Friday Facts #224 - Bots versus belts

[Avezo]

What about indroducing diagonal belts?

[Ethribin]

A thought that i had when reading the friday facts.

One way to balance out bots is to just limit how many you can put in a roboport.
like... instead of having 8 (I think its that many) slots for any robot, maybe make only two slots, one specivicly for builders, one specivicly for logistics.
Or maybe two for logistics, or such.

Maybe balacing the bots vs the belts is less about balancing the bots themselves but about the sytem surrounding them.

Sure, you could just build more roboports, but maybe make the energy cost or the building cost, or both for that matter, for roboports rather high? I always felt roboports are no real hassle to build. And considerign what they are, shouldnt they be a bit more of a hassle?
Maybe even make it so that you have to use a minimal distance between roboports (which would suck, but might help)


My idea here is that robots would be more used as a support for a belt factorie. So their mobility is more flexible then belts, but their throughput is lower. So it's smarter to mix and match. But I'm not sure if that would be good, seeing throughput is what you want from transportation.... idk.
Just a thought.


Also, sorry if someone already brought up this idea.
Dont have the time to read through the 24 pages of posts ><

[Jürgen Erhard]

When in doubt, model reality.

Bots behave the way they do because this game basically has no pathfinding to speak of. Not an accusation, just a fact, a technical detail in fact: imagine having to pathfind for 10000 bots. Yay.

And your solutions would make this hellishly more complex too, code-wise. Not to mention it would be the worse possible nerf.

[Ralph]

Personally, I am always concerned by belt compression, and would be glad if inserters had a built-in functionality for that. In some way, belts would be easier to handle so bots superiority would be less flagrant.

[FrodoOf9Fingers]

I think most of the points have been hit on the head, but I really don't understand the thought of removing a feature from the game for no other reason than to make those who don't use the feature feel better.

As an aside, I build with bots for the midgame, but my goal with bots is only transitionary. My megabases are almost always belt based, with few exceptions for getting items to tight spaces (a silo completely surrounded with beacons, for instance).

Buffing belts, as mentioned, is the best way to go. Add some more utility, maybe a dedicated lane splitter, a "buffer spot" to make arms pull items off of a piece of belt as fast as pulling from a chest, make a circuit network condition for "is this piece of belt moving / is it full" so we can better negotiate with belts. My 2cents.

[TheSuperSewcio]

I would just remove infinite researches for bots (adding more roboports will still be a thing) and if it won't help, then buff belts.
Bots are really broken only by these researches in my opinion. I can't imagine such thing for belts (infinite yellow belt speed yay!) so why bots should be upgraded over and over again.

[FrodoOf9Fingers]

Could also limit the number of bots that can deliver to/from a chest at a single time.

[bobucles]

I'd just like to point out that the slow moving, extreme capacity belt system already exists in game. Build a car, drop it on a conveyor, and use inserters to manage its inventory. Add throughput with more cars. Easy.

The item throughput of a "carveyor" is absolutely insane. A single belted car has 80 slots of inventory space, and the slowest belt can easily push a dozen around per minute. The main weakness is that inserters are too slow. Cars are a small target and there's a very low limit on how many inserters can manage their incredible inventory space. You really need the high speed capacity of a Loader to keep the items moving at any decent speed. A secondary concern is that it builds up a huge inventory in limbo. That's a big issue for a smaller base where every item matters, but an endgame RPM base doesn't really care if there's 2 rocket's worth of items in the pipeline. A huge inventory lag is a fair price to pay for moving 20 belts of items at once.

Now I know that you don't want to drop cars on your conveyor belts. First off you can't blueprint it, so that's a huge pain. Second is that you need circuit conditions to manage the car inventories properly, and that's way beyond the expectations of an ordinary player. But consider a different type system which is basically a mini rail with mine carts on it. Huge amounts of a single item fill up a cart, and it glides down to where it gets unloaded. There are many similar suggestions for some kind of "bucket conveyor" that can achieve extreme throughput for those parts of the base that need it. I know that ore and plates aren't difficult to manage on their own, but belts can't handle moving the sheer bulk that a late game base needs. A bucket conveyor system can solve this. It needs to move a LOT of items, but the MIX of items doesn't matter because it's a lot of one thing and the horizontal SPEED doesn't really matter because the player needs throughput.

