Friday Facts #224 - Bots versus belts

[sicklag]

.

[pjodron]

Why do people always seem to be either all or nothing? i use both in my bases and i see both as important tools.
allright, lets try and save roboports by meeting your arguments

Pro-bot arguments:
"The fact that they are so powerful, gives a very big sense of progression. They are behind a late game research, and gives you things that are very powerful and game changing."
"It adds large diversity to the game."

yes, while i agree in general i dont agree that they are that game changing. i still use treads because of the need of having realibly get stuff from A to B. also treads can be used as a visual idicator that you need to increase production or that someitng is wrong (i never count production time)

Anti-bot arguments:
"Bot basesareusually less complex and less interesting to look at, manage and expand."
I have seen players that makes complex bases with bots. but what would you think was a intresting way to manage and expand a base on? move pice by pice by hand?

"Because of their ease of use (and apparent ease of use), players tend to go to bots. We believe that belts are more fun, but we are guiding the player towards a less fun style of play."
Personal view, and sorry that i am having fun with roboports and treads at the same time. but this is also so very political way of doing stuff. we think that you are wrong and thus we will punnish you for doing so.
why not give out a carrot instead of the whip? (i know why, whips "because of their ease of use (and apparent ease of use)" are more common to use)

"When you learn that bots exist and how they work, building bases with belts just seems tedious"
why not turn this around abit. look on the other side. WHY does belt seem so tedious?

on a separate note.
fluff steam and its 1000character limit, does not even tell you that there is a limit nor how much you have written

[Neandertal]

This will probably not be a very well liked idea but if I was in change I would do the following.

Bots wont try to fulfil a request unless the bot can make the entire trip on one charge (Roboport -> Chest -> Chest -> Roboport). So bots wont even try to fulfil requests that need to move items from one end of the base to the other.
I would make a tutorial to make this behaviour clear.
This would force you to use buffer chests as intermediate stop points for long distance requests.
This will also solve the problem of bots standing in cue at roboports waiting to charge with items in hand.

Not sure if this will be easy to implement though.

Thoughts?

[Mario]

Hello Factorio-Team,
I personally think it would not be a good idea to take the logistics robots out of the game.
The simplest and fastest solution for all would be to limit the maximum number of logistics robots. And through research to increase the number.
For example:
Started with 50 robots at a maximum in use. Research at level 2 = +50 robots in use Up to 200 pieces at once. When the 200 are reached. the white research module is needed (the one from the rocket).
As a result, the end game is not influenced too much and the late game is directed more directional conveyor belts. Even the adjustment is made so easy (for later improvements).
In my opinion Factorio is not about what man needs, but what is possible.
That's my first thought on the subject hope he keeps you going.

P.S. am german # googel translator (bad at writing)

[greaman]

You don't really have to manage them.
Oh and they are too cheap. But thats more easily fixed.

I beg to differ.

Apart from a severe increase in energy needs you have to manage (which has become pointsless with the introduction of nuclear power), you have to

- manage numbers (and you can do cool stuff with circuit based loading/unloading of bots into/out of roboports)
- manage loading stations, a.k.a. roboports (unless you use mods for pur charging stations)
- manage paths (even though they are not destroyed anymore like in the past, you can easily have thousands of slowly creeping bots somewhere in the countryside)
{...}

There are specific problems that are way easier to solve with bots than with belts, i.e.
- mass storage (try to store 10 million ore with a chest/inserter setup - thats neither fun nor beautiful),
- quick sorting (leave it to the bots to sort materials and ort rather than filter inserters)
- very compact builds (which I use for items not being mass produced, but automatized for example)

and still, I regularly build whole areas in 'megabases' without a single roboport, i.e. my refineries (where only trains transport stuff in and out)

[Twi]

Disclaimer: I've read mooost of the topic, but not all of it.

The way I see it, you can't remove the ability to brute-force things with bots without removing the things that make bots nice and good. And I'd say there's nothing wrong with letting someone who wants to focus on other things throw together a 'meh' bot setup to take care of something real quick.

However, bots being the optimal way to take care of a lot of problems? That, at least, you can do something about, even if it's not the fault of bots. It might be a beacon problem, it might be a recipe problem, it's definitely a (lack of) belts problem.

So yeah. BUFF BELTS.

BUFF BELTS.

HAVE I MENTIONED BUFF BELTS?

