Friday Facts #225 - Bots versus belts (part 2)

[puffpoldi]

So i have a few ideas on how to try and balance those things a bit.

I'm currently not playing the game very much, but i've enjoyed it when i did.

You could add vertical splitter and make belts buildable under and above other belts (stacking the belts so to speak). Then you'd need to add a way to toggle viewing them and cycling through those 3 layers.
And then add either topshelf/underground inserter that insert from underground and topshelf belts. Making them either a different entity or making inserter stackable too.

Someone suggested adding carry weight. I think that's an interesting idea. But You'd have to define weight for everything that's in already. So i thought: why not use something that's already in there... like model size. Make Bots match arial transport via helicopter, in that everything that has a size of 1x1 or less when build can be transported by one bot. Everithing 2x2 by 2 bots. and everything bigger needs 4 bots. Or something like that. If you like you could make them transpot the actual model too. But i don't know if seeing a flying assembling machine would look awesome or was just annoying.

greatings

[Jürgen Erhard]

Hello there!

Tbh, I havn't read the entire forum atricle but I had a semi-long discussion with some friends of mine who worked in the gaming industry.

I agree with the factorio devs that bots in the current build are gamebreaking.

What freaking game are they breaking? I'm a fan of bots and I don't see what they're breaking.

It's "FREE PLAY" for a reason… or at least I thought that was the reason.

[live22morrow]

I agree with the factorio devs that bots in the current build are gamebreaking. The gap between belts and bots is way to massive and lead belts to absurdity.

I don't really follow? In the mid-late game, they're very strong, but need a lot of research and are very complex and expensive products to make, needing a lot of investment to get going. They do simplify layouts, but this is only a really big benefit if you're inexperienced at factory planning and didn't design your factory to scale.

At the megabase level, bots offer around 2x the space-throughput efficiency of belts, and allow for somewhat more compact beacon setups (maybe 20% - 50% more). This is in return for costing 5-10x more material to build, requiring more complex intermediates, and using a colossal amount of power. They're also more ups efficient, though the difference has been greatly reduced with recent improvements to belt performance.

[Mylon]

I thought a while about nerfing bots. But however you nerf their throughput it's so easy to compensate; "Throw more bots!"

But what if you make bots randomly drop stuff?

Even a 5% fail rate on a delivery would be a major wrench in perfectly timed setup. Such a nerf would make bot networks harder to predict and have an increased resource drain compared to belts. It would make bots a little more interesting than just another way to do the same thing as belts, but better.
Another twist could be to induce a harder penalty the longer distance the bot travels. Or a careful bot that's extremely limited but won't drop anything (for that precious satellite perhaps?)

https://mods.factorio.com/mod/Bot_Servicing

[vampiricdust]

Hello there!

Tbh, I havn't read the entire forum atricle but I had a semi-long discussion with some friends of mine who worked in the gaming industry.

I agree with the factorio devs that bots in the current build are gamebreaking.

What freaking game are they breaking? I'm a fan of bots and I don't see what they're breaking.

It's "FREE PLAY" for a reason… or at least I thought that was the reason.

I don't get it either, I thought logistic network was a straight up upgrade from belts. Far less cost efficient, but way more control over production and distribution. Instead of dealing with hundreds of splitters trying to get the right ratios, you can just set request amounts. Belts are really boring and their issues are obvious. Logistic bots come with a lot more subtle problems, huge power drains, and their flight paths have to be monitored to minimize wasted production times. Belts would move everything instantly freeing up assemblers to work, but with bots, they have to wait until the inventory is reduced.

There are so many subtle challenges to a logistic bot base that shifts things from a belt puzzle to a layout puzzle. Changing the positions of assemblers from long rows of the same thing to smaller modules of whole production chains, having to scale your power up rapidly, and placing roboports so bots don't way out of their way to charge. Belts are small scale thinking puzzles, logistic bots are large scale thinking puzzles.

[Koub]

The emptiest barrels have the loudest sound or how to obtain a Pareto efficient solution while the ghosts of past errors haunt you in your dreams?

I must admit, I'm a belter and everything that keeps those pesky bots in check is a good thing. I'm lazy and I use bots but I feel dirty. Burn them down. Cleanse the sky. Throw them off a cliff. Bury them together with those modules and their bastard beacon brothers. Bots were an error. Blueprints were an error. You must make the game harder, not easier. You must think smaller, not bigger. Make finite but solvable puzzles instead of infinite sandboxes. Make something count something. Let us make the factory an art in harmony with the terrain. Banish cliff explosives. Make it rough. Make it hard.

