So... Let's talk about bots, and how to fix them properly...

Post all other topics which do not belong to any other category.
Post Reply
redis
Inserter
Inserter
Posts: 44
Joined: Sat Aug 31, 2019 4:03 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by redis »

Adamo wrote: ↑
Sat Aug 31, 2019 9:00 pm
redis wrote: ↑
Sat Aug 31, 2019 8:30 pm
THe real issue here is that solar panels are OP. You should try Adamo Physics! It nerfs solar to an amount roughly based on the maximum theoretical throughput of a solar panel that size. It's still possible to make giant solar fields that work, but much more expensive. If you're powering your massive bot collections on nuclear power, then, what's the problem? You've won. Enjoy. Nuclear power is that awesome, even in real life: what's the issue? It's not infinite. You will run it down faster by using bots. But isn't that the point?

Sorry, but I will strongly reiterate that we should not be considering a hard limit on the density of bots. We need an option in this game that allows you to have infinite density of throughput, and the cost of massive electricity is a fair trade. Solar is OP, though, for sure.
There would be NO hard limit on the density of bots. It is possible to adjust number of bots each roboport can support or improve bots themselves with research (speed , cargo , recharge). You can have infinite research for bots making them very powerful. In fact it would only benefit ups instead of clouds of flying entities. It would simply put bots on the same page as belts where they are NOT unlimited.

Additionally energy in the game is just fine. It is logical that if space is infinite then resources are infinite and so is energy. Trying to make adjustments to energy to solve some specific issue with bot logistics will simply break other mechanics of the game or have too much impact. Bots trivialize the game making it not worth using any other ways of transportation if the goal is RPM or SPM, which are the most common goals post rocket launch. So, no I did not win by building nuclear plant.

User avatar
5thHorseman
Smart Inserter
Smart Inserter
Posts: 1193
Joined: Fri Jun 10, 2016 11:21 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by 5thHorseman »

Okay page 7 finally time to give my view.

Bots can be fixed very easily.

Make an early construction bot that runs on Steam, or doesn't fly, or something. Something you can get with Red science and make better with Green. Maybe have 2 recipes, basic bot that is terrible but works for free and another bot that uses basic engines in its "robot frame" and runs on fuel, and works at a speed based on what fuel you use in it like trains. It only takes 1 fuel at a time and moves like the basic bot when out of fuel, so it can limp back to refuel.

That's it. I fixed bots. The only problem with bots is you don't get construction ones early enough.

Next up, blueprints, and how to fix them by making them more ubiquitous.

User avatar
Oktokolo
Filter Inserter
Filter Inserter
Posts: 883
Joined: Wed Jul 12, 2017 5:45 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Oktokolo »

The "construction bots are too late" issue would indeed be fixable by having red/green science range limited construction bots needing some fuel like Nanobots, but with actual land-bound entities driving around to place things like Construction Drones (but behaving properly).

Late bots being the ultimate belt replacement is another problem. Solving it without making logistics bots useless is achieved easily by having bots needing time to pick up or store items and having a concurrency limit for containers. Bots would have to queue up to access entity inventories. Interaction duration and maximum count of bots wich can access an inventory at the same time are tuneable to any belt speed equivalent.
So the role of bots could be set to anything between "only usefull for those freak recipes where they make you need some sulphur in your science cluster" and the current "best option for low/mid range mass transport".

PTTG
Long Handed Inserter
Long Handed Inserter
Posts: 51
Joined: Sat Nov 18, 2017 7:47 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by PTTG »

Why not make it so that only one bot can use a logistics chest at a time, and it takes time for a bot to pull from a chest? This puts them securely in their role of distributing rare resources in a flexible way without overlapping bulk transport. It can be manipulated by developing more and more evenly distributed bot logistics chests, without making unlimited robot grids tenable.

rhynex
Long Handed Inserter
Long Handed Inserter
Posts: 63
Joined: Tue Apr 17, 2018 9:55 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by rhynex »

-- deleted because of misunderstanding --
Last edited by rhynex on Sun Sep 01, 2019 3:51 pm, edited 1 time in total.

