Belts vs Bots - A response to FFF #224

[BlakeMW]

If beacon range is four tiles then you can fit another beacon in. I have not tried any designs, but I suspect it will lead to beacon builds with less space instead of more.

I think for 4 radius to work well beacons would need to occupy 4x4 rather than 3x3. This could be considered a kind of fly in the ointment because beacon size would no longer be exactly the same as assembler size, but perhaps beacons deserve that because they are pretty darn powerful.

There would also be other ways to limit stacking beacon effects on assemblers: like outright capping speed to +400% just as efficiency is capped to -80%. That would put a hard limit on the usefulness of packed beacons.

[Zavian]

But even with 3x3 beacons, you could easily add a soft cap on beacon effectiveness by making each additional beacon less effective. eg the first beacon applies full effects, the second applies half effect, the third applies one third effect etc. If you feel that that is too big a nerf, remember that atm each beacon only applies half effect, so 4 beacons adds 200% speed, 6 beacons adds 300% and 8 beacons adds 400%. Under the above scaling 4 beacons adds 100% + 50% + 33.33% +25% = 208.33%, 6 beacons add 208.33% + 20% + 16.67% = 245% and 8 beacons is 274%. So the sweet spot is probably around 4 beacons, and there would be little incentive add more than one row of beacons. That automatically frees up where the other row of beacons would currently be for as many belts as desired.

Note that this would also mean that "normal" (non-beaconed) builds would show big benefits from even single beacon, assuming that one beacon can affect everything.

Edit: Played around with simulated 4x4 beaconed layouts (simulated with roboports) and I think for a row of 4x4 beacons, with range 4, each assembler is still affected by 4 beacons, but each beacon can now affect 5 assemblers. It's also worth pointing out that none of this addresses what I think is one of Twinsen's concerns, which is that increasing production by just stamping down more blueprints of assembler + requester + provider is a boring way to play, yet it's also the simplest and easiest.

[BlakeMW]

I also wouldn't mind soft-capping. In fact I'd make it a harder soft-cap: if a beacon is rated to serve 4 machines, and the beacon is servicing 8 machines, then the effect is plainly and simply halved (i.e. providing 25% instead of 50%), so for instance a speed3 beacon could never provide more than +200% speed total.

I think that between soft-capping beacons (or alternatively, assemblers) and just limiting how high module effects can stack, the limited module effects would be more likely to fly: after all, it's already implemented into the game (at least for lower bound) and it's a simple mechanism that is easily explained, and even the idea that say, an assembler can't run at faster than +200% speed actually seems pretty logical.

[m44v]

So now, in order to do the same task, you're forcing me to use circuits with my belts? What is with people wanting to push their subjective idea of fun onto others? So now I would have to construct an elaborate sushi train to perform the same task? Why is my subjective way of enjoying the game being punished?.

It's clear to me that nobody wants logistics bots removed, is disingenuous to argue that those that see a problem with bots wants to force bot players to play with belts. There's actually no agreed proposal, the most suggested so far:

• Change beacons.
• Cap bot's throughput by either adding collision or a chest access rate.
• Leave bots alone, add something that can compete and doesn't lead to a single trivial design.
• Add bot maintenance, so that keeping thousands of bots constantly working involves something more than just ploping more roboports.

Nothing like removing bots.

[Spatial]

It would be cool to add another stage in the evolution of belts. Something radically powerful with fairly high requirements, like a dedicated bus type of belt that needed power but had 8x throughput of a yellow belt, had unique appearance etc.

Another fun one would be sending items through pipes at huge speeds using pneumatic pressure from steam. Nuclear item delivery

[oracleofepirus]

There's at least one thing that we can do to limit bots without adding entities, a metric ton of programming (collision), changing other entities, or adding useless wait times.

Add a hard minimum distance between roboports.

Most new players aren't going to dump two roboports directly next to each other. It's entirely unmeaningful to a new player to not have maximum coverage. This is also exceedingly logical to anyone versed with flying. You need a ton of airspace for an airport to work, both for physical space and radio interference. There's a few airports in the world that are within 3 miles of each other, but none exactly next to each other.

This change affects every variable that a roboport can provide, without touching any other entity. I imagine that some of those values could even be buffed afterwards (I'm not a bot heavy player, but the number of charging ports per roboport is atrocious).

[js1]

As I wrote here viewtopic.php?f=38&t=56218&start=840#p333435, I think it's all about suspension of disbelief. I think bot megabases upset people (like me) because they're counter-intuitive; the optimal solution to production shouldn't be to airlift everything.