People should be progressively drawn away by the bot heresy and guided to the belt enlightenment by incremental punishment.
First use of logibot, he should have a warning message
If the player keeps on, he should receive small electric shocks, increasing with intensity if he doesn't get back on the straight and narrow.
Once the electric shocks are at maximum intensity, the punishment should continue by hitting or slashing the player with some weapon.
And if the player still doens't understand, burn them to ashes in cleansing fire to get rid of the heresy.

Or you could also ask yourself why is there so much noise with people asking to nerf bots in every imaginable way now that the devs have opened the pandora box, but almost no bot balancing issues were reported before (and I could find only one mod that did nerf the logibots that existed before this shitstorm, whereas there are literally dozens that BUFF bots in any possible way since months or years).

I wonder what would have happened if the devs had swapped bots and belts in their 224&225 FFF. Would have people overwhelmed us with ideas to nerf belts ?

NB : The first part is obviously a joke ^^. But the remaining of this post is serious. There have been more discussions in the past about nerfing belts than there have been to nerfing bots.

[vampiricdust]

Or you could also ask yourself why is there so much noise with people asking to nerf bots in every imaginable way now that the devs have opened the pandora box, but almost no bot balancing issues were reported before (and I could find only one mod that did nerf the logibots that existed before this shitstorm, whereas there are literally dozens that BUFF bots in any possible way since months or years).

I wonder what would have happened if the devs had swapped bots and belts in their 224&225 FFF. Would have people overwhelmed us with ideas to nerf belts ?

NB : The first part is obviously a joke ^^. But the remaining of this post is serious. There have been more discussions in the past about nerfing belts than there have been to nerfing bots.

Fully agree. I have always seen mods with more tiers of bots, infinite research for better speed, cargo, and power capacity, mods for faster charging, no charge bots, and better roboports with bigger ranges & charging spots. So many buff mods... I never used them as I thought vanilla had very balanced and effective research limits. They were easy to set up, but very difficult to troubleshoot & optimize.

The hardest part of these FFs is that everyone is jumping on the nerf bot bandwagon because there is a lot of love and respect for the devs. We just got buffer chests and then we get 2 FFs hinting at heavy bot nerfs... it is rather hard to take after getting something we haven't been able to do ourselves.

[Deadly-Bagel]

We just got buffer chests and then we get 2 FFs hinting at heavy bot nerfs...

We got one FFF hinting they wished they hadn't made bots so strong, quite different from nerfing, the second FFF was focused more on buffing belts than nerfing bots. They even went through examples on how they had considered doing this.

[Bauer]

Guys. It's only another few hours until we're seeing the next FFF.
So calm down, save energy, and wait until then before setting the Internet on fire.

[Aflixion]

We just got buffer chests and then we get 2 FFs hinting at heavy bot nerfs...

We got one FFF hinting they wished they hadn't made bots so strong, quite different from nerfing, the second FFF was focused more on buffing belts than nerfing bots. They even went through examples on how they had considered doing this.

The second FFF literally said in the conclusion that they think bots should be nerfed and listed a couple ideas they had for doing so.

[stm]

Concerning the comparision of throughput there are at least one, I would say two mistakes which were made.
Comparing a Belt bus of width 4 with a complete line of roboports is not really the whole picture. This is due two two effects:
1. You neglect the neccessary Energy production (which is true for all usecases)
2. You neglect the neccessaity for boxes. which could be replaced by an additional belt inside factory layouts (al least in most cases)
Thus the resulting efficiency factor is definately significantly smaller than two.
This comparison also does not consider other tricks like belt braiding (I don't like or use it, but it is available!)

These mistakes made in the metrology do not negate the argument that robots take away a huge challange concerning routing of items, though one could argue, that the danger of shutting down everything completely and not beeing able to control priority targets in case of shortages might compensate for that at least partially.

Concerning the proposed improvements of splitters: I would see them very critically, since they do take away a lot things which make laying belts interesting. Not only will they make filter inserters (and especially there stack counterpart) redundant, but most of these features can already be had by just adding one red or green wire to the belts In my opinion it would be better to include features like input and output ballancing, and perhaps the ability to connect splitters adjacent with each other to form larger balanced setups which do not require huge amounts of space. Just by adding additional Elements for line ballancing, belts would gain a lot in my opinion. Using the proposed improvements I fear that the easy parts of belt design get more easy (and boring) while the nigh impossible challanges (which are done a lot more efficient in many ways by bots) are not really touched. The item filtering feature of the new splitters for sorting beeing the only exception.
Stm

[olafthecat]

Post here instead of everywhere else please!