While I'm not really sure on how to solve some of their issues (getting sorting/many items in (more rainbow belts?)) and a lot of the ideas that might help have already been suggested (better sorting, less cludgy belt balancing), one thing I do have a suggestion for is throughput. (Beyond the obvious MOAR TIERS.)

Specifically, single-item-type throughput, because the phrase 'bot-based mining setup' will never not make me shudder, no matter how optimal it might be.

You see, the conveyor belts in Factorio are these dainty little things, right? They make sense for sensitive products on the factory floor. But for things you want to move a massive amount of, and you don't always care if there's a little mess? They don't really fit.

So have something I'd call a bulk conveyor. You build it like any belt, maybe even with undergrounds and splitters, but rather than a dinky little open belt, it looks more like a trench going into the earth, or at the very least it has some walls to stop all the junk from spilling out. It's kind of hopper-like in appearance, if hoppers were belts at least.

Mechanically, it's kind of like a cross between a belt and a pipe. It fits a single kind of item, represented as a value/quantity per tile/block/whatever rather than an individual entity on the belt. It moves said item quantity along automatically, like a belt. The actual velocity is slow, maybe even slower than yellow belts, but the amount that can fit into each space is immense. Not sure on the exact math, but I'd say it'd be at least a hundred, maybe hundreds of items per space.

If you really want to be sure you've dethroned bots for (single item) throughput, then give it an infinite research that lets you increase the maximum size. Call it 'bulk conveyor compression' or whatever.

(In terms of animation, well, I'm not a graphic artist, but what I'd probably do is make it 'like pipe animation, but slightly more complex'. You'd have your resting, open-belt animation for when there are zero items; one for when the number of items increases from zero (items entering the belt); one for when the number of items remains above zero (nice, active, full belt); and one for when the number of items drop back to zero (items exiting the belt). Again, think like big piles of ore coming rumbling down a huge conveyor, aesthetically. You could also have an option to 'cover' the belt with a hotkey, so that it gets a roof and people don't see the animation if they don't want to/are worried about lag.)

[someuid]

Wow. Tons of replies. Going to add to the chorus in the hopes someone will read and reply.

TLDR:
Don't remove logi bots
Buff belts via speed research to make them go faster
Tweak logi chests so they are more like assemble machines - limited item request types and add insert/remove delay for bots

=============================================

I don't think bots should be removed from the game. This is an open sandbox game. One of this game's primary features is easy modding. Features should not be removed to force some part of the player base to play a certain way because (1) this is an open sandbox game and those types are games are designed to let the player expand their game play in ways the designers and managers never though of and (2) its so easy for people to simply mod it back in that making such changes is really a waste of the programmer's time. The feature is already there so leave it be. I understand the need to tweak now and then but outright removing a feature seems like a terrible waste of time that could be spend doing other tasks.

That being said, the reason I go to bots is that belts 'annoy' me. You get three tiers of belts in vanilla and that's that. Building out a factory with tier 1 belts means manually replacing those tier 1 belts with tier 2 or tier 3 at a later date. That is boring and not automation (a core feature of this game.) You can not make belts faster in any way other than to replace them with a higher tier (1 -> 2 -> 3) and after tier 3 you have no other options to move more product into/out of a factory than add more belts and more belts and more belts, resulting in potentially moving around chunks of the factory to make more room for more belts. And this is why I go bots because after adding in the 8th belt of iron I don't want to add in more belts of iron. It seems like boring and tedious work.

So, possible solution: get rid of the 3 tiers of belts and give us some belt speed increase science to research. Let me drop down a belt system and a way to make it go faster and faster and faster without having to replace thousands of individual belts or letting me run into a glass ceiling that we have now, necessitating the need to tear up and move the factory again to add more throughput. I suggest this because this gives the player options (something an open sandbox game should be doing) rather than forcing them into one minded tedious game play.

Of course, part of the reason some people like belts is the challenge in snaking lines in and around assembly machines. Chests alleviate this issue by allowing as many different types of items to arrive at one point (the chest) that no belt system could do. Well, lets tweak this by modeling logistic chests after assembly machines. We have 3 tiers of assembly machines which are different from one another in their speed and number of different input items they can handle. So let's do something similar with logi chests. Make some different tiers of logi chests that are limited to the number of different input items they can hold. Tier 1 chests can only request/provide a single item (like copper plate) and it takes a logi bot 1 second to insert/remove the item from the chest. Tier 2 increases the number of item types to 2 with a .75 insert/removal time. Tier 3 can request/provide up to 3 items and have a .5 second insert/removal time.