[calut]

I have intel Q660 2,4Ghz, 4Gb memory. When i am at the point of making robots, my refresh rate is 40. So i am very happy when i delete all the belts, and reconstruct all my base. If u what to balance, make recepies for robots and roboports far more complex(nr of items and resource). This whay, if u build some megabase, u need more resource to get there, u feel u accomplish a megabase, make copy paste bp harder)
if u nerf or kill robots, u lose players with slow computer like me. All big comunites games work on slow computers.

[SirMai]

I am sorry, if this idea has been posted earlier. 25 pages was too much for me ;p

Using the energy consumption to balance bots is not the best way in my opinion.
Those who use Bots extensively will most likely use solar panels to power them. That's a one time investment and after that its without any costs at all.
In my opinion the solution may be to simply introduce "robot maintenance".
Every robot would need a "maintenance kit" after every X tiles flown/howered/robothingsdone ;p

The alternative would be to remove free energy.
In this scenario either solar panels or accmulators would need those "maintenance kits".

This could be balanced easily by changing the costs of a "maintenance kit".

[Faen]

I don't know is this would require too many changes to logistic robots' logic, but I think that playing with different overlaping logistic areas that have a limit to its number of active logistic bots would be interesting, it would make logistics more fun to play with and at the same time we would have more control over bots:

The basic idea is to assing each roboport to a logistic network(A,B,C...) the logistic bots on that roboport would then be asigned to that network, on the other hand we would have to assign chest to either one or two of these networks so only bots on that network can interact with them.

In the image below you can see an example of such networks, first be have network A (Orange) its an extensive logistic network that envelops all the factory, second we have network B (Cyan), its a small network that gives coverage to some asembling machines.

The logistic robots on network B gather the products deposited in the active provider chests (Cyan) and deposit them in the "HUB" chest (cyan/orange). This chest is very special, it is asigned to both logistic netwoks but it acts as a requester chest for network B (cyan) and as a pasive provider for network A (orange).
This way network B would have the task to gather products into the HUB while network A distributes them through the whole factory.



This kind of overlaping networks would be very effective doing specific tasks, you can make a network that gather resources from furnaces, another that send products from one area to another (a corridor), one that distributes low demanding items to all the factory...

That way you can begin having a limited bot network early game and develop an interconected system of logistic networks later, adding networks to perform new tasks for you, creanting exchange areas between networks and planning your factory based on this logistic areas.



If HUB chests feel too complicated to implement/use you can just use single network chests and send products from network B requester chest to network A active provider chests using inserters, although this would limit throughput and would requiere bigger exchange areas.

[Rylant]

To everyone that is saying if you find bots op then jut don't use them. This is one of the most stupid arguments I ever heard. By this logic devs should never bother balancing their games since if something is op just don't use it.
(do the daggers deal more damage per hit, have better range, knockback, critical, dps than sword just don't use them. Does the fire magic make every other element irrelevant doesn't matter just don't use it.)

Also this does limit the choice since there isn't any choice. You will always use bots unless you are doing no bots challenge run (and if you are then you will always use belts).

I think your argument here is equally flawed. The dagger analogy is less a game balance issue, and more a realism issue. A dagger that is powerful isn’t unbalanced; however, when daggers are more powerful than long swords, it is not realistic. People aren’t arguing from a realism perspective here. Let’s be real, there are more than enough unrealistic things happening in Factorio. Nobody is saying “hey, I can literally fit hundreds of nuclear reactors in my backpack which makes no sense. We need to nerf nuclear reactors now because it’s unbalanced.”

Right now, many people play with the intent of creating a mega base. To achieve a mega base, there is only one viable way to do it, and that is with bots. People have literally spend hundreds of hours creating their mega bases. I even know one person who is over 1000 hours on his save. Nerfing bots means that these people will probably have to start over, or at the very least, massively overhaul their creations. You don’t understand it when these people say “please don’t do this”?