Add a hard minimum distance between roboports.

This is a good idea. Although it's not exactly clear it's gonna entirely solve the problem that bots have potentially higher throughput than belts.

[ActofTreason]

What if I told you they can? You can filter items using static filter inserters, and you can balance belts with splitters. I'm not doing anything that'd be impossible without circuits; from a purely functional standpoint, all they do is save space. I don't have to hook anything to wires, no one does. I just do because it's fun. You said that bots can do something belts can't, and now you're saying what I do with belts is only possible due to my use of circuits. This isn't true. As I said, pure belts can do everything bots can, they just take up more space, and the more space they take up, the harder they are to use. This is partly why I use circuits, but at one point you start running into problems even with that. You run into the lategame, large quantities of items, large variety of items, lots of hungry assemblers, all with different needs. Trying to wrangle all the belts goes from a fun challenge to a chore.

Is it possible to fill up a train with 40 unique items using only belts and inserters? Using only one stop? I know this is getting a little pedantic, but I didn't mean "belts can't do what bot can do" in the sense you're talking about - with no parameters. A belt setup vs a bot setup might be the difference of about 20x the space and 20x the time, ie: bots can do what belts can't.

And I'm not forcing you to use belts. I make it a point in every post I make - you can keep using bots if you want, I don't care about those at all. They can be as useful and powerful as you want them to be, be my guest. But give me the ability to reach that same level of efficiency with a system that has more depth to it.

I'm not sure bringing the two systems in line would be possible. Late game belts could be buffed, but I don't think they could ever reach the level of bots without having to nerf bots. filling a train with 40 unique times will never be easier with belts. Unless you rename belts to bots and vice versa.

I’ve been thinking a bit about the issue of game progression with regard to this topic. A lot of the support for bots comes from the megabase builders who recognise that scaling up to massive production rates without killing the game performance is only possible with bots, and removing them would eliminate this type of gameplay. To me this is a legitimate concern in an open ended sandbox type game.

This is a very valid point, and I totally agree. Any nerfs to bots would artificially lower the maximum production possible at the very late stage of the game and would dampen fun of all players who both enjoy the power of bots, including the late stage megabuilders who are output focused.

You missed the part where Commander was designed as a casual format.

It doesn't change my response. A casual game can still be competitive, therefore it still needs balance to be playable.

The playerbase can and is splitting over this. If you think it's not, then you need to get your head out of the sand and take a look at why so many people have strong opinions about it.

Daid has a good response to this which I agree with.

It depends what you mean by split. They are certainly not splitting to the point of leaving the game (except in a few rare cases). I see it more as a friendly disagreement.

I'll add that players have strong opinions because they love the game. If Factorio wasn't as great a game as it was there wouldn't be enough people playing it as excessively as they are in order for this debate to occur. It's a testament to how awesome Factorio is.

Nothing in a game is supposed to function any certain way. It does so only by a designer's choice.

Well... Not exactly. Game designers have created game metagames by accident in the past, not by design. Same thing can be said for some mechanics, they don't always get used as intended. Obviously the mechanics themselves are exactly as the designer created them - that's how they got there. Designers don't always get full control the intended outcomes.

Irregardless, by "supposed" I was talking about expectations. Trains are expected to collide because they are solid and are supposed to be traveling on land. Bots on the other hand are assumed to be in the air, on a 2d plane they would move over one another, but not necessarily collide. This is what we would expect to see is shown in game, and it is. But obviously if the designer wanted to go against having any type of collision then that would be their call - it just wouldn't be expected.

If you think that bots are supposed to not have collision detection and that train are supposed to have collision detection, then you already have a misunderstanding of how some players will play.

I've already said multiple times that I think Factorio is primarily a sandbox game. I'm not making any assumptions to how anyone will play - I've even said all play styles are valid I believe? So you're misrepresenting my argument. What I am saying is that trains have collision, and bots don't. I'm assuming both are done for gameplay as well as immersion reasons. I am simply giving my best guess as to why the game may have been designed this way.

I cannot believe that your success in your board game wasn't luck.

This seems to be entirely outside the scope of our argument. I mean no ill will. I am simply trying to express a disagreement with your statements.

What's the optimal way to play Minecraft?

It is entirely dependent on your goals, just like Factorio. For example, both games would have optimal solutions to speedruns. Getting 100 spm is a player imposed goal. Likewise in Minecraft I could set myself a goal of getting as many diamonds per hour as possible - whatever goal I choose it has an optimum way to be done.