User avatar
5thHorseman
Smart Inserter
Smart Inserter
Posts: 1193
Joined: Fri Jun 10, 2016 11:21 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by 5thHorseman »

Oktokolo wrote: ↑
Sun Sep 01, 2019 5:27 am
Late bots being the ultimate belt replacement is another problem.
I don't believe I will ever agree with that statement. Bots being better than belts is exactly as much a problem as electric miners being better than hand mining.

User avatar
jodokus31
Smart Inserter
Smart Inserter
Posts: 1599
Joined: Sun Feb 26, 2017 4:13 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by jodokus31 »

I think logibots might be get a bit OP after game is finished (rocket launched), because they scale with space science research. I don't think, that it can solved without making them useless. Maybe some minor tweaks such as power consumption. Some hard limits are problematic. Or some things, to make them less viable for bulk items (ore, plates, etc..).
EDIT: Just thought about the stroked sentence. Logibots are maybe totally fine for mining (and on-site smelting), because it doesn't add much to the puzzle to make it belt based. And increasing mining productivity makes it hard to scale. Belt based designs get more interesting, if you have to handle multiple inputs and beacons.

The OP'ness appears after the game is officially finished and you already fought your way up to yellow science with belts. IMO thats the most interesting part of factorio and I can't get myself into serious megabasing. I'd rather install mods, to extend this "startphase"...

Belts don't scale with space science research. It would be cool, if we found a factor, which could be researched to make them competitive or slightly better.
The are already solutions with stackboxes, loaders, faster belts, wide chests etc. But it seems, they don't make it into vanilla.
Angels mods uses those coil/sheet items, which has ~4 times density of plates/wires, which is not a bad solution. But bots also profit from this.

User avatar
BlueTemplar
Smart Inserter
Smart Inserter
Posts: 2420
Joined: Fri Jun 08, 2018 2:16 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by BlueTemplar »

rhynex wrote: ↑
Sun Sep 01, 2019 10:28 am
why no one talks about nerfing zero-UPS solar power but bashes on bots?
They did, a few posts above. If you post a wall of text, maybe at least take the time to red the previous comments, at least on this page ?

----
redis wrote: ↑
Sat Aug 31, 2019 8:13 pm
The fact is that bots have nearly infinite throughput compared to belts and trains is very OP.
"Nearly infinite ?"

In that (in)famous FFF, kovarex deduced that logibots were 2-5(+?) times stronger than belts.
https://factorio.com/blog/post/fff-225
Image

Image
And if I'm not mistaken, logibots were nerfed shortly after that ?
And then belts were buffed in 0.17 too...
So what is the ratio now ?

(Also, I'd expect trains to be much faster than belts or even logibots...)

----

Also, yeah, while conbots and logibots share the same roboport, they have very different issues. (Even the roboport ranges are different !)
I don't feel that conbot discussions belong in this thread ?
BobDiggity (mod-scenario-pack)

User avatar
Oktokolo
Filter Inserter
Filter Inserter
Posts: 883
Joined: Wed Jul 12, 2017 5:45 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Oktokolo »

rhynex wrote: ↑
Sun Sep 01, 2019 10:28 am
first of all I dont think bots are OP.
If you don't think, they need to be fixed, then you probably are in the wrong thread here.
rhynex wrote: ↑
Sun Sep 01, 2019 10:28 am
this is easy to fix you know, put more chests around the assembler.
rhynex wrote: ↑
Sun Sep 01, 2019 10:28 am
a blue belt carries 2700 item/m. if we have a requester chest demanding 2700 iron plate, lets go for 1 sec drop penalty for each bot or chest, whatever. I shall keep concurrency number 1 for start. every bot carries 3 item at one time and it takes 1sec to unload. it makes 180 item/m. 2700 vs 180. this is serious penalty and I picked the shortest duration, 1 sec. grats, you killed bots entirely.
rhynex wrote: ↑
Sun Sep 01, 2019 10:28 am
in my above design where we put chests around an assembler like crazy we shall have 12 chests (assuming all of them are inputs) and we could have 12 belts anyway. 12*180 makes it 2160. still lower than a single blue belt. again, grats you have killed bots entirely.
It is either easy to get around the nerf or the nerf killed bots entirely. Both statements can't be true at the same time.
And nerfing bots to below blue belt speed in tight beaconed designs is not killing them. It is intended in stopping them to be the ultimate belt replacement everywhere.
They certainly would still be superior for distributing a huge assortment of items to multiple points or the player (wich probably would be exempt from the inventory access concurrency restriction for QoL reasons). They would still be able to outrun yellow belts easily and red belts with multiple boxes.

