than just a solution for higher throughput on the main bus. We also need it for assemblers in between beacon rows.

The problem with a uncompressed items -> assembler -> compressed items -> assembler ->uncompressed items is that it won't really address the problems of beaconed belt builds needing very high throughput. Think green circuits at 5.5 craft speed, with productivity modules that eats 33 copper wire, 11 iron plate, and produces 15.4 items per second (if you can keep it fed). When you want to have a line of 8-15 of those assemblers in between 2 rows of beacons, you want high throughput belts. We would need either support for grabbing uncompressed items straight off of a belt of compressed items, and transparently compressing items when inserted onto a compressed belt, or recipes that took compressed items, and produced compressed items. (The later is probably possible for a mod).

[GenBOOM]

high throughput is best done with 0 belts and 0 bots. take directly from a train into assemblers.

[Zavian]

Try fitting trains, inserters and assemblers between 2 rows of beacons so you can get 5.5 craft speed. There isn't enough room.

[martinez_21601]

I have been playing Factorio since 2012 and have built both belt bases and bot bases. I think Logistic Bots should be left just as they are, they are late enough that you need to make almost everything with belts before they open up anyway. It should be the option of the player whether or not they want to supplement their factory with bots or not. I think many of the recent changes have been bad decisions and I think at least for me the game is getting ruined instead of improved. I do not think the bots are "game breaking" because if you don't want to use them you don't have to. I also think that at least for me bots don't get used much until I already have a big factory that is belt based it doesn't make sense at all to nerf the bots. Over the past six years, I have spent thousands of hours playing Factorio over 500 hours just since the steam release and I have tried and adjusted to most of the changes I disagreed with even when you guys broke my Factory by changing the science recipes mid-build. Bots to me are a way to supply myself and create items that I need late game the fact that the technology comes even after processing units says they don't really have any effect on actual gameplay other than to make things easier for veteran players. This is a change I am speaking out about and if it happens I personally will uninstall Factorio and move on. I'm sure I am not the only one in the community that feels this way. Also, if the problem is you want to emphasize belts maybe do a research tree that speeds up belts or as I saw mentioned in an earlier post make some type of mini-rail system to eventually replace belts.

[TiMatic]

I like belts so much that I sometimes put cars on belts to increase the belt's total throughput.

I like that idea. I think there are some critical areas in my base that I can solve with car belts.

[martinez_21601]

I just wanted to add I think of bots as my reward late game for getting everything automated on belts first. Bots are not easy to get to especially for new players and once you have them you should be able to pat yourself on the back and feel rewarded for accomplishing the end game. Not many things are left to be built by the time you have bots anyway it's more of a tool to scale things up at that point in my factories anyway.

[GenBOOM]

Try fitting trains, inserters and assemblers between 2 rows of beacons so you can get 5.5 craft speed. There isn't enough room.

you dont need 2 rows. just add another train next to the first one. get double use out of your single row of beacons and spam trains.

[Killcreek2]

I actually think the base shown in the FFF #224 is actually a very nice design and took a lot of thought and planning to build. I have not looked at it in detail, but I suspect it has some very well planned out logistics networks to minimize bot travel, etc.

Agree wholeheartedly!
You can see the effort put into that factory just by looking at where each product is crafted relative to others in the production chains. It is far beyond "spam roboports & dump enough bots in". Props to ultramn.

& to the people who do NOT use bots "because they are cheaty" or "easy-mode": Have you actually used them on a large scale to the point where you understand their limitations & challenges? Making an IN-efficient bot factory is easy, but so is making an in-efficient belt factory...
Making an efficient factory using bots brings its own unique design & planning challenges, in addition to their already high power use & setup costs. Just watching a video of a knowledgeable players bot factory, will not really show you the amount of planning or understanding required to use them well.

I think putting a bot cap PER network would add some thought and planning to logistics bots. It may also be a somewhat simple fix.
/snip/
... there are precedents for bot caps in the game, your personal network limits combat bots and research increases these limits.

Link to original post.

This is probably the best suggestion I have seen so far to put "soft-cap" on bots during midgame, yet still allow endgame / megabase use. I have had a similar idea [but he beat me to posting it], so here are some rough example numbers to quickly illustrate:

50 active bots per network +10 per connected roboport. With lategame research to increase support per roboport, eg; +5 per level.
So, with 8 levels of research, it would be 50 +50 per roboport. With 17 levels, it would be close to the "ideal" charging ratio of 100ish bots per port.