Anything without a goal cannot be optimized. I believe that anything with a goal can have an optimal or equally optimal solution.

Both games come with many goals, either explicit or implicit. Players can opt to do all of them, or none of them. (edit: or some of them)

You're not increasing your own fun, you're only changing the relation of your fun vs. the fun of others in your favour. But the overall fun decreases.

imo, this is true. I think there is a reasonable discussion to be had for buffing late game belts. But I don't see any good arguments for nerfing or removing bots.

It's clear to me that nobody wants logistics bots removed, is disingenuous to argue that those that see a problem with bots wants to force bot players to play with belts.

I would argue that the topic as been at least broached, hence it's not disingenuous to argue against that point eg:"This feels hostile, and in Twinsen's case, straight up gets one riled up to the point of wanting the damn bots removed."

Irregardless, my comment was clearly directed at a very specific argument. I actually find it rather careless and insulting of you to take my comment directed towards one argument, and then apply it to everybody.

I don't know how to resolve what I was talking about without removing bots or fundamentally changing the mechanics of belts, hence my comment. Let's look at the context:

I made the statement that belts can't do what bots can do. The counter argument was the they can, it's just much more difficult. My thoughts were: I don't see a way to fix the complexity issue without fundamentally changing what bots are or removing them (effectively the same thing in this context). The problem isn't a throughput issue, but a space and logistical one. Therefore, my argument was that a fundamental change must occur to either bots or belts to bring them in line with each other. This is inline with what I said. Naturally I didn't tackle the bringing belts in line with bots as I don't think that's even a possibility.

Even if my argument isn't correct it was made honestly.

If you want to have an honest debate with me I'm more than open to it, but please to not attack my character by claiming that I am being disingenuous. I find that particularly insulting due to my religious beliefs.

[Tsuran]

I try to see things from a "beginners" perspective:
Factorio is a game about progression. Your primary goal for a long time will be to learn the game mechanics and create efficient setups. What seems super efficient in the beginning, might be terribly inefficient a few science packs later. So the game forces you to adapt your strategies and builds until you reach the current "goal" of it. I like the progress that you make with the different belt tiers. They are getting faster and the underground variants reach further. It is always a fun experience. Up until you reach the point where you realize, that logi-bots setups can make things a lot easier if you just have enough of them flying around. Usually you will come in contact with constructions bots first, as their role is clear and you instantly realize how useful they are.

Again think of the typical player: Currently you have to reach the point with research that enables you to use the bots in the first place, realize what you can or could do with them and set up the assembly to produce enough of them. Also the power requirements for a setup like this are quite considerable as you need A LOT of roboports. These are huge obstacles for the beginner.

I do not share Twinsens argument when he says that they are extremely easy to build. If you have the right setup, then everything is easy in Factorio. But that is the core mechanic of the game and it supports this kind of playstyle.

The question that Twinsens argument raises is a different one: Are logi bots supposed to be superior to belts? If no, then Belts could need a buff (compression, throughput, etc.) If yes, an idea might be to divide construction and logistic bots into different research tiers. If you want to promote the usage of belts, then it might be a good idea to make construction bots available sooner in the game, so huge setups (with belts) are easier to build.

One thing is clear: The things that are being discussed here affect the players that want to min/max with sometimes with insane setups. I do not think that this is a problem for a majority of the player base. Factorioans have quite a diverse playstyle. I know guys who think the game is boring without biters, while I despise them. Others hate rails while some would like to spawn as a locomotive. The great thing about Factorio is the fact that you easily adjust the game to your prefered playstyle. If you want to play without bots, then do so. You will run into problems if you want to create absurd contraptions, but you can reach the goal of the game anyway. Thats what I love about it. You are not forced to a specific way of playing. It might be a good idea, but if you want to try just do so.

Please continue to support diversity in this great game. So far you and this community are doing a fantastic job.

[Carl]

I plan to do my own little thread on this soon but coming at it from the belt side of things as there's a much bigger issue that bot throughput power or beacon power IMO.

But IMO if there's an issue with blueprinting with beacons. make it so no machine can benefit from more than one beacon, up the aura area and module capacity of each beacon, (say 6 modules and a 6 tile radius). The cuts down on the need for densely packed beacons. which makes belts a lot more viable all of a sudden.

[PurpleGreen]

Since Factorio is predominately a creative & problem solving game, it comes down to player preference.

