Friday Facts #224 - Bots versus belts

[Engimage]

First of all thanks for stating that you will NOT remove bots For those who did miss the statement:

Don't worry, logistics bots won't be removed from the game, mostly because it's a feature that has been developed and polished quite a lot. Also many players love the feature, and we've all become used to it over the years.

Many people (myself included) think that logistic bots are overpowered. Personally I try to avoid using them at all costs on every my playthrough.
However even being kinda biased fan of belts I am glad bots are here.
Yes bots are cheaty but sometimes even best engeneers want to have a shortcut.
I am glad that bots were moved high in tech tree so you need to deal with belts for quite a long time.

However bots are totally irreplaceble on megabase scale. I will point out several things.

On megabase scale outpost creation is really boring repetetive thing so there is almost NO option other from bot based mining. It is only a question of deployment time. Laying belts/balancers is fine for early/mid game but it is totally unacceptible for megabases.

Bots have their simplicity. But they also do have better throughput and are more UPS friendly. What can be done here? You can't change simplicity and you will not make them more UPS hungry so we are looking at throughput here. You can either lower throughput of bots or buff throughput of belts. Or both.

Currently bots have nearly unlimited throughput so that you can just mitigate everything by throwing more bots in. However for continuous throughput (like super dense beaconed factories) throughput IS limited by the amount of charge pads available so at some point if you lower bot carry capacity or increase charge time they will start to suffer forcing a player to spend more space placing more roboports which is totally fine by me.

Another thing here is belt capacity and/or difficulty of usage. Surely you should solve all current belt problems making them harder to use. Compression, sideloading, splitter inconsistencies etc.
But you mentioned overcompression. Maybe it is really good idea to think this way. I remember many suggestions about increasing belt throughput by introducing stuff like packaging (stacking items into a single package taking 1 slot on belt) and loaders. One idea is having something like Klonan's belt buffer which could accept such packages via inserter and unload them to belt and vice versa. Simple and cool. Or just accept items on one side and offload packages on another.

Maybe it would be a nice idea to unlock something like that really late game to counter bots? Maybe some supercompressed belt which can only be handled by loaders?

One more thing to look at is beacons. I can clearly state that namely beacons and the way they effect production is forcing players to go super compact and make bots overwhelm belts. While I do like to design belt based beaconed setups myself I can clearly see why bots dominate belts.

[FlyHigh]

Glad you're finally reaching the same conclusion I've reached a couple years ago: 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.
Cliffs have been a great addition to those of us who like the train&belt approach though. And the fact that this controversial subject makes an appearance in the FFs suggests the issue about OPed Bots will be addressed sooner or later. I can't wait for the imaginative ways you're going to work the problem

[Uxi]

I think the people who want get rid of bots haven't tried building a "proper" logistics based (mega) factory.
I my latest game, my "base" consists of over 40 separate logistic bots based production areas all connected by rail,
with 300+ trains ready to roam the tracks and that is without counting outpost or defensive areas.

The base narrowly escaped death when the fluid wagon was hit hard with the nerf hammer...
Removing logistics bots would make it stop working instantly... Maybe I could then redo everything to belts...
But I'm sure trains would then be considered OP and removed from the game, or?

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

Please don't nerf bots (logistic or otherwise)...

[vampiricdust]

Glad you're finally reaching the same conclusion I've reached a couple years ago: 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.
Cliffs have been a great addition to those of us who like the train&belt approach though. And the fact that this controversial subject makes an appearance in the FFs suggests the issue about OPed Bots will be addressed sooner or later. I can't wait for the imaginative ways you're going to work the problem

Why do you care what other people do in their factories? Me using bots doesn't hurt you, but you hoping they'll change bots will hurt me. This doesn't seem fair at all. Why not make belts better to help you and me and leave my bots alone?

[ili]

I'm sure this was mentioned before - the same way you have barreling, why don't you have packaging? Put a stack of items into a "box" and then send boxes on a belt? In case of metal plates, that would increase the belt throughput 100 times.

Like this mod, but limited to belts?
https://mods.factorio.com/mods/EmperorZ ... ompression
Can be a good way to make belt super effective...

[dinodod]

I think logistic bots should occur earlier and not be nerfed, but instead the main map generation should be moved closer to the rail world settings so as to make bots alone infeasible. On the current map settings I do admit they're a little too powerful, but I don't think that's an issue with the bots themselves.

I also think that belts should be improved. The problem I have with entirely relying on belts and avoiding bots is that subtle management of belts can be extremely frustrating. Take for instance the challenge of a 4x4 balancer. Yes it's been solved without reducing throughput, but if someone doesn't look up what other people have done, this (and many other problems) seem daunting Furthermore, if I'm not mistaken, a fully compressed belt will not always have a constant number of items on it per tile. This makes circuit network conditions really frustrating. There are some other features that would make logistic challenges involving belts much more satisfying, like a priority splitter which preferentially splits to one side.