The personal roboports mk 1&2 already have a hardcap on supported bots, in addition to the combat robot limit mentioned above. It would be great to have both personal & factory roboports unified to use the same "soft-cap" system, & they could even share the same upgrade research.


I think a major reason bots are seen as a "better" alternative is due to the new v0.15+ science update: item throughput requirements have increased dramatically, compared to v0.14. Bots could already cope with the increased demand quite easily [as could trains!], but belt throughput has not been updated to match & they are now struggling to meet the new levels of demand effectively.

I'll say it again: Please buff the belts!
A "Stacked belt" seems a great lategame option: if an inserter can grab a whole stack at once from a belt [as it can from a chest], rather than waiting to scoop up each item one-at-a-time at belt speed, it would help with the belt>inserter assembler feeding issues when using beacons [which is currently "resolved" by using bots to keep it compact, or with extra chests/inserters often at the cost of beacon coverage].
Ohh, the interesting sushi possible with a stacked belt & inserter/combi controls makes the mouth water...

[Riph]

While I hate to tell anyone "You can't have that thing you like," my personal dislike of logistics bots does make me think about what we'd have if that devtime had been spent elsewhere.
If they hadn't been coding logistics bots, would we have more varieties of belts? Programmable splitters that only grab the left/right side of the belt? Diagonal belts? Corner-sharped underground belts? The fact that logistics bots easymodesolve all the logistics problems means that there was never a true need for the above items, and they'd have been great to have.

[GenBOOM]

While I hate to tell anyone "You can't have that thing you like," my personal dislike of logistics bots does make me think about what we'd have if that devtime had been spent elsewhere.
If they hadn't been coding logistics bots, would we have more varieties of belts? Programmable splitters that only grab the left/right side of the belt? Diagonal belts? Corner-sharped underground belts? The fact that logistics bots easymodesolve all the logistics problems means that there was never a true need for the above items, and they'd have been great to have.

half of those are mods already. shouldn't be hard. its not about time to implement

[AlexTheNotsogreat]

I wonder if you could make it so that bots would have to maneuver around each other to avoid overlapping, like with biters/players. a means of nerfing them in a way so that "Just adding more bots" couldn't be the answer.

[WarpZone]

TL;DR: STOP IT. If you nerf all the different kinds of playstyles you don't use, you're going to end up with a game that only has ONE playstyle. Nobody wants that!

I am so SICK of seeing the same old development cycle on Steam, time and time again:

Step 1: A developer makes an awesome amazing early access game.
Step 2: Players get good at the game.
Step 3: Developers change the laws of physics so the players aren't good at it anymore.

Do you have any idea how PAINFUL this is!? I learned a SKILL! Your game TAUGHT me something! I felt SMART! Then you decided that since I wasn't playing the game YOUR way, I must be doing something wrong. So you're gonna change the rules and move the goalposts so that now I'm stupid and I have to start over from scratch.

Do you have any idea what that does to a player's motivation!? Now it's like, why should I learn belts? Once I get good at belts, you'll just take that away from me too, right? You'll go "Oh, belts are too powerful! People aren't using trains enough! We'd better nerf them!"

This is not how you design a game! You don't give your players something awesome and fun and amazing and then rip it away from them! You don't give the players three choices on how to build their base, and then nerf them until there's only One True Playstyle! I don't care what your in-house testers think is the right way to play-- THEY DIDN'T PAY FOR THE GAME!!!

You're paying them to come in and do a job! Why should they have more fun than your paying customers!?

But the worst part is everyone who agrees with you? IS ALREADY VOLUNTARILY CHOOSING NOT TO USE BOTS!

You're fixing something that isn't broken and never was broken!

If I'm overreacting, (and believe me, I am,) I suppose it's because I held out hope that FOR ONCE it was going to be different. Maybe this time, just this once, an indie developer on Steam would just make a good game and it would stay good. But then you added biter base expansion to prevent players from carefully engineering around the pollution and feeling smart because they beat the system. And then you changed the science recipes so players would have to build a bunch of random extra items in order to reach Rockets. (WHY!? Why not just add NEW types of science to the endgame, why did you have to go back and change the OLD stuff!?) Now you're gonna get rid of bots.

And everyone's patting you on the back and telling you what a great job you're doing because you took care to frame the argument in terms of "bots being OP" and players being "worse at the game than the developers" if they use bots. So now everyone's saying "Oh, I never use bots! I never use bots! I'm just like the devs, guys!" because they want to look cool and look like they're good at video games.

