Friday Facts #224 - Bots versus belts

[Aidiakapi]

As many have stated, I don't really see the point in nerfing logistic robots, because for more throughput, you build more bots, for more power, you build more solar.

I think there are two keys design points that should be kept in mind:

• Allow people to play according to their playstyle. Factorio has always been a game you can play in many ways, and having an option (such as logistics bots), doesn't force anyone to use them.
• Instead of just tweaking numbers to make one playstyle more tedious (because that's really all increasing costs will do anyways), provide methods of making the other playstyles more fun.

The primary issues with scaling up belts is:

• Tedious upgrades, transitioning between tiers of belts is extremely tedious, with basically no automated way of doing so in vanilla.
• The limited range on beacons make high-throughput builds with belts rather annoying, there are several ways this could be changed, such as a non-square transmission area (like 8x3, though graphically that doesn't make sense), capping effect bonuses, etc.
• Mediocre throughput, even the highest tier of belt just doesn't move many items/second, and you need many, many, many lanes to get to a serious throughput, which again amplifies the problem of tedious upgrades, and usability with module+beacon builds. Although an extra tier (maybe 30 items/sec/lane) would be nice, I think the idea of stacked belts have much more potential. Having two layers of belt (or effectively 4 lanes), and maybe even something like a vertical splitter would imo be amazing for making creative designs. That'd literally add an extra dimension.

[Mimp]

I therefore suggest the real issue here is free infinite power through solar, that allows one-time cost to mitigate a whole aspect of Factorio gameplay.

So all logistic bots need to be converted to either burner logistic bots or fuel cell logistics bots? Sounds good to me.

Edit: Roboports for logistic bots would need to be able to have a fuel buffer and refuel logistic bots instead of recharging them...

[GenBOOM]

belts are 1 dimensional. they are either lines or dotted lines on the plane that you play the game on; being normal belts or underground belts.

if you want to make belts more interesting make belts that can intersect on that plane and move upwards into the 3rd dimension of the game.
that would make them a constant rubix cube logistical problem to solve of how to manage those intersections and merges.
I want to build a layered base of belts. please make this possible.

nerfing bots, solar power, laser turrets in no way improves the game and how you interact and solve these interesting puzzles.

there is no solution like creating a new mechanic that provides content.
don't go down the road that other devs go down by alienating your player base by re-aranging existing content instead of creating something new.

[Tinyboss]

I think there's a lot of assumptions about bots and belts being made in this FFF. I can appreciate that some players don't enjoy bots, but removing them from the game is a jump that doesn't make sense.

I also believe that building belts is way more fun due to it's inherent complexities, challenges, and emergent situations.

I have trouble reading these arguments and not jumping to a strawman such as "I believe digging up resources by hand is way more fun due to it's inherent time investment, challenges, and emergent situations."

This.

Bah. Time investment, yes, but there's absolutely no complexity, challenge, or emergent gameplay involved with hand-mining. Strawman, indeed.

[ekimekim]

As my megabase has spread across the map, it's evolved into a well-struck balance between the three primary means of transport (bots, belts, trains):
* Bots for random small-throughput stuff the player might need or construct, or misc non-bulk items like satellites.
* Belts for local high-throughput transport, eg. from miners to train stations and from train stations to mass production lines
* Trains for cross-country bulk transport
I don't think bots are too OP right now. They're excellent over short distances, but hit issues with large, spread out networks.

As I see it, the tradeoff between bots and belts is ease/flexibility over throughput/predictability. In my grand base designs, I use belts exclusively and even pride myself on being bot-less - as you say it's more fun that way! But I don't feel pressured to use bots.

I wouldn't object to a nerf to bot throughput somehow, particularly over short distances, though as you say it's very hard to do properly because they're so easy to scale. Reducing speed or reducing carry limit wouldn't do much when players can just build twice as many. An anti-swarming measure would definitely work best but that's probably a) too drastic and b) too hard to code.
A few small tweaks might work to give massive, over-provisioned small networks higher cost. What about imposing an electricity cost for idle bots? Sure, electricity is cheap, but high power usage is the main disadvantage of beacons too, and they feel at LEAST as cheaty as bots right now.
But ultimately I'd suggest the best plan is just leaving them as is.