Rylant

[Mobius1]

To everyone that is saying if you find bots op then jut don't use them. This is one of the most stupid arguments I ever heard. By this logic devs should never bother balancing their games since if something is op just don't use it.
(do the daggers deal more damage per hit, have better range, knockback, critical, dps than sword just don't use them. Does the fire magic make every other element irrelevant doesn't matter just don't use it.)

Also this does limit the choice since there isn't any choice. You will always use bots unless you are doing no bots challenge run (and if you are then you will always use belts).

I think your argument here is equally flawed. The dagger analogy is less a game balance issue, and more a realism issue. A dagger that is powerful isn’t unbalanced; however, when daggers are more powerful than long swords, it is not realistic. People aren’t arguing from a realism perspective here. Let’s be real, there are more than enough unrealistic things happening in Factorio. Nobody is saying “hey, I can literally fit hundreds of nuclear reactors in my backpack which makes no sense. We need to nerf nuclear reactors now because it’s unbalanced.”

Right now, many people play with the intent of creating a mega base. To achieve a mega base, there is only one viable way to do it, and that is with bots. People have literally spend hundreds of hours creating their mega bases. I even know one person who is over 1000 hours on his save. Nerfing bots means that these people will probably have to start over, or at the very least, massively overhaul their creations. You don’t understand it when these people say “please don’t do this”?

Rylant

Agreed, removing/nerfing/blocking or doing something with bots isn't the result to improve belts, is like destroying all the apples hoping that oranges would grow red. You're not solving the problem, just creating another one.
Megabases simply CAN'T be done with belts. period. Hundreds of reasons clearly proves that belts can't handle late-game, which is why they can only carry 80 items/s they're ment for early to midgame (midgame read launching a rocket which is where your true logistic challenge will start).

[pleegwat]

To empower belts, I like the idea of introducing crates/pallets, which contain a full stack of any item, don't stack themselves, and cannot be carried by bots. Packing or unpacking could be done in a normal assembler. This potentially also makes belts more UPS-friendly, since the number of items moving around is much smaller.

Additionally I think some rebalancing of beacons, increasing effect range but adding diminishing returns, may allow more flexibility in layout and belt routing.

Increasing size of logistics chests could give a similar effect (since it would increase space required in bot setups) but probably would impact existing designs more than many players are comfortable with.

[Tercicatrix]

The strength of belts is that they move a quantity of items from point A to point B.
The strength of bots is that they are incredibly flexible and responsive to demands.

I think factorio would be best if it emphasizes these strengths, such that:
Belts are virtually always the optimum answer for moving a known quantity of set materials from point A to B.
and Bots are virtually always the optimum answer to problems that ask for flexibility, compactness, or coverage for 'spikes' in demand. (eg, supplying turrets with ammo, since running ammo belts to turrets has always been tedious and too simple to be fun)

And then, for advanced players, there should richness within edge cases where it turns out belts with circuit logic are better than bots, etc.

To help the strength of belts, I think other commenter's suggestions on stacking, etc, seems promising.

[BluetoothThePirate]

One thing I think we're overlooking is the fact that logistic bots are a fully automatic turnkey solution to a lot of problems that really should have multiple diverse solutions. The classic bot problem to me is complex late-tech items need a diverse set of things to make but are made in relatively small volume. That's a perfect use for a flying robot courier that carries a small volume of things. But when you have situations like "I can't get the incoming ore away from my train yard fast enough so I'll use a swarm of hundreds of dedicated single-purpose bots in an isolated network flying back and forth to move ore like ten tiles", that's a failure of the game to provide a better solution for unloading trains.

If the jobs people use robots for are split up into a diverse toolset, each of which comes with a set of puzzles to solve, that's a healthier game than one where there's the belt and inserter puzzle, the train and inserter puzzle, and the instant problem solving item teleporter not-puzzle.

