Friday Facts #224 - Bots versus belts

[ske]

Giving is easier than taking away. You are the ones with power to take. Mod developers usually give. The base game should feel tight but right, just wide enough to give a good experience but to put up a fight to get through all the way. Infinite resources and infinite capabilities in the base game don't sit well with me. Let the mod developers unlock those features. Neutering the bots will surely be met with harsh criticism. Not everybody will use mods now. But you might be able to do it in steps. E.G. you have a configuration checkbox for "supreme bots: enabled". Enable this by default at first. Nobody will notice. Then create a new game-mode "proper factorio" as a container where certain flags will be locked and "sandbox" where the flags are unlocked. This mode will be the mode that new players use to play it through in the release version. It will pose new challenges purely because of the restrictions. But for most players it will be more fun. Until that mode is developed well and liked, it will be optional. After that point it will be the default game mode and it will be required to obtain achievements and many servers will use it.

[technogen]

My vote is to remove the bots, but add a big (even huge) new feature to the belts that I had in mind, which will simultaneously make belts exponentially more fun and occupy the niche of bots (the one that involves curing chronic spaghetti belt syndrome).
I'm talking about realistic underground belts.
In a nutshell, I'd like to propose replacing the simple underground belt mechanic with a far more elaborate one, where players can dig underground, which is essentially the process of building a structure that teleports the player to a new "underground" surface. The new underground belts will be merely an entrance/exit that allows one to connect belts between aboveground and underground. The player will then need to manually lay out the belts in the underground surface to connect the entrance and the exit. This could be done recursively, where the player would be able to dig one level deeper. This essentially gives "infinite space" between any two points on any of the surfaces, which is exactly what bots provide, except this time, the player still has to manage the belt logistics to make it happen. The big bonus is in the incentive to build in an optimized an compact way, because unlike aboveground, the underground surfaces are full of rock, which needs to be cleared away to make room for belts.
Beyond solving the "bots versus belts" problem, underground layers have huge potential for amazing additions in the future, including more realistic ore generation, underground digging biters, seismic stations to combat them.

This is a tall order, but it's can easily double the amount of fun from building factories.

Regards,
technogen

[Yehn]

The topic of bots has been something I've been modding to adjust for a bit, but I think bots are just one facet of the bigger issue of "free/resourceless/infinite" power. The argument goes like this:

<good arguments>

I therefore suggest the real issue here is free infinite power through solar, that allows one-time cost to mitigate a whole aspect of Factorio gameplay.

I agree, but in practice nuclear is just about infinite, too. The reason to use solar over nuclear is UPS, not continuing cost. You have to expand your power when you expand your factory, but running out of uranium is about as urgent a concern as running out of sunlight. Unless you stay on coal/oil power, Factorio isn't really structured to have electricity "cost" anything.

As a random side thought, I've always found it odd that reaching oil isn't more of a milestone for power -- in terms of fueled power, that is, batteries are obviously a big deal. As some mods do, there was a chance here to have more powerful generators fueled from liquid products. But not completely OP like solar and less powerful than nuclear. Instead... all we get are fuel blocks to feed into the same old steam boilers.

[IronCartographer]

Kovarex_Box_2.png

That's pretty funny, but logistic bots aren't evil!

There's joy to be had in seeing the swarm descend on a train, resupplying your factory while...guzzling electricity...

Wait a second. Belts don't consume power... Nerf belts! Make offshore pumps take hand-cranking by pressing R and then supplying electricity thereafter!

[Kaufcraft]

what if bots didn't fly and/or required some sort of tracks? I suppose making them not fly would be super hard because of collision and all that, but maybe just requiring to build some tracks and/or other type of space demanding infrastructure for them to run along? i think something that occupies space could put enough pressure on designs to the point where it might become more viable to belt things in

I would like to develop on this idea. Lets restrict free movement of logistic robots between to points that this is only possible within the area of one roboport. If an item is to be transported within such an area the robot can fly to the item source and bring it to the item sink straight away. But if source and sink are in different roboport areas the robot has to fly to the source and than to the nearest roboport. From there it has to fly to the roboport in which area the item sink is. And than from there it can fly straight away to the sink. And to bet from the source roboport to the target roboport it has to fly to each roboport in the way. And then between the connection lines of the roboports there should be no obstacle that prevents the line of sight between such that buildings. Some things like rails and belts should be allowed. This would force the player to put more sophistication into the logistic network and maybe to build more roboports than would be necessary at the moment to bypass obstacles.

