Factorio Forums

I wasn't around for loaders, considering the changes that went over the game in the meantime, would they still be considered "OP"?
What about them was "too strong"?
Producing a compressed belt with just 1 entity doesn't seem all that game changing to me just like the splitter filter and priority settings that people hailed as "maybe too op" didn't change all too much in the grand scheme of things.

Would loaders be an option now, bringing them back to give belts a new tool in the light of the recent "bots vs belts" debate?

welcome to the forums!

why not enable the recipe via console, and decide for yourself?

impetus maximus wrote:welcome to the forums!

why not enable the recipe via console, and decide for yourself?

Code: Select all

/c game.player.force.recipes["loader"].enabled = true
/c game.player.force.recipes["fast-loader"].enabled = true
/c game.player.force.recipes["express-loader"].enabled = true

If those are the same things that are in the creative mode mod, they seem to be broken and do nothing.
There's a filter setting and it doesn't do anything.

I thought they were chests that auto-spill their content on a belt?

I've been using belt loaders to test beaconed setups, just make sure that the "belt" side is pointing at the belt, and the other side is pointing at the chest! To load into a chest you need to put it down the wrong way, then rotate it (like an Underground Belt).

Personally the problem for me isn't them being "OP", it's that they are yet another tool in an already cluttered pile of tools. I never use Burner Inserters, and the first two tiers of Furnaces are stepping stones only, come midgame they all just clutter the crafting window. It's also more stuff for new players to try to get to grips with, I mean we already have 7 types of Inserter plus three types of belt in three different tiers, adding to that doesn't seem like a good idea.

Loaders would almost entirely replace Stack Inserters, which can move 12 items 2.31 times per second or just under 28 items per second (when moving between two inventories). An Express Loader shifts 40 items per second to or from a belt, for which Stack Inserters are even slower. Stack Inserters would become another obscure tool most players don't use, but you can't remove them because they're needed for high-end Speed Beacon setups.

I don't know about the feasibility of it, but IMO Stack Inserters should be able to grab and drop from/to both sides of a belt simultaneously. This would wildly improve their practicality with belts, but ideally we would want two Stack Inserters at full research to be able to compress a blue belt. 12/40 = 0.3 seconds to drop all items using both lanes, however it takes 0.433 seconds to make a full turn so 0.733 seconds to drop 12 items is about 16.37 items per second. To compress a Blue Belt between two of them, that needs to be 20 so either stack size or rotation speed would need increasing (or both) but this would also affect chests... Still, I think it would be worth it. assuming speed remains as-is, stack size would need to be increased to 18 I think. Hmm.

With what we have currently you need about 3.5 Stack Inserters to compress a belt. This is most painful when unloading from trains, you need a lot of infrastructure to then transfer all those chests to belts - it takes ~14 chests to make a typical 4-belt bus (if split correctly - easier to use 16) and twice as many Stack Inserters. Resources often aren't the issue but splitting such a large number of belts can be a pain - most reliable way is to use 16 chests, split each pair then sideload (so you have 8 lane-balanced belts), split those 8 down to 4 then use a standard 4-4 bus balancer. Whew! The belt infrastructure ends up being bigger than the station itself!

Deadly-Bagel wrote:I've been using belt loaders to test beaconed setups, just make sure that the "belt" side is pointing at the belt, and the other side is pointing at the chest! To load into a chest you need to put it down the wrong way, then rotate it (like an Underground Belt).

Personally the problem for me isn't them being "OP", it's that they are yet another tool in an already cluttered pile of tools. I never use Burner Inserters, and the first two tiers of Furnaces are stepping stones only, come midgame they all just clutter the crafting window. It's also more stuff for new players to try to get to grips with, I mean we already have 7 types of Inserter plus three types of belt in three different tiers, adding to that doesn't seem like a good idea.

I see, I thought they were chests with a hole that stuff comes out of but they're much more clunky to use.
So if they were themselves chests with an input and an output without all that 3 tier nonsense, they'd be useful, right?
Stack inserters loading from a train into one of those things wouldn't have the problem of wasting time waiting you were talking about and would also auto-balance both lanes of the belt so the train can't not get unloaded unevenly.
Ore running directly into the chest without the need for another row of inserters would simplify train loading stations.

Trains are where the biggest belt clusterfuck tends to happen which is why so many people use bots for unloading so this would be a good thing, right?

Trains are where the biggest belt clusterfuck tends to happen

Yap, and I'm tired of it. Im using the belt buffer mod to unload trains, and only for that.
It makes it a bit too easy. I'd like to see a litte nerf to it. But I really hate the ugly mess that happens without it.