I think that if anything, belts should be improved to be as satisfying to keep up with the ease of implementing a well-designed bot network.

"but if someone doesn't look up what other people have done, this (and many other problems) seem daunting"

You completely missed the point of the game. 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. You only want a simple "click and solved" solution to building. You don't want to think for yourself and solve problems but rather expect others to come up with the solutions and then use them for your own builds without understanding how to use them.

[Kyralessa]

Factorio should incentivize belts instead of punishing players by removing logistics.

How about, in addition to adding the "green belt" (faster than blue belt) I mentioned previously, making it another infinite research item, where each tier of science (including rocket science) makes the green belt just a bit faster?

[Faen]

I personally don't like logistic bots moving items between chests, it feels too much like cheating, I normally only use requester chest to suply fuel for my trains.

That said I don't want to see bots nerfed, I would like to see belts buffed in the lategame, at that point its gets frustating to manage belts throughput, every time you need to make items in bulk you will waste a lot of time solving belt's bottlenecks, most times arriving to bizarre or not very satisfing solutions, also a lot of times your inserters are unable to put items onto non-filled belts.

So I would like to see a new lategame belt (green belt?) that solves these two problems.

TLDR: New lategame belt with Insane throughput and able to accept items from inserters as long as its not full.

[cpy]

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

[Mash]

My randomness;

-Research reduced roboport range?
-roboportstation as roboports? robos. only transporting from/to modular(robo capacity)train-stations and players,
And robo. docking belts with a single filter, with holding capacity of one bot that requires dock/undock time, so robo. major item throughput requires space to equal the train.
-robo. radio activity attacks biters? At some point 3x3 tile worms emerge out of the ground randomly very close to the signal core egde base, radar included.

[DerGraue]

I love this FFF simply because it is controversial and also because some ppl can't read and start crying for no reason

@Twinsen

You asked in what universe I'd rather like to live:
My answer is: universe 1, the one with logistic bots as they are now.

- Bots give me more options to build my base. I do not have to just rely on belts, I also have the option to go to bots. And with the recent buffs to belts (UPS wise) they are way more viable now, so bots are not the only lategame option. In the game that I am planing right now I have parts of my base planned with belts, which I will use early and mid game and later on I plan to switch to robots. Just to have this option is a nice thing.

- Bots also have disadvantages compared to belts. They need research and a lot of energy, which you dont have early game. So it's a progression you can work towards.

- A bot base is not a no brainer, I have seen many bot builds that were just horrible.

- I think bot bases can be really fun to watch, it's just a preference, I guess.


@ devs:
You should ask those controversial questions more often in your FFF or just in general to see what the reactions are. IE the fluid wagon change seemed to piss of many ppl, if you had asked before that maybe you had chosen differently. I personally don't care because I never used fluid wagon separation, but many other people seem to care. Moar communication = moar better!

[wutname1]

I would love an option to enable the bots you describe.

Bot based bases always feel so dull.

[raidho36]

Me gripe with bots is that they remove the challenge. It's true that it's easier and quicker to design and build things that way, but that very aspect is makes it uninteresting. Building a factory IS the fun part of the game, and bots remove it neatly. Bots feel a lot like 100% refined sugar cereal: there's nothing wrong with it, it doesn't tastes bad, and it does everything it's supposed to do. The only downside is that you can't really enjoy it past age of seven.

So don't use them. I don't find belts to be a challenge, they are just tedious, time consuming, and you still have to build a factory. Your argument is as childish as you tried to make bots sound.

So bypassing playing the game is supposed to be a better, more mature way to play the game? Why not have player start with rocket launch kit so the game can be finished in under a minute? Why not have all recipes unlocked from the start and cost nothing to produce the items? Same reason applies to using belts over bots. Imagine if in Skyrim you had a key that opens all locks in the game? That would be boring and lock picking aspect would vanish.

Also. Imagine someone used that argument against you. Don't like guns? Don't use them. Don't like TV programs? Don't watch them. Don't like poor road safety? Don't ride cars. That argument is dumb.

[Sworn]

Add an option to the start of the game (no bots allowed), release and collect info during maybe 2 or 3 month. Than you will have some data to work with.

For me, that is true the bot's are simple to powerful, if think of bots like "Drones", than you can't have 100k bots loading itens at the same time and flying all over the place.
You will need a station to at least load itens, not just a random chest any where in the map, and probably but not required, an unload place.

But if you go for a more complex bot system (load, unload station) you will end up with a flying train system that carry less itens.

My opinion about nerf/buff. Don't do any. Buffs is a nerf, you just need to swap your point of view, and vice-versa, if you buff one sine the other was nerfed, and if you nerf one side the other was buffed.