Alternatively one could consider a limitation of how many robots can deliver or get items to/from chests per second.

[MeduSalem]

The throughput per area seems like something that could be tweaked by adjusting how often bots have to charge and how many bots can charge per roboport. Seems you could push the numbers in such a way that the space required for the roboports would become problematic. I already have this issue today anytime I've tried doing train offloading with bots, after awhile I end up with all the bots charging because I cannot stuff enough roboports into the space.

I tend to use my Chunk-aligned Module System I developed:

v3 Grid Chunk Borders.jpg (433.41 KiB) Viewed 9236 times

2x2 area

v3 Super Module 2x2.jpg (413.38 KiB) Viewed 9230 times

I can stuff as many Roboports in the hazard concrete area as I want... in any spacing I desire to deal with the bot density. So any change to the charging behaviour/bot density can be outsmarted.

[bobingabout]

"I don't like logistic bots, so I've convinced the rest of the dev team to remove them from the game."

I hate you.

You've already locked logistic robots, or their accompanying logistic chests behind higher and higher tier science packs, surely you've already gone far enough.


Either way, unless you make it impossible to do so, I'll still just mod them back in.

And a 3rd type of Robot (Does some jobs of construction, and some jobs of Logistics) doesn't sound like a bad thing. Perhaps merge them into a single type with flags such as...

Code: Select all

can_repair = true
can_construct = true
can_provide_requester_chests = true
can_provide_player = true
can_attack = true

Basically, if the first 2 are true, it's a construction robot, if the second 2 are true, it's a logistic robot. construct and provide player sounds like a good starter robot, and a decent mid-way point between construction and logistic.
And yes, add an attack zone to the logistics network, and when enemies enter this zone, attack robots go and shoot at them. turning on that tag will obviously require you to specify a weapon for the robot, and should be able to be configured to require ammo(Like most weapons used by a player, or vehicle, and the gun turret), or energy(Like laser turrets, and personal defences). and although I don't think it's needed... just shoot (Like combat robots, or worm turrets)

[Encrypted]

Hi just my 2 cents.

I think that you have already done good work by making logic bots easy to get but the requester chest research expensive. So that personal logistics is earlier then base wide. I also think bots are too powerful.

For this I can think of a few things to make it more difficult.

• add space science to requester chest so it moves that to late late game
• dont let the bots fly. Make them ground bound and have to have brick or concrete to drive on
• add collisions to bots so they cannot swarm. Or limit same time access to chests so only 1 bot at a time can access it

There are of course more options but these are the ones I like

[Jiehong]

Keep bots, but make belts being able to carry containers of stuff?

Pipes are like belts for liquids. Fluid barrels are a way to move pieces of fluids around.

Similarly, we could have a belt version of each belt (behind 3 different research) to allow belts to transport containers of stuff as well.

A given container could carry the equivalent of the smallest chest worth of things. That would greatly increase the transport density of belts by a factor 100 for example.

[Therax]

You can count me firmly in the anti-bot camp, and the answer for my own personal play-throughs is simple: I don't research Logistic Networks. However, that doesn't mean I think a nerf is either warranted nor would it be effective.

In terms of game balance changes, I'd advocate for leaving bots as is, but significantly buffing belts in the late-game. I'd suggest:

1) making belts have predetermined item "slots" such that inserters can transparently create fully compressed belts.

2) Making a fully compressed belt tile produce consistent results from tick-to-tick when read to a circuit network.

3) Increase the compression on belts, by allowing items to be placed closer together. This can also be something that improves with technology as a player progresses through the game. (This may be related to the mentioned "stacked" belt tier.)

4) Consider increasing beacon range and decreasing beacon effectiveness. One of the major downsides of belts in the late game is how difficult it is to get multiple belts of throughput into a beaconed assembler line due to there only being room for one belt on each side. Belt weaving techniques mitigate this to a degree, but that's really an unintuitive and advanced technique.