A better plan might to be buff belts instead of nerfing bots. I'd love to see more options for higher throughput belts. This doesn't necessarily mean faster like the current yellow->red->blue progression.
Some ideas off the top of my head:
* Roboports currently scale logarithmically with infinite science due to bot speed research (fun fact, on my main map right now, the next bot speed upgrade costs 8.192 MILLION science!). It'd be nice if belts had an equivalent upgrade.
Maybe this could be a form of stacking, but one easy and potentially useful one would be research to extend underground belt distance. This would be a good candidate for the exponential research costs, since that way it won't ever realistically go above level 10-15.
* The main thing that slows me down with belting is good balancing and routing. An end-game item to make that easier/more compact would be nice - a 4-way splitter or something?
* An interesting and thematic consideration with an extra belt tier would be making it explicitly NOT able to be inserted onto / taken from. So you have to think in terms of capillary belts leading to arteries if you want to make use of them.

[purdueme91]

Allow people to play according to their playstyle. Factorio has always been a game you can play in many ways, and having an option (such as logistics bots), doesn't force anyone to use them.

This is the most important point. Just because someone thinks belts are more fun or you're lazy if you use bots, let me make that decision on how I play.

[Cleany]

The premise of "Bots vs Belts" is flawed in my opinion.

Bots can replace the function of belts in terms of moving items between assembly machines, but most people only do this to experiment. Not all, most.

What bots are used for mainly can be described as late game stock and base management. Making sure that there are certain amounts of certain items in certain places. And for construction bots, repairing and placing buildings.

What is being talked about being removed (hypothetically of course as we all understand) isn't bots, it is the logistics system, the late game systems of managing the items accrued during the many hours of gameplay, for the purposes of managing a large base/factory.

Personally I find it rewarding that there is a late(ish) game goal of being able to automate how, and where things are stored. Importantly, this is entirely separate from the challenge of creating production systems using assemblers and belts.

To illustrate this, my large factory consists of 2 things. A "base", walled and protected areas within which I can have dedicated areas of production. And production areas, mini factories producing things like science, oil products, power, military stock (bullets, robots etc.), bots, circuits and modules, you get the idea. Both of these things work in tandem.

What seems to be the suggestion here is that it would be more "fun" (I thought the quotes appropriate though I didn't like using them) to try to achieve late game large base stock and systemic organisation using only belts, and I can say that I strongly disagree.

What is fun (for me anyway) is looking forward to being able to have bots replenish items used when hand crafting, remove items that you don't want (e.g. wood, plastic) automatically and take them to where they are needed. Being able to supply several items from one set of stock, e.g. a central store of level 1 speed modules. Being able to manage the automation of levels of spare rocket fuel so that you can use it to power backup boilers. These and countless other things are just a part of the game, and to suggest that it would be fun to replace these functions with belts, or to lose them altogether seems wrong to me.

The problem as I see it is this: There are different types of fun, and the fun behind being able to automate a large base which serves to defend, contain, and organise the factories within it is being radically overlooked. I hope that we can all appreciate the challenges and fun that all aspects of the game bring to the table, the mathematical challenges of factories with belts, the tinkering fun of train signals, the satisfaction of using circuits to automate things, the gratification of completing a organisational structure to the layout of your base, and the sheer fun of being able to bring all of those parts together with the logistics system!

So let's be clear what is being discussed here - the nerfing or removal of the logistics system. (Bots only serve it, think of them as staff if you want!). For me, the logistics system, and the bots that tie it together, complete the late game - turning it from a set of similar challenges (factories creating various things), into a cohesive whole, letting you make a "base".

For me, without the logistics bots the logistics system doesn't really exist. Without the logistics system, it would be extremely tedious and ugly to move things around using belts and trains. There would be no satisfying end to being able to create a finished item, a thing of beauty that can exist on it's own, and defend your factories from biters and lethargy. Full automation and defence is surely the goal of your little chap when he landed, after all he is only one man. His only real defence from tedious and eternal micromanagement of a world of biters, cliffs and annoying trees are logistics bots!

[Caine]

if you want to make belts more interesting make belts that can intersect on that plane and move upwards into the 3rd dimension of the game.

Do you mean just adding elevated belts or are you talking full-scale dwarf-fortress style Z-axis here?

[Tardan]

Never ever nerve logistic bots!

i play factorio many years and over 1000 hours. i build in many different styles. with mods and without mods.

first i learn to build with belts. later i mixed belts with bots. than i tried to build mega factories:
- mega factories with trains/belts (you can also copy/paste this factory like a bot factory. you must only build some trains manually)
- mega factories with trains/bots

i like both mega factories. i love these thousands of flying bots at a sorting train station. i like wired spaghetti belt construction.
i also liked to learn the different ways of playstyle. why you think about to take away our fun with the game! everyone can play without logistic bots.


one thing is correct: with logistic bots the game is easier. with blueprints the game is easier. with nuclear the game is easier. and so on. you want to nerve everything and kill different ways of automation? really?

but one idea you can discuss: disable logistic robots and beacons until the first rocket. in the second game enable beacons. when in another game the player start a rocket, than enable logistic robots. give new player a sense about the rocket and motivate to start a new game.