Train stations are a perfect place to use blueprints. Build the station once, copy it, and use it over again.

Loaders are basically "inserters but faster". They aren't very essential to the game. A loader with an inventory might be neat, but it's not very different than using a chest and inserters in any configuration you like.

@Bagel:
some 12chest-4 belt unload ideas. viewtopic.php?f=194&t=58728

bobucles wrote:Train stations are a perfect place to use blueprints. Build the station once, copy it, and use it over again.

Loaders are basically "inserters but faster". They aren't very essential to the game. A loader with an inventory might be neat, but it's not very different than using a chest and inserters in any configuration you like.

Yeah apart from making things slightly more compact, it won't really change much, that's true.
I was just wondering what could be done to give belts a boost and loaders came to mind which I thought had been removed for being "too powerful" but I guess it was more along the lines of "too clunky".
Faster belts or stacked belts seem to be the answer.

Maybe if splitters could compress 2 belts into 1 by stacking the items, that would do. Or something.
At least it wouldn't require another tool.

disentius wrote:@Bagel:
some 12chest-4 belt unload ideas. viewtopic.php?f=194&t=58728

In all the suggestions (at least those in the first few posts) there are two chests that will unload twice as fast as the others, as such the chests will run out in half the time and belt throughput will be halved. I don't consider that to be a complete solution, I'd rather run through 6-4 balancers (a 1-2 train can do this by unloading only on one side).

My current game involves 15 belts of each Iron and Copper Ores, I've got two stations for 4-wagon trains with 10 chests per wagon so 80 chests. I think it was then sideloading to 40 belts, running each through 4-3 balancers (30 belts) then merging them 2-1 for my required 15.

dood wrote:I wasn't around for loaders, considering the changes that went over the game in the meantime, would they still be considered "OP"?
What about them was "too strong"?
Producing a compressed belt with just 1 entity doesn't seem all that game changing to me just like the splitter filter and priority settings that people hailed as "maybe too op" didn't change all too much in the grand scheme of things.

Would loaders be an option now, bringing them back to give belts a new tool in the light of the recent "bots vs belts" debate?

I love the loaders but they are very powerful.

1) they can transfere a full belt replacing a multitude of inserters
2) they can filter at full belt speed
3) much simpler to build than a multitude of inserters
4) they fill an assembler or furnace up to the stack size instead of just 2 cycles worth of inputs, meaning the assembler produces a stacks worth of output ahead of time
5) no need to load a buffer chests from belts with multiple inserters and then stack inserters to fill a fast assembler to keep up with demand, directly feed a full belt into the assembler

Personally I use Deadlocks loader mod but the cost of loaders might be to low for the power they have.

How often do you use Stack Inserters with that mod mrvn?

Deadly-Bagel wrote:How often do you use Stack Inserters with that mod mrvn?

For every train station. Because loaders don't work on trains.

mrvn wrote:

dood wrote:I wasn't around for loaders, considering the changes that went over the game in the meantime, would they still be considered "OP"?
What about them was "too strong"?
Producing a compressed belt with just 1 entity doesn't seem all that game changing to me just like the splitter filter and priority settings that people hailed as "maybe too op" didn't change all too much in the grand scheme of things.

Would loaders be an option now, bringing them back to give belts a new tool in the light of the recent "bots vs belts" debate?

I love the loaders but they are very powerful.

1) they can transfere a full belt replacing a multitude of inserters
2) they can filter at full belt speed
3) much simpler to build than a multitude of inserters
4) they fill an assembler or furnace up to the stack size instead of just 2 cycles worth of inputs, meaning the assembler produces a stacks worth of output ahead of time
5) no need to load a buffer chests from belts with multiple inserters and then stack inserters to fill a fast assembler to keep up with demand, directly feed a full belt into the assembler

Personally I use Deadlocks loader mod but the cost of loaders might be to low for the power they have.

But are loaders more powerful than their competition? Which are bots. Loaders are still belt-based technology, so even at 100% efficiency, they would be weaker than bots in theory.

https://www.factorio.com/blog/post/fff-225

With this in mind, we can safely say, that robots are at least 2 times stronger then express belts, but in real factories, it is much more as belts need lot of other parts and are rarely used as ideally as robots would be, so my private guess would be that robots are currently around 5+ times stronger compared to belts.

The main advantage of belts is that its easier to balance a train-station. Bots grab from the nearest provider chest, which means you need active-provider chests + backpressure sensors on your train stations if you want decent "balancing" compared to a belt-based unloader. But the size and space savings of bot-based train stations are pretty obvious to anyone who has tried it. I don't think people were making mega-bot sized megabases when loaders were first considered to be OP.