[Gentaei]

Comparing Bots and Belts:

Bots:

pro:
- infinte throuput per tile since bots overlap
- allows very small and organised builds
- very fast on short to mid distances
- no delay by item pickup/drop from chest
- can stack the maximum amount of beacons per assembler
- only needs an high enough request in the chest to work fine
- researchable with Tier 3 scince (blue)
- cheap research for the first speed levels (+75%), also low tech need (T3 again)
- endless research for bot speed
- quite cheap to make bots, roboports and logistc chests

con:
- take energy to operate
- need recharging
- logisticsystem only with T5 scince


Belts:

pro:
- steady throuput independent on lenght
- no need for energy to operate

con:
- compressing a belt to 100% throuput can be tideous, also inserters cant compress without overfilling a splitter
- limited throuput per tile (13.33, 26.66, 40 per second)
- placing itmes on the belt needs time and can only be done on one line (half a belt)
- pickup from belt needs time (compare chest with blue belt, the diffrence is huge, especially with stackinserters)
- needs more space and balancing a XX line can be difficult
- blue belt needs T4 scince (purple)
- cannot reach high speed beacon coverage due to the space requirement and throuput cap


Nerfing Bots is basicly working against their deisgn. More energy? No Problem! Build more Power! Cant carry that much? No Problem! Build more!
By Design they scale very much with little effort, almost exponentaially, while belts scale linear. Also Bots support the 1 million speed beacons build while belts can't. (you can clearly see it in the given Pick of FFF, someone blueprint the base and tell the ratio of beacons/speed module to assembler)
If you want to nerf Bots you could increase the needed amount of research for the speed boost, give it more levels with smaller increments maybe? This should help flattening the midgame a bit. Its also easy to swap to bots quite fast and abandon belts, without even touching blue belts much. You could also increase the amount of resurces to make logistic bots since they are quite cheap, maybe use blue circuits for them?

You should also consider buffing belts in lategame. You simply face the throuput lock of 40/s on a blue belt, with all the limitations of inserters and splitters. Adding a T4 Belt with much much higher throughput would be a good step. You can fiddle 2 blue belts throu a single line with belt weaving of the undergrounds, but its no big use for crafting.
Adding a "two handed" Inserter that can place its items on both sides of the belt would help a ton with the loading of belts to get more througput. You could even top it of by giving it the priority above items on the belt.

Also consider a cap for the beacons bonus speed. thats probably the most gamebalancing thing. a t3 assembler with craftingspeed of 8 and 40% bonus prod is hefty. Try feed it with 100% by belts in the right ratios without throuput clogs. It is possible yes, but needs more time for planning.

Bots simply have the upper hand by using less space with high prod speed. These builds are also very compact. (good choice for the pick in the FFF ) You can build a megabase without speed beacons and bots, but its way more timeconusimg and tideous with belts. Since we all are lazy we simply take the easiest way to get the job done, and bots are the easy way.


Personally i think they are fine, they work better with the given tools (beacons) and as long as they do they will be the choice to go. How about you challange some of the Megabesebuilders to try out Belts and Trains only and see how it works?

[golfmiketango]

hahaha



http://lapsedhistorian.com/get-blower-l ... ing-tubes/

this is actually a good tech alternative that requires steam power and solves the problem of ammo, repair packs, and other small orders.

+1, brilliant idea.

[m44v]

Right now, many people play with the intent of creating a mega base.

The majority of players don't aim for that, is just that mega base builders are overrepresented in the forums and Internet in general.

To achieve a mega base, there is only one viable way to do it, and that is with bots. People have literally spend hundreds of hours creating their mega bases. I even know one person who is over 1000 hours on his save. Nerfing bots means that these people will probably have to start over, or at the very least, massively overhaul their creations. You don’t understand it when these people say “please don’t do this”?

Factorio is early access, those players knowingly signed up for this when they bought a game that is in a state of development. Building a mega base is also a player imposed objective, not by the game. If building megabases is detrimental to game mechanics because simplifying logistics is a requirement, then it might not be in the developer's vision for Factorio.

[Weresmilodon]

Well, that's an... explosive topic, isn't it? 25 pages in a day? Not surprising, really.

From my perspective, it's not a question of if bots are too strong. The Bots vs. Belts comparison will always indicate that bots are better, simply because belts are just so limited in comparison. Bots don't need nerfing, belts needs buffing. Not just faster belts either, we need a bulk-transport belt that can compare to the throughput of a bot network over long distance, making the belt a good option. Also, you might want to seriously think about the Loaders again.

Obviously, nerfing the belts throughput by reducing compression is in no way benefiting belts either. But i don't need to tell you that.