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

[Gemoron]

Hi, first of all I want to note that I didn't read all the pages beforehand and I think I won't reply to any following ones. I'd just like to post my opinion for the devs to read and concider.

I don't mind Bots existing but I like to build my factories based on belts supportet by Bots to evade useless spagetti, e.g. bringing science packs to the research area. To define my playstyle in the Timmy/Jonny/Spike (Go big/Self expression/competetive) I guess I am mostly a Jonny with a bit of Timmy.

First: I do love the new splitter mechanics and I have felt so long that the new types is how they should be supposed to work. smart definition on where items should go and priority functions without having to hack them in with sideloading will allow more compact controll of the factory item flow. This will also be more accessable to new players without having to dig through reddit for some ominous black magic sorters.

Personally I think Robots shouwln't be nerfed as it is free for everyone to usethem or not or how to use them. Yet I think belts require a buff and the new splitter function is a welcome one. Another possible buff was shown, but was discarded long ago, but they are visible in your screenshots: loaders. I think they should be developed to the final level. they offer a higher input in chests or factories for a fair tradeoff which is you require more space to place them, e.g. Belt->Inserter->factory requires 2 tiles. A loader needs the Belt, a splitterand a loader, typically making the same effect 3 or 4 tiles wide which makes beaconing more difficult. On the other hand, loaders would allow fast inline-storage. I usually get loaders from mods, but I'd like to see them in classic Factorio as well

[mtilsted]

How about letting bots use up a resource other then power. Something lubricant based such as a "robot cleanup kit"* would be the obvious choice. That way you can still build a robots only base, but robots would not seem like the obviously best choice for anything.

Ps: Personally I play without logistic robots and chests.

*Could cost for example 5 lubricant, and a gear and be required each time the robot have been flying for 4 minutes.

[Durabys]

Hi, first of all I want to note that I didn't read all the pages beforehand and I think I won't reply to any following ones. I'd just like to post my opinion for the devs to read and concider.

I don't mind Bots existing but I like to build my factories based on belts supportet by Bots to evade useless spagetti, e.g. bringing science packs to the research area. To define my playstyle in the Timmy/Jonny/Spike (Go big/Self expression/competetive) I guess I am mostly a Jonny with a bit of Timmy.

First: I do love the new splitter mechanics and I have felt so long that the new types is how they should be supposed to work. smart definition on where items should go and priority functions without having to hack them in with sideloading will allow more compact controll of the factory item flow. This will also be more accessable to new players without having to dig through reddit for some ominous black magic sorters.

Personally I think Robots shouwln't be nerfed as it is free for everyone to usethem or not or how to use them. Yet I think belts require a buff and the new splitter function is a welcome one. Another possible buff was shown, but was discarded long ago, but they are visible in your screenshots: loaders. I think they should be developed to the final level. they offer a higher input in chests or factories for a fair tradeoff which is you require more space to place them, e.g. Belt->Inserter->factory requires 2 tiles. A loader needs the Belt, a splitterand a loader, typically making the same effect 3 or 4 tiles wide which makes beaconing more difficult. On the other hand, loaders would allow fast inline-storage. I usually get loaders from mods, but I'd like to see them in classic Factorio as well

The solution to bots is NOT nerfing bots. It is nerfing chest access speed/volume and at the same time also allowing inserters faster ouput out of chests. Either through side inserters or through research upgrades increasing their maximum overall (not just for stack inserters) items per arm swing. Both would have to be implmented: Nerfing chests and boosting inserters. Not nerfing robots.

[quickshot0]

I think the main issues are Bot access to Logistic chests and the inability of more then four (eight if you combo yellow and red) inserters from accessing a single chest. Bots are simply TOO good at taking and giving items to a chest and inserters are simply too weak. A thousand bots are able to take form a single chest. In real-life? That should be utterly impossible because they would crash into each other. Only four stack inserters are able to access a chest. In RL we have robotic arms who can in a swarm of dsozen or more can dance around each other when working on a single vehicle in car factory

So a solution:
1. Only one or several (but definitelly not an infinite number or even a high double digit number) bots can simultanously access a single logistic chest.
2. Side Inserting and Long Arm and FIlter upgrades to all Inserters: DELETE Red Inserters and Filter Normal and Stack Inserters from the game. Add several research upgrades that add Long Arms, Filtering and Side Inserting as otpion to Yellow, Blue and Green Inserters. This allows for Long Armed Filter Stack Side Inserters in late game: Capable to take or insert from chests from from tiles away like a red inserter and from any angle possible, this guys and gals:

I made a post on this previously. But better inserters like that don't just allow for better belts and chest use, they actually allow you to change the end game. You can actually eliminate belt for short transport moves up to 7 blocks or 13 blocks away with 1 chest. This allows you to reformat entire blocks of your base in to inserter only type sections.