[Eelshock]

What if bots were able to transport some items but not all?

So you could update your lategame factory to move all your ores around with bots, but you'd still need belts for intermediates and science packs.

[Squeegy]

I have a BOTS vs BELTS idea that I think allows logistic bots to stay, while encouraging belts.

It seems where bots should be discouraged is in automated production. So if all ingredients for science pack production are being delivered by bot, that should be a bad idea. But for more rare events like blueprint construction/deconstruction, player enters logistic network for resupply, or a rarely used train station needs a top off on it's fuel, it seems like bots could help without breaking the game.

The goal here is to create a factory that is a beautiful machine that turns ores into rockets and science, and the VAST MAJORITY of "gears" in that machine should not be bots.

CONSTRUCTION (FLYING?) ROBOTS

These are awesome. They make you feel like a god, and you still have to carefully design what you construct. And play based logistics just eliminates a ton of tedium in a large and complex base.

* They handle all jobs they do now, without any nerfing.
* Add ability to do player based logistics.
* Research speed and cargo capacity as they are now to make player resupply and construction faster and easier.

LOGISTIC ROBOTS

These handle all constructed entity <-> constructed entity item deliveries. While useful, the goal here is to make them a poor choice for speedy or high throughput delivery.

* They take some to pickup or dropoff items, and only one Logistic bot can pickup or dropoff at a time. (Maybe like 10 ticks? Improve with research?)
* They handle all jobs they do now, excluding player resupply.
* They move slower and/or carry less, or perhaps move MUCH slower 1/6 the speed, twice the capacity to make them feel "bigger" and justifying their slowness
* Make the recipe much more expensive.
* Maybe (I'm less sure of this suggestion...) give them a max range (improve with research?). They cannot take any items farther than their range. To get items over long distances you have to setup waypoints of daisy chained buffer chests.
* Maybe move speed is correlated to bot charge level (or max range), so that short distances are speedy, but gets slower and slower over longer distances discouraging high throughput long range deliveries.
* Maybe logistic bots collide with others, causing congestion if too many are around the same chests. (performance allowing, of course)

They would be slower and more expensive, which can be overcome by building more. But THE KEY is that they take time to load/unload and they BLOCK logistic access to that chest while it's happening. Building more logistic bots will not help throughput if the chests involved already have a logistic bot docked on them 100% of the time. You could, however, build more logistic chests to increase throughput but now you are paying for throughput with very precious footprint, which is a far more interesting cost. Perhaps logistic chests could be bigger (1x2 or 2x2) to make the footprint cost even higher.

Conclusion

These changes should make construction and player resupply awesome and convenient, as they are now, and logistic bots become useful for infrequent, low throughput automation. And if you wanted a bot based factory, you should have to dedicate significant real estate to it, you should want to put inputs and outputs at least somewhat close to each other, and it should be its own logistical challenge to get it working optimally. But the KING of high throughput automated delivery of items should ALWAYS be belts.

But some automated tasks seem ridiculous to belt up. Keeping trains stations full of fuel, they just need a rocket fuel delivery every once in a while. A requester chest supplying assembly machines making rarely used placeable buildings. The ONLY thing I use iron ore for in my main base is concrete, I don't to make a belt of that. These are things that are only worth automating if the automation is easy, but they don't affect the overall science/rocket throughput of the factory.

[Omarflyjoemacky]

The same players who are saying "nerf the bots" are the reason why we don't have loaders in the vanilla gameplay.

Players deserve a choice. They can choose to use bots or not as it stands. If you remove them, then ... no choice. That's not fun.

[Atox]

I really don't think bots should be removed since they are the only option for true megabases. I propose the following changes:

Personal Construction Robots earlier in game.
Logistics Robots require Blue circuits and/or Efficiency Module
Requester Chest much later in research.

[GenBOOM]

if you want to make belts more interesting make belts that can intersect on that plane and move upwards into the 3rd dimension of the game.

Do you mean just adding elevated belts or are you talking full-scale dwarf-fortress style Z-axis here?

as someone mentioned undergrounds belts are not really underground, and they for no reason have a distance limitation imposed upon them.

I think elevated belts that do not allow the player to walk on them and require some kind of support structure to be held up would be a much more interesting mid game, and bots can be pushed further back and it would seem only natural.

also, bots and belts are inferior to trains in every way in late game anyhow. while bots vs belt fight is going on the trains are doing all the work.
I agree beacons are too simplified as well. I would welcome an update.