Now, I know I've made two different type of suggestions here - one is to not take out a feature (keep the logis and make belts faster via research) and the other looks like i'm trampling logi chests. But I think there is a fine line between trampling a feature via removal (getting rid of logi bots) and tweaking something to be more like other items already in the game and people are happy with (making logi chests have limits like assembly machines.) And the suggestion doesn't end there either. Just as we can buff assemble machines with beacons, we could also buff logi chests with some sort of local device that you wire up to a logi chest to increase its stats. And that could be an excellent gateway to getting more people into using green/red wires.....

[Escadin]

I'm pro bot.
For one things it's because of that sense of progression. The fact that they let you scale so much higher is what gives them a place in Factorio. I don't like this transition because it isn't smooth but building up to high science pack tech and then building entirely new factory from scratch, with twice the size and 8 times the production of your old one is fun.


You say:

Because of their ease of use (and apparent ease of use), players tend to go to bots. We believe that belts are more fun, but we are guiding the player towards a less fun style of play.

Absolutely doesn't hold true for me personally. I do like the challenge of belts and balancing them is lots of fun.
However, I don't find anything "fun" or "engaging" in having to draw a belt line from one end of my factory, right across the main bus and anything else, back to the other side just because I happen to need one additional item there and ONLY there. It messes everything up for a triviality that is not actually worth redesigning your base over (I'd rather not automate that particular item then at all - or carry full inventories to a chest there manually). That's not an interesting challenge and I simply need to quick-fix it.



On the other hand, Bots are by no means omnipotent. Their AI is pretty basic and there are barely any tools interfere with it and guide it. In fact, sorting out logistic priorities and balancing distribution is actually harder with bots than with belts, simply because the lack of tools and mechanics. I assume you did this and stayed indifferent to any proposed changes because you didn't want to buff bots any further. Unfortunately, what you achieved at the same time is deriving them of any depth in their mechanics. And then you say belts are so much better because they offer more challenge and fun.

This point has been discussed several times in this forum and I'd like to remind you of one concept I've been trying to push across all those discussions:
Logistic Bots could have their own interesting and challenging late-game puzzles, their own depth and thus stand equal to belts in how engaging they are to the player - both for fun and for difficulty.
We just need to more entities to influence their AI, create balancers, priority lines, isolate transport routes and everything we can do with belts. After all, both belts and logistic bots share the same requirements, the same logistic demand of your factory. There is no reason they can't be equally engaging and challenging but different in their execution or with varying pros and cons.

Thus, I'd rather have you increase the control we have over bots and (at the same time) the challenge in using them correctly.


Late to the party... as always.

[seePyou]

This will probably not be a very well liked idea but if I was in change I would do the following.

Bots wont try to fulfil a request unless the bot can make the entire trip on one charge (Roboport -> Chest -> Chest -> Roboport). So bots wont even try to fulfil requests that need to move items from one end of the base to the other.
I would make a tutorial to make this behaviour clear.
This would force you to use buffer chests as intermediate stop points for long distance requests.
This will also solve the problem of bots standing in cue at roboports waiting to charge with items in hand.

Not sure if this will be easy to implement though.

Thoughts?

This! 100 times this! Also, a previous thought, that if belts seems tedious, FIX the belts! It's like saying "Vegetables are good for you, but chocolate is preferable, so we are going to ban chocolate".

[bigflyer]

I tend to steer clear of huge numbers of logistic bots. Like you, I think they seem a bit cheaty.

Usually I use them for 4 purposes:

- to supply me
- to supply a group of factories I have producing odds and ends (eg factories, electricity poles etc, so that the construction bots can build things easier)
- to send stuff that I have trashed from my inventory to suitable places (ie, mainly trees)
- for the odd case where I have messed up the belts and I can't easily get a belt from one factory to another which needs its output (this is something I try to avoid).
(The exception to the last case which doesn't bother me is supplying fuel to trains. It'd be a nightmare to route belts to all these.)

I tend to have no more than a couple of hundred bots by the time I'm bored with a world and start again.

BUT, I haven't yet had the stamina to build a huge mega factory producing hundreds of satellites or whatever using mostly belts. I think the best I got was doing a satellite every 5 minutes.