I did that with such inserters once and it was actually pretty fun as it made you think about new efficient layouts later in the game. Belts ultimately got changed to mid range transport options, instead of moving goods to every single assembler. (Also obviously one or two inserters is less of a UPS impact then then the two inserters and one bot that a bot base uses)


I personally think that as such that making inserters much better in the late game might actually be a big help in the late game. Especially as it would potentially not need anything to be made 'worse', while at the same time allowing one to reinvigorate gameplay by making all new factory layouts possible that you couldn't do previously.

[Drone]

One idea to fixing bots v belts is to give bots a pickup and drop-off time, during which other bots are unable to interact with the chest. This limits the throughput of them without ending up with the situation where you "just build more bots" because that doesn't work without adding more chests (which depending on the design may be impossible). Making so that chests have a 0.25 second pickup/drop-off time (for example) still leaves bots useful for constructing blueprints, providing the player with resources and moving modest quantities of goods around (for example for late game automation of construction of barely used buildings such as pumps and tanks or similar) but limits their ability to render belts obsolete for moving vast quantities of goods around.

You can still work around this (for example, by having multiple logistics chests fed from splitters, or similar) for those who demand to use bots for everything, and it would still be effective depending on the design, but it would severely limit the ability of bots to be competitive with belts in situations where belts should be the logical choice (mid-range, bulk movement of goods)

[ILsauro]

Could stack belt be a solution for late game belts?
White: normal belt
Yellow: new entity that pushes items down a "layer"
Red: two layer belt (iron on top, copper on the first underground layer)
Blue: three layer belt (steel on top, iron on the first underground layer and copper on the second)

You could have a research that increases the maximum number of layers.

Sorry for my bad english (it's not my motherlanguage)

[Durabys]

I think the main issues are Bot access to Logistic chests and the inability of more then four (eight if you combo yellow and red) inserters from accessing a single chest. Bots are simply TOO good at taking and giving items to a chest and inserters are simply too weak. A thousand bots are able to take form a single chest. In real-life? That should be utterly impossible because they would crash into each other. Only four stack inserters are able to access a chest. In RL we have robotic arms who can in a swarm of dsozen or more can dance around each other when working on a single vehicle in car factory

So a solution:
1. Only one or several (but definitelly not an infinite number or even a high double digit number) bots can simultanously access a single logistic chest.
2. Side Inserting and Long Arm and FIlter upgrades to all Inserters: DELETE Red Inserters and Filter Normal and Stack Inserters from the game. Add several research upgrades that add Long Arms, Filtering and Side Inserting as otpion to Yellow, Blue and Green Inserters. This allows for Long Armed Filter Stack Side Inserters in late game: Capable to take or insert from chests from from tiles away like a red inserter and from any angle possible, this guys and gals:

I made a post on this previously. But better inserters like that don't just allow for better belts and chest use, they actually allow you to change the end game. You can actually eliminate belt for short transport moves up to 7 blocks or 13 blocks away with 1 chest. This allows you to reformat entire blocks of your base in to inserter only type sections.

I did that with such inserters once and it was actually pretty fun as it made you think about new efficient layouts later in the game. Belts ultimately got changed to mid range transport options, instead of moving goods to every single assembler. (Also obviously one or two inserters is less of a UPS impact then then the two inserters and one bot that a bot base uses)


I personally think that as such that making inserters much better in the late game might actually be a big help in the late game. Especially as it would potentially not need anything to be made 'worse', while at the same time allowing one to reinvigorate gameplay by making all new factory layouts possible that you couldn't do previously.

Perhaps add the devs option of creating stack bbelts?

Technology which would turn all belts into stack belts.

*You would research this and automagically all inserters, splitters and side-loading (also mining drills) would be able to make stacks on any type of the 3 belt tiers we have now.

*You don't get any new entity, but you get the choice between upgrading all of your belts with this technology and not having to do any extra work, or tearing up your base and replacing with robots.

*When you get the research, all of your belts would start buffering extra items, which some people might find annoying.

*If we could code this and find a visual solution which wouldn't take too much work, I would prefer this option.

[Caine]

Currently we have three main tools for solving logistic problems: belts, trains and bots. Strictly speaking we also have pipes, but for simplicity we consider them as liquid belts.

Ideally these all have their strengths and weaknesses such that there are always tradeoffs and thus decisions to be made. To see how balanced it is, we need to have some metrics for comparison. I postulate the following metrics to be of importance:

• Throughput: The number of items we can transfer per second.
• Flexibility: Can it handle multiple resources easily? Is it fixed infrastructure or does it dynamically adapt to changing circumstances.
• Space Efficiency: How large is its footprint. A low space efficiency is much harder to fit in (e.g. in beaconed setups).