Right, but I disagree that balance is a small consideration in Factorio just because there's no competition, when there's a mechanic that is a one-size-fits-all solution and is simple to use it, it trivializes the creative part of the game. Bot bases are reduced to rows of assemblers flanked by beacons and roboports. Player's preference is usually the path of least resistance, not of the most fun.

this is just one way of building a bot base .. if its fun for someone to do it this way ... let them do it. but there are other ways to make use of bots in the base. and the rows and rows of assemblers you describe ... where is the difference between boring straight mainbelt bases ? its also highly uncreative and not really challenging to just stack belts and side split for ... yeah rows and rows of assemblers.

and here we go ... belts andbots - both can be boring and non challenging, if one chooses to play this way , having fun .

[Koub]

Just to be sure everyone here is aware of the fact : the subject of this topic is no longer accurate.
viewtopic.php?f=38&t=56519
And there has been a lot written in this new FFF that may light things differently.

[Caine]

Can we add (#224) to the title? Such that the context does not get lost.

[lyallp]

Just a note, it's been a little while since I played, so maybe my comments, below, are no longer relevant.

I think that bots should be nerfed such that only a limited number of bots can access a chest or construction item at one time.
It's completely unrealistic that 100+ bots can all get stuff from a chest or perform construction in the same instant.
This not only makes them more realistic but brings things back in line with belts, you can't get dozens of inserters processing against a single chest.
Limit bots to the number of inserters could do, when it comes to pickup/dropoff/construction. Make them queue up and take a few seconds to load/unload/construct.

[PTTG]

Alternatively, make it so that any one bot can only move one item. You could even have a way to reconfigure them with wires, but in order to fully botomate your base, you'd need to have a complicated programming core in addition to all the bots.

[ActofTreason]

Just to be sure everyone here is aware of the fact : the subject of this topic is no longer accurate.
viewtopic.php?f=38&t=56519
And there has been a lot written in this new FFF that may light things differently.

Perhaps the title can be changed to "Belts vs Bots - A response to FFF #224" ? And then the conversation can continue in light of the new information. Not sure if I can change the title, or if it needs a mod. (Edit: never mind, got it sorted. Thanks.)

----

When I wrote my initial post I wasn't sure how many responses or views it would get. I'm glad for the thoughtful discussion. I'd like to reiterate that my intention was never to be antagonistic towards anyone, the players or the developers. Even though we might disagree, the fact that we are having such impassioned discussions is a testament to how great Factorio is. The developers have clearly put a lot of thought and consideration into their game and design choices - the proof is in the pudding.

I'd also like to thank Twinsen for broaching the topic, I'm sure he's received a fair bit of flack for initiating something that ended up being so controversial. But I would like to point out that if it weren't for that we wouldn't be having this discussion. I think the dialogues that were had both here and elsewhere have been informative, and I have certainly enjoyed reading and discussing them.

I'd also like to reiterate why I find this discussion so interesting. I don't think this is something that has an easy solution. Any changes were almost always going to result in some winners and some people who lose out. To try and navigate through such a tricky landscape is tough and poses an interesting challenge. I think the best argument to be made in a situation like this is one in which there are very few people who lose out, and has as many winners as possible, but that's easier said than done.

Anywho, because I'm an opinionated soab I'd like to go through the latest FFF (#225) and share my thoughts:

I'd like to start with the conclusion made by kovarex:

I even believe that robot-only Factory should be possible and not useless, I just don't believe, that they should be 5+ times stronger than anything else. ... Players would still be able to build robot only factory, belt only factory or combination of those, but the strongest strategy would be to combine all types of transport, each for the part where they are the strongest.

As a sandbox game, Factorio should offer many viable strategies. The closer belts and bots are to being in-line with one another, the more incentivised players are to use either on of them based on preference. This is good design and I agree the two should be brought more in line. If I understood, there is potential here in the changes for belts to supersede bots in terms of throughput? This one question alone is very tough to answer and would likely deserve its own post to go into all the depth and pros and cons of each. But, because the FFF doesn't tackle that question explicitly I'll hold my thoughts for now.

Secondly providing a situation where each of the three transport options has a situational use where they are most optimal is also good design imo. This promotes transport diversity (belts, bots & trains) in optimal factories and gives each option a time to shine. This in turn flows on through to providing insentive to players to use each of the three transport options. The three options being closer in line punishes players less who deliberately choose not to opt for the most optimal choice for any given situation.