[The Eriksonn]

I am one who like belts and trains. I still use bots for resupplying myself and blueprinting stuff. And also for small minority items that i dont feel like making a new belt for. like Alien artifacts when those exsisted.

I also like to make factories that use combinators to calculate how much is needed, and you cant really do that with bots, but that is off the Point...

If you want to reduce peoples intrest in bots, try instead to improve the intrest in belts by making them somehow more fun than they already are(impossible???)

Edit: if you look at the achievments we have 6 that are robot related but none at all that are belt related, so a small "fix" would be to include achievments that are belt related: for example if you have many splitters and undergrounds compared to belts is an area you get "spaghetti" achievment.

[Sonik-HSC]

Hello... Well merging 2 types of 1 bot kind be a good idea..maybe game can get more performance. With 1 kind of bot we never more need to think in wat do we need to build. But keep the 1º bots build then they will work on logistic...
About advance chest i think we only need 3 "provider chest"+"storage chest"+"player personal chest ( this chest only suply the player )" .

About prioraties the bot go first take itens on the storage chest to keep always room for possible storage! and the if there isnt the disiere item they will go to the provider chest

1More thing inst possible to merge the 2 rail signal in one??? This can give all huge advantage on the performance of the game!

Happy 2018

[Cleany]

CODING ANALOGY:

Imaging belts as coding a subroutine and logistics bots as the main program that calls the subroutines.

The methods used in the main program sometimes overlap the coding of a subroutine, and vice versa, but importantly the main program is mainly an organisational challenge, and coding a subroutine is mainly a mathematical/algorithmic/efficiency challenge.

Imaging removing the modular nature of programming? That is what removing the logistics system would do to Factorio. Instead of a series of beautiful factories all doing individual jobs, virtually independently of each other - there would basically be one huge factory, one huge spaghetti (with train "buses" lol). Would you ever consider programming in such an environment, where individual elements of the whole could not be satisfactorily separated from each other where necessary?

Why?! But why!!!

[entity279]

Logistic bots means boring factories made of fieleds of assemblers each with an output chest and an input chest .
(Let's also stop for a second and remember that logistic bots make the awesome player logistic/trash slots work. Such a useful and fun feature)

Saying "just don't use bots if you don't like them" is dismissive. What about multiplayer? Also when a solution is clearly superior a competitive player is feeling forced to use it regardless if she likes it or not.

Did not read all the posts, but saw some defensive posts attempting to save the bots from an upcoming Armaggedon. Since this is not coming, thouse posts are not contributing to the discussion, which is a bit sad.

Read posts saying that instead of nerfing bots, belts should be enhanced. I like that thought a lot. There should be a competition in endgame too, players feel better when they can choose competing, reasonable solutions to a problem. Currently there's not much competion as anything with belts takes much more time to set up. Belts should have (a) new feature(s), not supported by bots. Whish I knew what this feature could be.
(Very crazy thought: how about limited range belt portals, that teleport the items from the intput to the output . Underground belts on steroids. You said it yourself, robots basically already provide teleportation anyways. These portals will just be cheaper power wise, and interface naturally with belts)

As for the nerf, bots do feel inefficient in a way. In that whatever problem factorio throws at you, you just approach it with the same swiss knife. Maybe the game should not punnish the bots themselves, but the inefficiency. I can think of pollution + biters as that punnishment.
While solar pannels generate 0 polution, they still take up a lot of space. Similarly to how cutting trees helps pollution spread, all the space filled with solar pannels should encourage the pollution's spreading even more. It should thus matter (to the point where the difference is visible and tangible in some way), pollution wise, that the one factory layout has a significantly larger footprint than another.

So I'd like to see, in the very very late game, that players would choose between building an all out logistic robots factory with noticeably stronger biter attaks and building a safer factory with more belts than bots.

[IronCartographer]

Huh.

I'm surprised no one has mentioned this.

Want to make bots more exciting? Make them crash when they run out of power like they used to, but non-destructively (even if the item stays on the ground where they landed, potentially cluttering belts...although losing power over water might be fatal...)!

Then if you don't power your bots properly, you have to either scoop them up yourself, or hire yet more bots to do the clean-up and stuff the exhausted ones back into storage.

Thankfully filtered deconstruction planners exist now, for picking up such trash items on the ground.

[Mithos]

I also think that a big, monolithic logistic bot base probably is an easy and less challenging style of play. But instead of removing logistic bots entirely or just nerfing their skills, you might also consider to nerf the ability to build big, monolithic and boring logistic systems. If there was a limit for the total number of logistic bots in a network, the player would have to construct several smaller networks and manage their resource I/O. Or you could limit the number of roboports in a network and, therefore, charging spots.