I think that currently the situation looks like this:

Current.PNG (9.16 KiB) Viewed 6636 times

It is immediately clear why bot-only bases are so viable.

Ignoring specific solutions for the moment, what do we want this table to look like? As an example take the following table which is nicely balanced, each solution is strong in one metric, weak in another and average in the remaining.

Example.PNG (9.55 KiB) Viewed 6636 times

Now we are in a much better position to evaluate individual changes to belts and/or bots. Do they take us in the right direction?

I want to pose some questions to the community

• Are these the proper metrics? If not, which ones should add/remove and why?
• What do we want the table to look like?

[Nick-Nack]

Logistic bots already cost more than they add compared to belts. Any further nerf without a cost reduction would further make them cost a ridiculous amount for what they add. You could make 25k blue belts just from the research cost difference alone. Each bot costs 3.4 times as much as each additional belt and roboports are another 13 some blue belts ea h. Using bots now is already 35% less cost efficent by Kovarex's throughput numbers just based on research costs.

Ressource cost just isn't the right measure, because of automation: As a player it doesn't matter how much it costs, because the factory will pay for me. What matters is time cost, because I as a player have to pay that myself, and time cost of bots is currently way less than for belts.

And yeah, I've mentioned it before and lots of others wrote the same thing, but just to make sure the devs see it:
Limiting bots per second that can access chests is the ultimate solution.

EDIT: I like Cains metrics, and limiting bots per second that cannaccess cehsts would effectively lead to the second table he posted.

[crysanja]

Hmm maybe bacons and radars should project a no logistics field.

[fishycat]

If you really follow that path, that you, for some unknown reasons choose versus the end of last year, to the end, there will be a point in the future where you look back and say, yes, that was a big mistake!

I trusted every single decision of yours in all that time, whether i liked it or not - but in this special case, I feel you are wrong and don't even notice it.

Please, take a step back, think it all over and come to the right decisions.

[hitzu]

I'm not sure how would you make belts more space efficient other than totally getting rid of inserters in belt setups and embedding a direct loading from the belts into containers and assemblers.

[Durabys]

Currently we have three main tools for solving logistic problems: belts, trains and bots. Strictly speaking we also have pipes, but for simplicity we consider them as liquid belts.

Ideally these all have their strengths and weaknesses such that there are always tradeoffs and thus decisions to be made. To see how balanced it is, we need to have some metrics for comparison. I postulate the following metrics to be of importance:

• Throughput: The number of items we can transfer per second.
• Flexibility: Can it handle multiple resources easily? Is it fixed infrastructure or does it dynamically adapt to changing circumstances.
• Space Efficiency: How large is its footprint. A low space efficiency is much harder to fit in (e.g. in beaconed setups).

I think that currently the situation looks like this:

Current.PNG

It is immediately clear why bot-only bases are so viable.

Ignoring specific solutions for the moment, what do we want this table to look like? As an example take the following table which is nicely balanced, each solution is strong in one metric, weak in another and average in the remaining.

Example.PNG

Now we are in a much better position to evaluate individual changes to belts and/or bots. Do they take us in the right direction?

I want to pose some questions to the community

• Are these the proper metrics? If not, which ones should add/remove and why?
• What do we want the table to look like?

Space Efficiency of Belts is increased by the last proposed dev option:

Technology which would turn all belts into stack belts.

*You would research this and automagically all inserters, splitters and side-loading (also mining drills) would be able to make stacks on any type of the 3 belt tiers we have now.

*You don't get any new entity, but you get the choice between upgrading all of your belts with this technology and not having to do any extra work, or tearing up your base and replacing with robots.

*When you get the research, all of your belts would start buffering extra items, which some people might find annoying.

*If we could code this and find a visual solution which wouldn't take too much work, I would prefer this option.

Thoughput of Bots can be limited heavilly if you add a maximum limit for how many bots can access a single chest in one moment. Which is basically what I propose.

Space Efficiency? You should count the neccessary space and resources to build needed Solar Fields and Accumulators or Nuclear Mining&Refining&PowerGeneration into the such an equation too. That is easy to change too. One or multiple of these options: Increase power demand. Nerf the Kovarex Process. Make Solar panels much bigger and requiring more space per panel. Or you could also decrease the roboport logistic radius like you proposed (which hamper Beacon Set Ups). Or increase the size of the robotport.

I'm not sure how would you make belts more space efficient other than totally getting rid of inserters in belt setups and embedding a direct loading from the belts into containers and assemblers.

The devs proposed upgrading all belts with one research into stack belts as the ir final option that they like the most.

[quickshot0]

Perhaps add the devs option of creating stack bbelts?