What I would probably do about it is:
- buff belts a lot
- Make it possible to get a much higher throughput on belts. You have upgrades to make bots faster & carry more. Do something similar for belts. Belts start getting annoying when I NEED to have lots of parallel belts carrying the same thing so I can get the throughput I want. They're fun in early game when one belt can supply most of my factory. Later, not so much. So, let blue belts be upgradeable
- Make it easier to sort items off belts in late game (eg intelligent splitters)
- Be able to easily split left/right belt contents onto two belts
- Better belt balancing
- Belt compressors
- maybe allow belt circuit connectors count 'items per second/minute' rather than just items on that segment of belt at the moment (would allow me to build more intelligent balancing)
- limit bot counts without upgrades. So, you could allow small numbers of bots in early game and make it so that it's only in late game that people can have thousands of bots

[seePyou]

get rid of the 3 tiers of belts and give us some belt speed increase science to research. Let me drop down a belt system and a way to make it go faster and faster and faster without having to replace thousands of individual belts or letting me run into a glass ceiling that we have now, necessitating the need to tear up and move the factory again to add more throughput. I suggest this because this gives the player options (something an open sandbox game should be doing) rather than forcing them into one minded tedious game play.

Of please let that sink in please dear whatever please!!! Replacing belts is my No1 gripe with the game right now and a vastly huge part of the reasons I go bots!

[meganothing]

So why not introduce some way to manage them. Maybe managing priorities and limiting them to routes.

While belts are managed on the physical level (where you place them and in which direction) logistic bots are more on a logical level (where to get stuff from and where to put it). And all that is (quite extraordinarily) managed by the game itself. No real way to optimize/develop strategys there. You set the amounts in the requester chests and thats it.

In my opinion, thats where you should make changes. Make bots managable (changable behaviors, prioritize some goods) and limit them more (let them break down if they run out of power with the need to be retrieved), so they become more fun in themselves.


TL;DR I like the bots, but i want them more responsive to user choices.

I really like that idea. For example:

1) it could become possible or neccessary to assign bots to specific chests or chest groups (similar to train network). Sure, that would first make them more powerful but also would mean there finally is an optimization game you can play with logistic bots.

2) or (similar to 1) you have to give them a start and end point and they can only access chests in a small distance around start and end point.

Both solutions would not bother megafactory-builders in the slightest (blueprints would set them up like it is done with assemblers), but add a game to them in earlier stages.

3) Even more interesting would be if they could be given a short state machine aka sequence of operation diagram where you could specify their behaviour. But that would be a major new game system, a lot more programming work for Wube.

[rldml]

get rid of the 3 tiers of belts and give us some belt speed increase science to research. Let me drop down a belt system and a way to make it go faster and faster and faster without having to replace thousands of individual belts or letting me run into a glass ceiling that we have now, necessitating the need to tear up and move the factory again to add more throughput. I suggest this because this gives the player options (something an open sandbox game should be doing) rather than forcing them into one minded tedious game play.

Of please let that sink in please dear whatever please!!! Replacing belts is my No1 gripe with the game right now and a vastly huge part of the reasons I go bots!

Basically a good idea. Make the first three logistic techs for upgrade like they are now (13/s, 26/s and 40/s), but represented in only one instance you can place, and after them: infinite science to make them even faster (e.g. +1/s per tech level) - with the premise the game logic can handle this.

But the problem i see are the costs of the belts: If they are just stackable upgrades of the same thing, they all would have the same costs to construct them. To balance this, the default belt have to become more expensive. That would be a problem in early game. Perhaps you can make the recipe more expensive over time, how do we prevent that smart players produce huge amounts of belts BEFORE they develop the next upgrade tech? This would become a viable strategy and is the exact opposite of what i would want as a game developer...

It would be much easier, just to remove yellow and red belts and start with blue belts from the beginning. That would reduce the unwanted replacing-times, but you lose a part of complexity of the game. And that is one of the key feature why we play this game...

Greetings, Ronny

[Avezo]

I don't think that faster belts are the solution. Not that it's a bad idea, but animations of inserters would look rather stupid at higher than blue belt speeds IMO.

If we get faster belts, loaders/unloaders are neccesity. Even to the point that perhaps faster belts would be too fast for inserters and would work ONLY with loaders. This could introduce additional challenge/annoyance because you'd have to plan faster and slower belts ahead, instead just using fastest belt everywhere.