But i would probably want access time to be below 0.5 seconds to not make them feel too laggy and your calculations (wich i was too lazy to do myself) show that even with concurrency 1 the nerf just doesn't work.
Incorporal flying bots seem to literally be conceptually overpowered in a way that makes them immune to nerfs even for the most easy to nerf use case of tightly beaconed designs. If nerfing does not work there it works nowhere. Case closed.

rhynex
Long Handed Inserter
Long Handed Inserter
Posts: 63
Joined: Tue Apr 17, 2018 9:55 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by rhynex »

Oktokolo wrote: ↑
Sun Sep 01, 2019 3:33 pm
rhynex wrote: ↑
Sun Sep 01, 2019 10:28 am
first of all I dont think bots are OP.
If you don't think, they need to be fixed, then you probably are in the wrong thread here.
ok cool. I misunderstood then. just make your changes as "mods" and dont touch my bots. thanks.

then should not this discussion be at viewforum.php?f=34 ?

redis
Inserter
Inserter
Posts: 44
Joined: Sat Aug 31, 2019 4:03 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by redis »

Adding time to access chests does not eliminate the problem. You can create more chests instead. I want to reiterate this proposal

Allow ROBOPORTS to be placed only with NO OVERLAP of their range. You would not be able to place roboports very close to each other because of "interference".

This will reduce number of charging stations and increase distances to fly to recharge lowering throughput. To keep it possible to improve bot throughput more research bot upgrades could be added like more cargo, charge time and speed . If this is not enough then also more bots could be allowed in the roboports. Moving roboports further apart will reduce number of bots flying and and improve UPS. The biggest issue is spamming roboports strategies allowing to create clouds of bots with unlimited throughput in the late late game.

To support existing bases this needs to be done as optional feature, which could be disabled with a checkbox or mod.

User avatar
Oktokolo
Filter Inserter
Filter Inserter
Posts: 883
Joined: Wed Jul 12, 2017 5:45 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Oktokolo »

rhynex wrote: ↑
Sun Sep 01, 2019 3:47 pm
ok cool. I misunderstood then. just make your changes as "mods" and dont touch my bots. thanks.
Only devs could touch your bots and it really doesn't look like they are planning on doing that.
Bots are the most UPS efficient logistics solution by design and devs will not do anything that makes UPS tank for megabases. They will not add any checks or AI/pathing improvements for that reason.
They also have already seen that nerfing bots by just adjusting the existing parameters is impossible without making them annoying for player deliveries too.

So don't worry: Nobody will touch your bots.
rhynex wrote: ↑
Sun Sep 01, 2019 3:47 pm
then should not this discussion be at viewforum.php?f=34 ?
Yes. The thread starter maybe had (and maybe still has) hope that devs will touch bots if they see some compelling way to do it though.

User avatar
Oktokolo
Filter Inserter
Filter Inserter
Posts: 883
Joined: Wed Jul 12, 2017 5:45 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Oktokolo »

redis wrote: ↑
Sun Sep 01, 2019 3:57 pm
Allow ROBOPORTS to be placed only with NO OVERLAP of their range. You would not be able to place roboports very close to each other because of "interference".
Gapless logistics coverage is a good thing to have when using bots for player deliveries.
A minimum distance between ports wich is lower than the logistics radius might work though.

redis
Inserter
Inserter
Posts: 44
Joined: Sat Aug 31, 2019 4:03 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by redis »