You're creating an echo chamber so you can get away with removing any feature you want. Frankly, it's pretty disingenuous. But you know what? THE METRICS DON'T LIE! You have the numbers! You KNOW people do use bots! Well I've got news for you! People use bots because they WANT TO. If they didn't WANT TO, they would stop DOING it. Just like YOU DID.

People should be allowed to play the game however they want. Just because you CAN remove a feature or nerf a playstyle doesn't mean you SHOULD. Maybe everybody doesn't want sprawl. Or spaghetti. Maybe some people like building a bot network and sitting back and watching with satisfaction while it runs itself. Maybe someday that will get old! But that will happen for every individual player at a different time and date.

All players will not simultaneously get bored with bots at the same time when a certain update comes out.

Therefore you shouldn't remove all the bots from the game forever in a certain update.

Please. I'm begging you. It hurts. Don't do this to me again. I had hope and now you're killing that hope. Please. Just stop making your game worse.

This is Subnautica all over again...

[vampiricdust]

Blue belts move 40 items per second. Logistics bot with a capacity of 5... that means to compete, you would need 8 logi bots per second carrying those items for however long it takes to reach their destination. I have already done the math that 1 logi bot equals a little more than 3 blue belts. So for each 8 logi bots flying, you could have 27.8 blue belts. Then, because bots need to charge AND fly back to the starting point, you need more than double the amount of logistics bots for any given distance to maintain that.

Logistic bots are INCREDIBLY EXPENSIVE. There is a reason speedrunners don't use logistic bots, they're incredibly expensive. Combine this with the expense of roboports and the expense of expanding power production without risking blackouts due to these guys eating 1mw each, bots are not "overpowered" and they are not "cheaty". They are a huge investment. For every 8 logi bots, you could build a 1-5-1 train or 138 rails. You could build 478 yellow belts for each 8 logi bots and move 6,373 items per second twice as far as the bots can without any power consumption.

The only advantage bots have is operating in tight spaces where bulking out on belts is just not an option. Belts vs Bots is bulk vs space. Using bots for anything you have the space to use belts for is incredibly wasteful and incredibly inefficient. Yeah, bots can empty a chest in seconds, but the cost of the bots to do that is crazy. 48 stacks of 100 is 4,800 items, it would take 960 robots carrying 5 items to do that at a resource cost of 172,320 per chest you empty at the same time. That is 515 blue science packs worth of logi bots.

Belts are dirt cheap compared to bots and anyone who says blue belts are expensive is wrong, because they are dirt cheap for their item per sec efficiency.

[Caine]

I felt SMART! ... now I'm stupid

Ehhh... okay... I am sure that there is a joke in there

[Mestiff]

I don't understand the point "we think about removing something from the game because something else in the game is more fun for us", and you are right to keep logistic bots in the game.
I play with other people, some don't use bot at all because they don't like it and play only with belts, some use more bot than anything because they love it... but never we talked about one being too much powerful than the other because it's two different transportation system which have pros and cons and we play Factorio the way we like.
It's like thinking removing the train system because it's faster and can transport more items than belts.
The actual energy consumption, cost, science position/cost of bots and the roboport logistic zone size are now well balanced for the bot utility and doesn't need another change.
More nerf on bots will break the bot system, forcing player to use only one short distance transportation method (belts) and will make Factorio more linear (which is bad) and lacking diversity.

More pros:
- Bots accelerate the building pace the moment we really need to expend faster. Nerfing them will make building specialized outposts more tedious and repetitive, aka boring.
- Using a large army of bots tends player to use nuclear power to entertain it. Otherwise nuclear power isn't really needed and don't worth the time and effort.

People from universe where Factorio never had bots would be happy to have more diversity in game, especially when it doesn't remove anything else in the game, people who don't like the use of bots would be indeffirent because this addition wouldn't change the way they play and enjoy the game, they'll just ignore the bot technology like everybody who don't like bots do.

[hectar]

For the reason that you have stated as bots being OP, one of the reasons is that you are able to make hundreds of them easily. You could put a hard limit on the number of logistics robots per network. This way it will prevent them from relying completely on bots for production logistics, at most for simple, not always needed things.

[VrycePB]

I've only been playing for less than month, but I'm really enjoying this game. I like the idea that you're asking the community for their opinion on such a game changing topic.

In the FFF you mentioned 2 universes. One with Logistic Bots and one without. Why not have this as a togglable setting that can be changed when creating the world in the first place. This way you could turn them off if you wanted to play a world more focused on belt usage or turn them on if you wanted to play with the little flying sherpas.

I'm sorry if this is already been suggested, I just didn't want to sift through 26 pages to see if it had been.