A common request is to add loaders back into the game, but make them use a sizable amount of electricity to use. Maybe only add express loaders in Logistics 3 (which is locked to yellow science anyway), so you can give belt-users another buff compared to bots.

In any case, its clear that the devs are looking at the "most fun way" to buff belts. Maybe the "ideal base" should use all resources available: maybe bots really are the "best last mile", but belts can be used (with pallets / containers) in some designs. That way, the hybrid belts + bots wins over at the end. If inserters were working on pallets of items at a time (each pallet == 1 stack), kind of like barrels / unbarreling, then belts would be the best mid-range entity (assuming bots were unable to move pallets).

I really like that pallet idea in that FFF post. It seems like the most fun way to solve the problem of bots vs belts. And everyone wins at the end of that. And really, that's the core issue at hand with this Loader discussion, right? (Because enabling of Loaders is simply another "belt buff" to be considered in the future).

dragontamer5788 wrote:I really like that pallet idea in that FFF post. It seems like the most fun way to solve the problem of bots vs belts. And everyone wins at the end of that. And really, that's the core issue at hand with this Loader discussion, right? (Because enabling of Loaders is simply another "belt buff" to be considered in the future).

Yeah, in this type of game, the only limiting factor is game performance so there is no reason not to buff the other options up to a similar level of a perceived stronger one.

So far, the proposed ways to "nerf" bots, say, by limiting carry capacity to 1 would only serve to hit game performance in the knee because people would just end up with more entities flying around to get the same result.
Same as with the rail tanker nerf that cut its capacity down to a third to create parity with tanks, all that did was more train trips which only served to hit the game performance. Simply tripling tank capacity for parity would have done but alas.

That kind of performance destructive way of balancing things in a game where UPS is precious seems really backwards to me, so something that not only gives people a new tool to play with but also cuts the number of moving parts down would be ideal.

dragontamer5788 wrote: But are loaders more powerful than their competition? Which are bots. Loaders are still belt-based technology, so even at 100% efficiency, they would be weaker than bots in theory.

https://www.factorio.com/blog/post/fff-225

Lets see:

- loaders don't consume tons of power for recharge
- loaders don't carry items all the way across the base just to turn around 1m from the goal to head back for a recharge
- loaders don't disappear somewhere else because an iron plate must be moved way across there and the local bot is busy
- loaders don't all bunch up in one place waiting in line for a recharge with 10000 other recharge points available elsewhere
- loaders don't stray over alien territory getting shot down and loosing the 4 nuclear reactors they carry

Overall bots just plain suck for any volume transfere. While a billion bots can take stuff out of a chest if you figure in the recharge time they are a lot slower than a belt. Or rather their speed correlates with the number of rechrage points you can make them use in parallel. Unless you go insane with recharge ports all over they are actually quite slow.

And given their random nature you never know where the bot will go next. They are unreliable for a constant sustained throughput.

Imho bots can only reasonably replace belts in small networks. Which means lots of little satelite factories that have their own logistics network and a big cluster of recharge points close together. Or for low volume items where the time till arrival doesn't matter.

On top of all that: bots can't put items into an assembler or furnace. So you still need the inserter or loader for the final transfere. Bots are in no way an alternative to loaders. Inserters are. And they clearly beat inserters.

mrvn wrote:

dragontamer5788 wrote: But are loaders more powerful than their competition? Which are bots. Loaders are still belt-based technology, so even at 100% efficiency, they would be weaker than bots in theory.

https://www.factorio.com/blog/post/fff-225

Lets see:

- loaders don't consume tons of power for recharge
- loaders don't carry items all the way across the base just to turn around 1m from the goal to head back for a recharge
- loaders don't disappear somewhere else because an iron plate must be moved way across there and the local bot is busy
- loaders don't all bunch up in one place waiting in line for a recharge with 10000 other recharge points available elsewhere
- loaders don't stray over alien territory getting shot down and loosing the 4 nuclear reactors they carry

So you don't build enough Roboports and you don't have enough bots. Its pretty simple really: instead of building 4-lanes of blue belts going somewhere, build a lane of roboports to provide ample charging space. That solves almost all issues revolving around roboports, the last is a minor layout concern. (Minor in comparison to learning how to use balancers, splitters, filters, and the gross amount of space blue belts take up in endgame builds).

Robots can "stack" infinitely and have no collision detection. They're UPS friendly, and in developer-sponsored tests, they perform 2x better than belts (yes, even with loaders AND the "have to carry themselves back" issue).