Nice try
[Koub] merged into main topic.

Sorry, I'm just sick of people talking about it literally everywhere.
I knew that this existed, it's just that others don't realise that it does.
I don't really mind what happens, as I have said multiple times, but I do like discussing it.
Up to a point.
Then it gets...frustrating if anything else.
Might abandon the forum until it calms down

[waisinet]

Why dont you just add an option to deactivate bots when starting a new map?

[The Eriksonn]

Might abandon the forum until it calms down

Or do like me: dont write much but constantly observe and read what others are saying, thats also fun in my opinion...

[StachelJoe]

I really like the idea that every way of transportation has its advantages and disadvantages and to build an optimal factory, you would have to use all of them.

Trains for the cheapest and biggest throughput on long distances,
belts with medium throughput but more flexibility than trains
and bots with less throughput than belts but more flexibility.

So for a max output factory, you would use trains for the ores/plates, belts for the items like green circuits or gears and bots for items that don't need much throughput, like blue circuits, but would complicate the belts too much and therefore use too much space. There would also be more different approaches in how to build big factories, depending on which items you chose to move by belt or by bot. And i think to limit the throughput of bots by limiting the amount of items per time that can be put into a chest could achieve that. And maybe a belt buff/ the discussed splitterbuff. I never build a factory so big/optimized that it became a problem for me, but certainly it does for other people.

And after reading my own post to make sure it does make sense, I'm sorry for my english

[uint]

I don't get it either, I thought logistic network was a straight up upgrade from belts.

I think what people mean by saying that bots are breaking the game, and why devs want to nerf them, is that bots weren't meant to be an upgrade from belts. They were meant to complement belts, just like trains complement belts in cases where using belts would be esoteric (or just make life easier and more pleasant). Of course the lategame where belts are redundant is also a valid game mode, as shown by all the people who use bots to build their great factories. It's just not something the devs envisioned, and I think they have their right to make the game they want to make.

Maybe devs can keep the "old" bots as a separate game mode, and rework them as they see fit for the main game.

[mrsimple]

Hi!
If you guys want to nerf the bots, why not increase their energy intake in a non-linear way based on the working drones in a roboport-net? Or limit the available commands you can issue/second in a roboport net? Combine it with decreasing the roboport net command/sec based on the distance of 2 different roboports networks. The closer they are, the less command they can issue. This would not allow the roboports to grow without limits.
Even if the details of my ideas are flawed, the concept would be to make the throughput function non-linear while not affecting the belt throughput. I keep thinking about WiFi thechnology where you can only have a limited amount of users before the speed will be slow or in mobile-phone services where you can only give service to a limited amount of devices.
Belts have a non-linear throughput (even though they have the big advantage of not needing electric power), because after a while you won't be able to put items on the belt, so increasing production won't increase the actual output. It has a natural boundary. Right now bots don't have this.
Any thoughts?

Edit:
You could also introduce minimum roboport distance since the bottleneck with bots is the recharge time. You can bottleneck their energy supply, so if you put in more bots, you don't increase the throughput. Combine this with the tweaking of bot recharge time.

[meganothing]

We got one FFF hinting they wished they hadn't made bots so strong, quite different from nerfing, the second FFF was focused more on buffing belts than nerfing bots. They even went through examples on how they had considered doing this.

The second FFF literally said in the conclusion that they think bots should be nerfed and listed a couple ideas they had for doing so.

And I want my youth back. Doesn't mean I have a way of doing that. They said each idea had major drawbacks and I don't think they will implement anything that just replaces one problem with another one.

[Engimage]

Again a simple way to nerf bots for enormous throughput but still not break megabases as this is a real concern.

Make bots consume greater amounts of energy based on active bots number in the network.

The more bots are flying the progressively more every bot consumes. This is This should not be a too complex development wise and will kinda punish usage of bots for too high throughput.
It will reflect the increased bot flight calculation CPU load that will consume more power.

Or just leave them alone for good. Without quality changes such nerfs will not reach the target.

[Aunte]

Make bots consume greater amounts of energy based on active bots number in the network.

Image an energy-issue, some bots move slower... now even more bots have to become active and it gets worse...