=> My suggestion are the following:

1º) Make the logistic bot system more challenger, like you need to have some kind of station for the "drones" to arrive (and the stations take some extra space to), they can't came all at the same time and pick up itens at will. Like the station should have a number of slot available for bots to land, load/charge (you can't just charge flying over random places) (like the charging time, just some bots can charge today, the same should be applied to load itens), and have time to load itens. Than you may have more technologies that increase the number of slots/loading time or have more tiers of the "station" that has more slot and load faster.
With it, you will need to route itens to your stations than have the bots delivery it for the destinations.
Bot's shouldn't be allowed to switch stations at will. You have an station, the bot belong to that station only.
And you should be able to control the activity of the station with some conditions (enable/disable/set something, etc, etc) like the inserters and roboports have.

2º) About the belts, not only the belts but the inserters to, if you are going for "belt base" system, than the inserters options today are very bad. For a belt base we need better inserters like some mods add, not extreme overpowered like super mega fast, but at least fast long handed, stacked long handed, and filtered long handed. For belt base you need more then 1 or 2 line of belts.

3º) Like bot's have research, belts should have to, like the speed could be researched, "increase speed of the belt by X%" like 3 or 4 level of research for each level of the belt (like 10% each, so the yellow don't become faster than a red, etc, etc), Increase underground distance by 1.

4º) larger chest for the inserters to pick up itens. Like a 1x2. 1x3 chest

Obs: About the constructions bots: Don't remove constructions bots, at least for me, they are fine, actually you should be able to get them a little bit early, or start with one or 2 bots and a place for than to "rest" they only pickup itens in the "rest" place. But for now they are fine in the way they are implemented.

[jodokus31]

I think, the main problem of logi bots is, that they are the only way to go, if you want it really big. Belts can't compete with them. Not throughput-wise and not UPS-wise.
Trains vs. belts/bots is no question, because, trains are capable to do large distances the best.
What if, belts would be potenially the best way to go, if you want to have very high throughput for short distances in end game (How it can be realized is another question). So you have to choose: if you want the easier setup of bot supplied factory with less thoughput, or the best throughput with belts, but the hassle of set it up correctly.

I think, nerfing logi bots too much, is not an option, rather have end-game alternatives with belts.

Proposals for buffing belts:
- Add loaders very late in the tech-tree
- Add near inserters.
- Research for belt speed (if possible)
- Make chests bigger in size on the map (not in storage), to be able to connect more loaders/inserters to one chest.

[bobucles]

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.

[Caine]

Moar communication = moar better!

Few developers communicate as well as the factorio team. I do not think it is fair to claim they do not communicate enough.

If you disagree then I dare you to purchase Exanima

[Hurley]

This is the first time when I feel that I need to express my feelings and my opinion.
For me personally removeing or limiting / reducing bots capabilities will make the game less fun or even unplayable.
Making them harder to build, maintain (for example by changing for electic powered to fuel powered) or limiting items carry count will mess them up.
I bought the game about a two years ago. I've played it for almost 450 hours so far. For the first few months I didn't used bots at all but eventually when I started to use them I've liked them so much that I can't imagine this game without them.
If you don't like to bots or you think they are to powerfull simply don't use them or limit your usage (not through game limitation but your own choices) to point when It's acceptable for you.

So far I've perceived all of the changes that were introduced into the game as expected features, needed optimalizations, great enhancements.
I felt that all of the changes made the game more fun.

But if you introducing something into the game allowing players to use it, love it and then you are taking it away from them you make them angry as me now, by even starting considerations about removing the bots or limiting them.
It would be the same as you would remove magic from MMORPG fantasy game when half of the players are mages.

Buff the belts and leave the bots in more less same state as they are now. You can always add research to allow each belt type to store/carry more items. Simillar to bots tech.

[roothorick]

What I'd suggest:

• Bots should cost much more. Bot frames should be around 2-3x the materials, and make logistic bots take 2x PU instead of ACs.
• Chests themselves should be a bottleneck. It takes some fraction of a second for a bot to extract or place an item in a chest, and only one bot can be interacting with a chest at a time. Bots will prefer chests with fewer bots waiting/en route over closer chests. Requester and buffer chests with the fewest active delivery jobs will be dispatched to first. This allows you to compensate simply by adding more chests. (Also reduces their UPS advantage slightly )

This keeps them more than viable for megabases while complicating megabase design, and if you really need an army of logistics bots, you'll have to work for it. When in doubt, make people go bigger Carrying capacity buffs become much more than simply reducing the number of bots you need, as it dramatically improves per-chest throughput.

I also agree with buffing belts. Here's an idea: belt speed buff researches. Each tier adds a 25% buff. Caps out at 200%, making blue belts push 120/s (1 per lane per tick, so you hopefully don't have to refactor how belts interact with each other). One research scales all three belt types proportionally, with red and yellow belts capping out at 80/s and 40/s respectively. They will of course impede movement even more, to encourage using construction bots and player logistic slots, and dedicating space to walkways/roads.

Give inserters speed buff researches as well, so they can keep up with the faster belts. Again, scales all inserters proportionally.

PS: Construction bots shouldn't be a shortcut around mining time for trees and rocks. They should sit on top of the tree/rock and grind for a bit, similar to how repairing works.

[sicklag]

.