Robots need chargers. Each roboport provides 4-charging ports. A solid line of roboports provides the maximum amount of charging in a given direction. Despite being the strongest strategy in the game, robots still require a bit of brain power to use correctly, which is why I'm personally fine with it. But the main issue is that belts (even with loaders) require MORE brain power to use correctly, and yet only offer 1/2 the bandwidth of any proper bot-based build.

Proper bot builds are incredibly easy to design, blueprint, and copy paste. This bot build is 5000 science-per-minute (5-rocket launches + all the infrastructure to support that science going at once). Yes, its a lot of power, but the simplicity and regularity (especially useful for blueprints) is outstanding.

To accomplish the same with belts would require 236 parallel iron plate blue belts, and 123 copper belts. That's 236 perfectly compressed, perfectly balanced blue belts of iron.

Have you ever made a 128-belt balancer? Its way harder than just slapping down more roboports and inserting more robots.

But hey, prove me wrong. There are a TON of 5kspm bases out there using bots + trains, no belts at all. If you think belts are superior, build a 5kspm base and report back to me. You'd be the first one to do it.

Here's a great overview of the size that these bases require: https://imgur.com/a/7aAMy.

--------------

Alternatively, maybe you can just replicate a single outpost. Which would be way easier.

https://i.imgur.com/zAxhRct.jpg

This bot-based mining area creates 677 gears per second. That's 17 parallel perfectly compressed perfectly balanced blue belts of gear output, fed by 967 iron plates/second (25 blue belts), fed by 805 iron ore/second (20 iron ore blue belts).

That means those bots are performing as if they were 20 + 25 + 17 parallel lanes of blue belts (assuming perfect compression and perfect balancing). Feel like you can accomplish that with belts? I'd love to be proven wrong here.

dragontamer5788 wrote:

mrvn wrote:

dragontamer5788 wrote: But are loaders more powerful than their competition? Which are bots. Loaders are still belt-based technology, so even at 100% efficiency, they would be weaker than bots in theory.

https://www.factorio.com/blog/post/fff-225

Lets see:

- loaders don't consume tons of power for recharge
- loaders don't carry items all the way across the base just to turn around 1m from the goal to head back for a recharge
- loaders don't disappear somewhere else because an iron plate must be moved way across there and the local bot is busy
- loaders don't all bunch up in one place waiting in line for a recharge with 10000 other recharge points available elsewhere
- loaders don't stray over alien territory getting shot down and loosing the 4 nuclear reactors they carry

So you don't build enough Roboports and you don't have enough bots. Its pretty simple really: instead of building 4-lanes of blue belts going somewhere, build a lane of roboports to provide ample charging space. That solves almost all issues revolving around roboports, the last is a minor layout concern. (Minor in comparison to learning how to use balancers, splitters, filters, and the gross amount of space blue belts take up in endgame builds).

Robots can "stack" infinitely and have no collision detection. They're UPS friendly, and in developer-sponsored tests, they perform 2x better than belts (yes, even with loaders AND the "have to carry themselves back" issue).

Robots need chargers. Each roboport provides 4-charging ports. A solid line of roboports provides the maximum amount of charging in a given direction. Despite being the strongest strategy in the game, robots still require a bit of brain power to use correctly, which is why I'm personally fine with it. But the main issue is that belts (even with loaders) require MORE brain power to use correctly, and yet only offer 1/2 the bandwidth of any proper bot-based build.

Proper bot builds are incredibly easy to design, blueprint, and copy paste. This bot build is 5000 science-per-minute (5-rocket launches + all the infrastructure to support that science going at once). Yes, its a lot of power, but the simplicity and regularity (especially useful for blueprints) is outstanding.

To accomplish the same with belts would require 236 parallel iron plate blue belts, and 123 copper belts. That's 236 perfectly compressed, perfectly balanced blue belts of iron.

Have you ever made a 128-belt balancer? Its way harder than just slapping down more roboports and inserting more robots.

But hey, prove me wrong. There are a TON of 5kspm bases out there using bots + trains, no belts at all. If you think belts are superior, build a 5kspm base and report back to me. You'd be the first one to do it.

Why would I even try to balance 128-belts of iron ore? I simply dump 128 compressed iron ore belts into lines of furnaces to get balanced belts of iron plates. Even if they somehow get unbalanced then one line of furnaces will take a little pause. So what? In the end they produce a few belts of green circuts or red circuits. Easy enough to balance them at the end and ignore the 0.0001% loss.