5) Perhaps adding a tutorial on belt weaving might be in order?

[purdueme91]

I agree with a lot of the posters that if you think robots make it too easy then don't build them. I am on my first 0.16 Vanilla playthrough and I am setting up my last science pack and still haven't set up any kind of logistic network. I love bots but I am making it through at a good pace without them.

That being said, there are other mods out there that really really need to have bots. Try to complete a Bob's playthrough without bots. The belt spaghetti would be un-freaking-believable (I tried once. Won't do it again). Or Angel's or SpaceX or Seablock. It would make some of the most popular mod packs unplayable.

My idea to preserve them but making them less powerful is following through the way it is in Bob's but even more stringent. Different tiers of robots (4-6). Each robot tier has improved cargo carrying and speed (remove at least some of the speed and cargo upgrade research tiers). To build next tier, you need to use previous tier robot. Each tier is locked behind its own science level. Remove the infinite robot speed research. Maybe you add a speed module to upgrade a robot. Not sure if would go to needing the brain and the arm like in bobs but have an improved frame that needs to be added to the previous robot.

Now it's no longer trivial to just research something and BAM your logistic system can carry twice as much stuff. Seems more realistic too to have to physically upgrade a bot for it to go faster and carry more stuff.

[IronCartographer]

You've already locked logistic robots, or their accompanying logistic chests behind higher and higher tier science packs, surely you've already gone far enough.

Easy there tiger. Logistic bots are here to stay.

For what it's worth, I fully agree with the quoted text.

The devs need to remember that the playerbase isn't entirely composed of super-elite optimization maniacs. Bots are satisfying in their own way, and can be avoided or modded out if desired.

Heck, a lot of us have never built on a large scale, and spend most of our time watching the community rather than playing the game!

[Simanova]

Dont punish bot users. Lets try some new ways.

Here is my idea of solving this problem:

Create crawling bots like spiders on the ground. The bots should have the same core, but different tools.
Some of them are workers, other bots are defenders. Every bot is limited to carry one item, regardsless of its use.
Upgrade the late game experience with hive mastery.

This would evolve the game. Replace capsule throwing.
A lot of players would use belts for transport, and bots for building, defending, and minor logistics.

Something like this.

[Caine]

4) Consider increasing beacon range and decreasing beacon effectiveness. One of the major downsides of belts in the late game is how difficult it is to get multiple belts of throughput into a beaconed assembler line due to there only being room for one belt on each side. Belt weaving techniques mitigate this to a degree, but that's really an unintuitive and advanced technique.

Without a limit of how many beacons can influence a factory people would just place more beacons and still move items in by bots.

Perhaps bots should be moved into an officially supported and included mod. That way players will need to toggle it on to keep existing functionality but vanilla provides the intended experience.

[Bokrug]

The problem I have with this is the assumption that belts are more fun than bots. Different people find different things fun. I think a solution would be to make belts more attractive for later game. I believe others have suggested it already, but diminishing returns for beacons. It would mean builds would effectively have more space to fit belts in. People try to maximize beacon effects and that pushed out any extra space that belts might fit in. Maybe that means tweaking modules to get the same effects with less beacons. Maybe a specialized beacon that acts as a mk 4 module? Forcing players to build larger because of a beacon nerf will just hurt ups and anger people.
Another part of this would be to make bots more expensive. Raising power requirements and material cost will mean it takes a lot longer to make a full switch from belts, and because costs will always become trivial it will have little effect on large bases. I think also reducing the stack size of bots might help. Ports have a limit to amount of bots they can hold, but I've never heard of it being an issue. With smaller stack sizes it would require more ports, which would help with the increased power demands anyway, while increased cost of ports would further help to reduce the speed of a full conversion.

[Armentia]

I've skimmed the crazy number of messages that have appeared since this FFFs posting so recently. Just wanted to put my two cents in as well.

I don't like the bots, I don't use them. Others like them, they use them. WORST CASE, you remove them and had an official mod for the game that keeps them in.

If you want to extend the logistics part of the game, add packaging. viewtopic.php?f=80&t=19343 is the conglomerate topic I found in the suggestions forum. Items into crates, crates onto pallets. This will (hopefully) help with the UPS issue of giant belt driven factories.