Oktokolo wrote: ↑
Sun Sep 01, 2019 4:10 pm
redis wrote: ↑
Sun Sep 01, 2019 3:57 pm
Allow ROBOPORTS to be placed only with NO OVERLAP of their range. You would not be able to place roboports very close to each other because of "interference".
Gapless logistics coverage is a good thing to have when using bots for player deliveries.
A minimum distance between ports wich is lower than the logistics radius might work though.
Yes, definitely no gaps. So I agree may be some small overlap in range could be allowed, but roboports must be spaced out from each other. Currently it is possible to place roboports without their range overlap and still no gaps.

User avatar
Oktokolo
Filter Inserter
Filter Inserter
Posts: 883
Joined: Wed Jul 12, 2017 5:45 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Oktokolo »

redis wrote: ↑
Sun Sep 01, 2019 4:13 pm
Yes, definitely no gaps. So I agree may be some small overlap in range could be allowed, but roboports must be spaced out from each other.
A minimum distance from other ports should be enforcable in build event handlers.
Now we only need someone less lazy than me to actually mod it. ;)

Hannu
Filter Inserter
Filter Inserter
Posts: 850
Joined: Thu Apr 28, 2016 6:27 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Hannu »

5thHorseman wrote: ↑
Sun Sep 01, 2019 10:43 am
I don't believe I will ever agree with that statement. Bots being better than belts is exactly as much a problem as electric miners being better than hand mining.
It depends on your point of view. In my opinion belting and piping stuff is interesting logistic problem and also aesthetically nice thing to do. Immaterial bots are cheap cheats, which make everything trivial to build and ugly and unrealistic to see. Possible problems are not interesting and limits are practically limits of my computer and not limits of my creativity. I like to run around my base and see when my complex belt systems and all those inserters work. It gives much more than simple number which tells how much my base produces in minute. I like also to draw electronic circuit boards with drawing software instead on actual PCB planning tool (of course only simple 2 layer PCBs). It is quite much like belting in Factorio.

Of course I see that different opinions exist and not all like to plan belts or see the beauty of spaghetti. There will not be any solution which would make everyone happy. Megabase builders are too used to have easy immaterial bots and they seem to be the most important customers for devs. I can handle the situation with personal rules (I use logistic bots only for malls, not in main material flows (except maybe satellite or nuclear fuel capsules).

User avatar
5thHorseman
Smart Inserter
Smart Inserter
Posts: 1193
Joined: Fri Jun 10, 2016 11:21 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by 5thHorseman »

Hannu wrote: ↑
Mon Sep 02, 2019 12:24 pm
5thHorseman wrote: ↑
Sun Sep 01, 2019 10:43 am
I don't believe I will ever agree with that statement. Bots being better than belts is exactly as much a problem as electric miners being better than hand mining.
It depends on your point of view. In my opinion belting and piping stuff is interesting logistic problem and also aesthetically nice thing to do. Immaterial bots are cheap cheats, which make everything trivial to build and ugly and unrealistic to see.
Like I said, we are not going to agree. I'm cool with that. I'm not cool bots being nerfed even more because of it.

Italian Cuisine
Inserter
Inserter
Posts: 24
Joined: Sat Mar 02, 2019 1:24 pm
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by Italian Cuisine »

redis wrote: ↑
Sun Sep 01, 2019 3:57 pm
Adding time to access chests does not eliminate the problem. You can create more chests instead.
Creating more chests is the intent, and it's a good one. By adding more chests, you cannot directly increase the throughput to single dense assemblers, like you can by adding more bots and a bunch of semi-distant roboports.

More chests means dealing with inserters and belts in the mix, if you want to combine the throughput of multiple chests together. That's a good combination of Factorio gameplay elements that allows you to increase throughput while still requiring managing physical and logistic scalability. The current bot mechanisms avoid that, and nerfing charging stations still wouldn't place any demands of scalability at the actual point of item consumption.