One great thing about belts and the new splitters in 0.16 is that you can filter and prioritize. Something you can not do with bots at all. There is no way to say 75% of the iron plates should go to that requester chest and only 25% to the other one.

Note: Even if belts are superior that doesn't mean that they are more UPS friendly. A 5kspm base in 0.15 with belts kills your UPS and I would say that is the first reason why nobody does it. Heavily beaconed setups are another. Belts take space and with bots you can probably squeeze in another beacon or two and that is worth more. All of that (including building a solid line of roboports) is so far past building your first rocket that is doesn't compare to normal gameplay.

mrvn wrote:All of that (including building a solid line of roboports) is so far past building your first rocket that is doesn't compare to normal gameplay.

Even blue belts are beyond your first rocket, because 31.5 iron per chunk of belt is way too expensive compared to 11.5 iron per red belt chunk. (2x red belts give you more throughput than 1x blue belt, and 1x blue belt costs on the order of 3x red belts). So loaders, blue belts, and all of this discussion is implicitly about the endgame and abilities above and beyond your first rocket.

Ultimately, bots scale. They are the ultimate endgame strategy, at the moment. This has been proven again, and again, and again. Fact is, you cannot build a belt-solution as simple or as easy as these copy/paste bot strategies. The fact that bots have superior throughput to boot is just icing on top.

Why would I even try to balance 128-belts of iron ore? I simply dump 128 compressed iron ore belts into lines of furnaces to get balanced belts of iron plates. Even if they somehow get unbalanced then one line of furnaces will take a little pause. So what? In the end they produce a few belts of green circuts or red circuits. Easy enough to balance them at the end and ignore the 0.0001% loss.

If you like loading trains evenly, you'll want a balancer. That delay on belt#35 will cause your trains to be delayed. Bots on the other hand, have so much "burst" throughput that they can deliver everything from your buffer-chests to the train-station no problem, and don't need a balancer.

I edited in a simpler challenge, since building 5kspm does take a lot of time. How about a single facility that processes 677 gears per second? By my calculations, it will require a minimum of 20+25+17 gear / plate / ore parallel belts. Just really, think about it. How would you build such a facility? How would you weave belts between those beacons? How will you deal with inserter slowdowns (belt->inserter is way slower than chest->inserter)?? How do you "balance" the output between miners? Belts under normal circumstances would probably be backed up and definitely can't handle productivity-boosted speed-beaconed-miners.

I mean, maybe you do have an idea. Maybe belts really are superior to bots. If so, you'd be able to create a demonstration, yes? Frankly, I'm not seeing how belts could possibly ever compare to this build.

One great thing about belts and the new splitters in 0.16 is that you can filter and prioritize. Something you can not do with bots at all. There is no way to say 75% of the iron plates should go to that requester chest and only 25% to the other one.

Seems like a job for trains if I ever saw one. But I'm not entirely sure what the point would be?

At the small level, requester chests have much finer grained control. This requester chest gets 200-ore buffered. That requester chest gets 50-ore buffered. click-click, boom. Done. Bots handle the rest (assuming you have enough bots and roboport charging stations). On a broader scale (if you REALLY needed a base-wide percentage split), you can handle that by managing train deliveries towards sub-bases. (This sub-base gets 3 train deliveries, that base gets 1 train delivery).

Prioritization is as simple as using active provider chests instead of passive provider chests. Bots prioritize active chests above all other tasks, and you can use that to clear out high-priority areas (ie: train stations).

Its a foregone conclusion that belts need a buff in the endgame. That's why we're talking about loaders, pallets, stack-belts and other such things to rebalance the game towards belts. Players spend more brain power on belt solutions, so they SHOULD be rewarded for their efforts. Loaders would make building balanced and compact belts way easier, AND they're already programmed in the game (no developer effort! It'd be just a config-change to upgrade them to "default" settings).

But I think "pallets" would be the best belt-buff for this game. Keep bots as the "last-mile" superiority as they are, but at least have belts be the undisputed king of mid-range distribution. Unless the devs feel like fixing the "Chest->Inserter" vs "Belt->Inserter" slowdown issue, bots->chest->inserter->Assembly machine will always be the best "last-mile" infrastructure.

Yeah, bots are better at stupidlarge structures than belts. Always have been.
That much is hard to dispute. They make things much easier to rearrange and expand.

My only point of contention about the comparison is people slapping a fixed throughput number on bots -whose entire thing is a malleable, dynamic throughput- in order to compare them with belts.
That oversimplifies things a bit too much imo because it is exactly that quality of dynamically applying their throughput to an arbitrarily sized network that makes them so great to begin with.