[Dranzer]

I feel like it's an issue of scalability.

For one, belts take up space. In a belt-based factory, If you want to double your bandwidth without overhauling your design, that can be incredibly difficult or costly. Each belt, each inserter, they all take up space. There are some hard limits, and later on things get complex. Overhauling a factory to account for increased demand can be more painful than interesting.

Bots, on the other hand, do not take up space. If you want to double the bandwidth of a logistics system, you dump more bots into the roboport. At worst, you have to build new roboports. Sure there are power costs, but power is a problem that can be completely separated because powerlines also don't really take up space. The scaling issue of logistics bots is primary one of distance, but guess what - logistics robots naturally allow far more dense factories anyway, so the problem they encounter is the very problem they're designed to solve!

The solution to all this, if you're really looking for one, is to take a look at how items move around belts, and realizing how limited some of those options are. Everything takes up space. If you really wanted to fix the problem, you'd need to look into things like single-tile configurable filter-splitters. I'm talking about a single tile object that can take belt input from any of its four directions and dispense - selectively - in any direction. I'm talking about inline belt filters and chest loaders. I'm talking about single-tile balancers and splitters. I'm talking about the ability to stagger items within a single lane. I'm talking about killing off inserters in the lategame because they create more problems than they solve.

Or I guess you could make magic teleport chests that run on wires like the circuit network, but that doesn't really solve the problem of belts being difficult to work with.

[bobingabout]

Dont punish bot users. Lets try some new ways.

Here is my idea of solving this problem:

Create crawling bots like spiders on the ground. The bots should have the same core, but different tools.
Some of them are workers, other bots are defenders. Every bot is limited to carry one item, regardsless of its use.
Upgrade the late game experience with hive mastery.

I must admit that the fact that robots can just fly point to point without having to avoid obsticles is part of what makes them so powerful, but the lack of a need for path finding is also part of what makes them viable, otherwise they'd slow the game to a crawl. The subject of bot path finding has come up before (Such as moving from port to port, and staying inside the logistics zone, rather than just flying in a stright line across a logistic void full of enemy bases)

[Anemos]

Hi guys. I just registered here so I would be able to write under this post.

Bots are great and people love them. It's a late game option that is executed nicely and releaves some of the grind of playing the game too often. It's a meaningful and unique feature that makes the game stand out both in gameplay and in character.

Giving the player late game power is not a problem, unless there is no good amount of challenge to be had. Being able to manage your factory so well frees up time for the player that needs a new type of sink that, in my mind, should be the late game threats themselves. Building quicker things should be answered by the game with the proper opposition, or even late game conditioning in a satisfactory manner.

In short, bots are amazing, it's the late game that somehow feels a little dry. It just needs one more thing, some final situation that will make this a true masterpiece. Could it be one more type of alien, an environmental challenge or a simple extra game mechanic, this could be brainstormed by the team. I have absolute faith in you guys doing the best for the game.

I love Factorio, I recommend it to everybody I talk with and this is my input to you.
Thanks guys and keep up the great work. You're making something really special.

[ske]

What can logistic bots do that belts can't?

* Move items between arbitrary places.
* Move many different kinds of items in a coordinated fashion.
* No restriction on the amount of items per time unit.

If you have enough bots and the energy supply, I don't know a single thing that would be worse than belts.

Making bots carry less or fly slower would only increase the number of required bots. Everything else would stay the same.

Making bots only carry items locally might be hard to understand and would only solve problems partially. Usually bots currently are mostly employed for local transportation.

The one thing that could change this easily is rate-limiting chests (and roboports). It even makes sense. Currently there are bots overlapping all the time. Rate-limiting only applies to charging after they already have done their work. Think about how it would be if the only one bot can land on a chest each time. It would be as slow as a yellow inserter. This would change at least some designs like train unloading. But with other designs you could simply add more chests to an assembly machine or smelter because they only need like 4 yellow inserters for input and output. Still, I think this would be the most straight-forward way.

How else could bots be made worse than belts in most situations while still keeping an advantage and adding to the gameplay in some distinct areas?