--

Regarding early game construction bots, my preference would be not bots, but a Robotic Construction Arm that you install on your armor. Say it has the reach of a long-handed inserter, and can construct items from your inventory into locations immediately around your character. There would be no long distance construction, and no mass parallel construction, but would simply automate the actual placing & production selection of your blueprints, as you walk through it.

I think that would be a great early game automation assistance, as your multi-assembler units get larger, without breaking too much into OP territory that early.
It's not spaghetti!

foamy
Filter Inserter
Filter Inserter
Posts: 432
Joined: Mon Aug 26, 2019 4:14 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by foamy »

Italian Cuisine wrote: ↑
Wed Sep 04, 2019 11:20 pm
redis wrote: ↑
Sun Sep 01, 2019 3:57 pm
Adding time to access chests does not eliminate the problem. You can create more chests instead.
Creating more chests is the intent, and it's a good one. By adding more chests, you cannot directly increase the throughput to single dense assemblers, like you can by adding more bots and a bunch of semi-distant roboports.

More chests means dealing with inserters and belts in the mix, if you want to combine the throughput of multiple chests together. That's a good combination of Factorio gameplay elements that allows you to increase throughput while still requiring managing physical and logistic scalability. The current bot mechanisms avoid that, and nerfing charging stations still wouldn't place any demands of scalability at the actual point of item consumption.

--

Regarding early game construction bots, my preference would be not bots, but a Robotic Construction Arm that you install on your armor. Say it has the reach of a long-handed inserter, and can construct items from your inventory into locations immediately around your character. There would be no long distance construction, and no mass parallel construction, but would simply automate the actual placing & production selection of your blueprints, as you walk through it.

I think that would be a great early game automation assistance, as your multi-assembler units get larger, without breaking too much into OP territory that early.
Honestly I like all of this; problem is, right now, you get bots more or less simultaneously with the first piece of armour that includes an equipment grid of any sort.

redis
Inserter
Inserter
Posts: 44
Joined: Sat Aug 31, 2019 4:03 am
Contact:

Re: So... Let's talk about bots, and how to fix them properly...

Post by redis »

Italian Cuisine wrote: ↑
Wed Sep 04, 2019 11:20 pm
redis wrote: ↑
Sun Sep 01, 2019 3:57 pm
Adding time to access chests does not eliminate the problem. You can create more chests instead.
Creating more chests is the intent, and it's a good one. By adding more chests, you cannot directly increase the throughput to single dense assemblers, like you can by adding more bots and a bunch of semi-distant roboports.

More chests means dealing with inserters and belts in the mix, if you want to combine the throughput of multiple chests together. That's a good combination of Factorio gameplay elements that allows you to increase throughput while still requiring managing physical and logistic scalability. The current bot mechanisms avoid that, and nerfing charging stations still wouldn't place any demands of scalability at the actual point of item consumption.

--

Regarding early game construction bots, my preference would be not bots, but a Robotic Construction Arm that you install on your armor. Say it has the reach of a long-handed inserter, and can construct items from your inventory into locations immediately around your character. There would be no long distance construction, and no mass parallel construction, but would simply automate the actual placing & production selection of your blueprints, as you walk through it.

I think that would be a great early game automation assistance, as your multi-assembler units get larger, without breaking too much into OP territory that early.
This (limiting chest access) is better than it is now for sure because you are limited how many chests you can place. It still promotes spamming of robports and now chests too. Limiting roboports distance is a better approach.

In fact, I think this is how devs meant the game to be played ( and it is how it is played in early stages when getting to the first rocket). It shows those lines when roboports connect with each other when you are placing them. I do not think they thought about someone placing a square of 5x5 or more robports in one place (in fact there is no point in it early on). It is just a loophole they "forgot" to close allowing to create massive number of bots in a small area.

Devs only think about the game up to the point of the first rocket and late game balance is completely screwed up with this bots bias. You should not need to have so many bots in the first place. It is better to compensate with large cargo if needed. Better for UPS and same goal of throughput can be achieved.
Last edited by redis on Fri Sep 06, 2019 1:21 am, edited 1 time in total.

Post Reply

Return to β€œGeneral discussion”