The changes discussed only really begin to impact late late game, but this is where the issue ultimately lies. Hopefully this provides some longevity to belts and has them being used more in the late late game, and not just bots.

I enjoyed the read put together by V453000 on different solutions to belt throughput. The idea of a new belt or belt research that allows for belt stacking is a particularly interesting concept and I think it has merit. I would want to trial it out before commenting too much on it further. It's all in the implementation.

I can't wait to try out the belt buff. It looks awesome and I think it's an elegant solution to some filter inserter setups that do the same thing.

It's tough to play devils advocate or offer constructive criticism to the #225 FFF post as I agree with a great deal of the principles of it. It's sad to see bots get further nerfed, but I can't think of a compelling case to argue that they should remain untouched. If bots remain more optimal than belts then nothing really changes overall. Nerfing bots will potentially have an effect on the optimum output of those seeking to "min/max" their megabases, but ultimately the "maximum" output can bet set at any number, it's completely arbitrary. Imo, what matters more is how in-line everything is.

[Jürgen Erhard]

I think Twinsen has his head stuck in postgame land. He's seeing players finally reaching Factorio's breaking points in the post post post game, but treating it like a flaw in the main campaign. It's not. Players are reaching those breaking points because their objective is to literally break the game. The distance they have to go to reach those breaking points is nothing short of ridiculous, and the devs should be proud that players have to try so damn hard to finally break something. But step back and put things into perspective. Any balance discussion outside the game's main objective is literally, LITERALLY beyond the scope of Factorio. It's fun to see how far the post game can go, and if you have crazy new tools that help players to go even further beyond that's great. However, game balance in the post game is ultimately not that important. It's certainly not worth game sweeping changes that compromise the enjoyment of the main game.

"postgame": most people would call that "endgame", and it's a disease. All the cries of "but Factorio has no endgame!" THIS ISN'T WoW!

I generally agree, BTW, with all your points.

[BlakeMW]

And there has been a lot written in this new FFF that may light things differently.

When I read the latest FFF it changed my thinking very little. For me the thing which make bots a no-brainer is that they enable Prod3+Speed3 Alternating Rows. If bots charge slower I'd just place more roboports. I suppose at some point the expense in roboports might make me change my mind, but at least at the moment having say 4 rows of roboports instead of 1 row wouldn't be a big deal.

Though one other passing remark was that it would be balanced if belt throughput was tripled: when i thought on this it's largely true. Take a 4x prod3 assembler boosted by 8 Beacons: for electronic circuits the assembler has a throughput of roughly 50 items per second! Even with some optimization around not belting copper cable it works out that 2 blue belts can only sustain around 3 such assemblers. And it's completely true that if belts had higher throughput, they could sustain more assemblers, at 3x the throughput you could have enough assemblers in a row to start getting some good synergy.

But I think more strongly than ever the best solution would be to cap beacon effects (in some manner or another) with or without a change in beacon radius. This would level the playing field between "sparse" and "dense" builds. As it is at the moment bots are enablers of beacon abuse, if we take that away then bots just become a convenience.

[infinitehubgear]

Thanks to Tyson for starting off a quality thread!

On the subject of end-game balance. One idea floating around on the forums that I've found appealing is limited chest access, it does have a bit of realism it it. Instead of making it a permanent nerf, the player could research faster chest access until its negligible. I'm not sure what the game performance implications are but queuing for charging at roboports already seems cpu efficient. Maybe one of the devs can clarify?

I have found it satisfying when a robot speed research is completed, you visibly see the difference. Perhaps a similar satisfaction will be found when noticing robots accessing chests faster. With this a player has to work harder to get the full power of bots that we have now, but its not denied to them in the far-end-game. Ultimately it would not solve the belt-bot imbalance of the far-end-game, but it might give another dial to tune this imbalance long enough to matter.

[anarcobra]

I'm not saying you don't deserve that simpler option. It allows you launch a rocket or a few, fair enough. But it's really disheartening when people like me pour hours into making elaborate systems only for the game to push us in a direction we don't want to go.

I really don't understand how the game is pushing you to use bots when you don't want to. Didn't you have fun designing your sushi belt? I assume yes, otherwise why do it. But why does the existence of bots bother you so much? Is it because you are envious of other players who do it fast and easy? Even if bots weren't there, other players would simply use dedicated belts. Would you feel pushed to dedicated belts? Not trying to be mean, but I just don't understand this push people seem to feel. I think you should be proud of your design, but not be angry that other players are choosing not to go that route.