When it's set like that, belt speed (together with loaders) could be subject to infinite research.

Hovewer I think it would be bad to introduce yet another type of belt, to me they feel a bit bloated already. I say - turn blue belts into this instead, they are already crazy expensive, would feel good if you could make them even more powerful through infinite research.

I said something similar somewhere on the forums lol

[meganothing]

TBH if someone comes and open a topic with the name "whatever" vs "whatever" he/should be warned cuz of possible community split.

On the other hand one could see what you are doing as blaming the victim.

It's like: Look the bad jew said something now there is the list everyone can read again...this is negative public exposure.GUY THIS IS SOMEHTING YOU SHOULD DO IN A PRIVATE ROOM. you are not a good mod. If you feel attacked please think twice about WHAT I MEAN. I don't wanted to attack you.

Now you really lost it, sorry. Godwin's law is confirmed again: "“As an online discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches one.”.

Warpzone was warned without directly addressing him in a post by Koub. He continued with another melodramatic post bringing on a direct warning. I know of another games-forum where his account would have been (very publicly) disabled a long time ago. Koub, I admire your restraint.

[GenBOOM]

I just want to be clear.
My post earlier was not about attacking the devs, I wanted to provide an analogy that could be easily understood by anyone and the perspective of how this is being handled from the bot players point of view while throwing in a star wars reference.

There have been a lot of constructive and creative posts about solutions other than changing logistic bots that would actually improve the game for everyone if the devs don't listen then its clear they do not care about all of their current players, its as simple as that.

[Koub]

The thing is : If half the players want to keep (or even buff) the logistic bots, and half of them want to get rid of (or violently debuff) them, which half should the devs upset (if they were really considering to do something, even radical, for good) ?

[brunzenstein]

The thing is : If half the players want to keep (or even buff) the logistic bots, and half of them want to get rid of (or violently debuff) them, which half should the devs upset (if they were really considering to do something, even radical, for good) ?

Thats a simile as trivial question:
The latter half should simply refrain from using bots.
The first one can

So everybody is happy

[Tricorius]

The thing is : If half the players want to keep (or even buff) the logistic bots, and half of them want to get rid of (or violently debuff) them, which half should the devs upset (if they were really considering to do something, even radical, for good) ?

I never use rail tankers. I rarely use half the stuff in the military tab.

How many posts have I made on the forums asking for all of those features to be removed? How many posts have I made asking for those things to be punitively nerfed into the ground?

(Spoiler: I’m pretty sure it was none.)

I simply don’t understand the personal world view of “if I can’t have it the way I want it, no one should have it”.

I don’t see this as an either/or situation. In fact I’ve never read topics suggesting there was a violent divide amongst players until this topic. I’m sure it is probably here in the forums...somewhere. But I haven’t seen it. Which likely means that 90%+ of the Factorio playerbase isn’t aware of the division either.

Honestly, I think the devs kinda opened their own can of worms here.

I would be *very* surprised if half of the playerbase hated bots. (I would believe maybe half of the forums community hates bots, but this gets us back to the microcosm of a microcosm issue I have with the forums.)

Does anyone *realistically* believe that 600,000 people actually hate robots in Factorio?

I’d be willing to bet that the large portion of that (theoretical) 600,000 playerbase simply wouldn’t use them if they didn’t like them. And they would instead play the way they wanted to play. (Read: sandbox game)



P.S. I’m willing to be proven wrong by actual data. I have been curious if the devs actually collect any metrics. When I have to report on the “popularity” of a given feature in the software I write, we collect metrics information.

How many games do people start?
How many of those games do they use belts in?
How many of those games do they use bots in?
When is the average time / tech level / whatever when bots are constructed?
What is the belt vs bot distribution over time?

This is a solved data science issue.

If these results come back with “90% of people stop playing as soon as they reach robots, never build another thing, and stop playing the game altogether”...then yup, I agree, the devs have a very real problem.

Somehow I doubt the metrics data would reveal the above scenario though.

[Jarin]

The thing is : If half the players want to keep (or even buff) the logistic bots, and half of them want to get rid of (or violently debuff) them, which half should the devs upset (if they were really considering to do something, even radical, for good) ?

Pretty sure that buffing belts and inserters won't upset very many people.