For instance, right. Though the idea of an enclosed express belt system with higher throughput could also work. You could then have small unloading areas from the main belt for each larger assembly area. The enclosed system also allows for it to be more CPU efficient, as you don't have to render-track all the the goods to the same detail any more between each unloading area. Though I'm not sure how people would think it looks compared to normal belts. (One could explain it in game as something like an air reduced area, where things can go faster due to lower air resistance)

I would note it's probably best to not exceed the trains ability to move large amounts of goods though, it really being an intermediate option instead. I suspect just doubling current throughput would be sufficient for that already really, as the bot throughput example is a highly idealised case and doesn't work that well in practise.


It seems to me atleast that this is a fairly reasonable option to having all game elements continued to be used, while not introducing some kind of major limitations to any of the systems. And rather just further expanding on their scope of two of them. This means everyone can also continue to build bases like they want, but that people chasing the 'most efficient' high productive designs now have some new options to play with that will likely lead to new designs.

So better gameplay by making inserters strong enough to outcompete bots at short ranges and to make belts better in real world mid-range options. Thus pushing bots to be used more for difficult logistical issues rather then many of them.

[Frontrider]

I haven't built any big bases yet (if I think about it its probably because of the lower belt performance, combined with my refusal to use only bots)

My production lines usually ends up looking like this:(after a few games)
-train input
-belts for supply
-bots:
-to provide temporary fixes (when the only thing matters is to get that product made, before doing the upgrade I actually need)
-move overflow/trash (SOmetimes have to)
-filter (will use splitters when they get in)
-manufacture "obscure" materials what I don't need a whole lot of.

I haven't looked at numbers, but to me the bots actually look scary when I think about lag. (even tough the numbers say otherwise)

[Caine]

I'm not sure how would you make belts more space efficient other than totally getting rid of inserters in belt setups and embedding a direct loading from the belts into containers and assemblers.

I have seen several proposals over the last week that you could interpret as increasing space efficiency of belts:

• The new splitter. It potentially makes sushi belts more feasible, thus making it easier to route multiple inputs to a factory. This is an increase in flexibility at the cost of throughput which reduces the need for multiple input belts (freeing up space).
• Belts in the air, such as e.g. cable cars on wires or by stacking belts (as opposed to putting stacks on belts).
• Configurable inserters (e.g. pick and drop at an angle, which reduces belt placement constraints).
• An underground world layer (like factorissimo) where belts can be routed freely below structures.

Changes will typically affect more than one metric as all the building blocks are interconnected.

[hitzu]

I'm not sure how would you make belts more space efficient other than totally getting rid of inserters in belt setups and embedding a direct loading from the belts into containers and assemblers.

The devs proposed upgrading all belts with one research into stack belts as the ir final option that they like the most.

This would affect throughput. Space efficiency is how much space you need to build a production setup. Now the more complex the recipe the more complex and large the belt setup is (unlike the bot based setup which is always 2 chests, two inserters, and a power pole per an assembler). Beacons make this even worse.

[Durabys]

I'm not sure how would you make belts more space efficient other than totally getting rid of inserters in belt setups and embedding a direct loading from the belts into containers and assemblers.

The devs proposed upgrading all belts with one research into stack belts as the ir final option that they like the most.

This would affect throughput. Space efficiency is how much space you need to build a production setup. Now the more complex the recipe the more complex and large the belt setup is (unlike the bot based setup which is always 2 chests, two inserters, and a power pole per an assembler). Beacons make this even worse.

I posted this on Space Efficiency underneath that:

Space Efficiency? You should count the neccessary space and resources to build needed Solar Fields and Accumulators or Nuclear Mining&Refining&PowerGeneration into the such an equation too. That is easy to change too. One or multiple of these options: Increase power demand. Nerf the Kovarex Process. Make Solar panels much bigger and requiring more space per panel. Or you could also decrease the roboport logistic radius like you proposed (which hamper Beacon Set Ups). Or increase the size of the robotport.

[Durabys]

Could stack belt be a solution for late game belts?
White: normal belt
Yellow: new entity that pushes items down a "layer"
Red: two layer belt (iron on top, copper on the first underground layer)
Blue: three layer belt (steel on top, iron on the first underground layer and copper on the second)

You could have a research that increases the maximum number of layers.

Sorry for my bad english (it's not my motherlanguage)

Oh god. THIS! So much fucking this.
Devs please? Can you allow for same colour UG belts to always connect only to the Undeground exit you just created and not to any other? That would allow same colur stacking of UG belts!

[promaty]

How about a tier 4 belt that has a huge throughput but is too fast for inserters ? It can be specifically used for transporting bulk materials over main bus, then you split it into slower belts for production?