Belts vs Bots - A response to FFF #224

[Caine]

The main thread seems to be mostly an "voice your opinion" dumping ground which moves very quickly. There is some discussion weaved in there, but it is hard to find.

I say, don't merge. I would personally prefer a couple of separate threads investigating a specific direction than one place where everything gets mixed. That said, you are the moderator. It is your judgement call.

[ActofTreason]

In this case the breaking point for bots is way way WAY beyond the primary game. Saying that bot bases are OP is like saying that my RPG dream team of max stat lvl 99's is horrendously overpowered. It's certainly true, but it's not a design flaw and it doesn't mean the game's balance is broken. It just means that the player beat the final boss, killed all the super bosses and essentially reached the upper limits of what your game can do.

If I understand you correctly, then I totally agree with you. I think we might be working with slightly different definitions of balance, but ultimately, we agree.

I'm not sure I would agree that bots have a "breaking point". They are simply powerful, they don't fundamentally break the game. They are simply the dominant strategy later on , but ultimately the game still works and nothing is "broken" to the point of it being unplayable. Players are just incentivised to choose bots over belts, but both are still valid depending on your goals.

Belts have the ability to fully replace bots, trains, and, if we consider barreling, pipes as well, in every situation imaginable. This is a fact. What sets trains, pipes and bots apart from belts is that they solve certain problems more efficiently, i.e. by providing more throughput, being more space-efficient, and/or easier/faster to set up. Bots have the unique attribute of being superior in all three of those aspects once bot construction is up and running and ample power supply is provided. There is practically no need to build belts in the lategame.

This wouldn't pose much of a problem by itself, after all nobody is complaining about solar or nuclear superseding coal/solid fuel power generation in the lategame either. The issue that at least I personally have with bots is that they do not pose an interesting design challenge, and if you choose to rely on them (which by the way, the motivation is very strong lategame what with bots being superior efficiency-wise), your Factorio experience will go from designing elaborate belt balancers based on various item-to-item ratios, circuit-controlled or otherwise, to just plopping down roboports and colorful chests everywhere. The design challenge is just not there anymore, and if you care for that kind of thing, the game becomes an extremely deprived version of itself. I believe your article calls this the "creative node", whatever that means. Bots give you none of that, belts do. All the while belts aren't an efficient, versatile transport system like the bots are lategame, which causes people to feel like the game is setting up an expectation of them abandoning belts for bots, and with them all the fun they were having. This feels hostile, and in Twinsen's case, straight up gets one riled up to the point of wanting the damn bots removed.

Fun is subjective. A change to nerf bots or to buff belts will be different, but I see no compelling evidence for it being better overall. A possible example of this could be a player who enjoys the increased output from bots, and doesn't like having to mess around with "elaborate belt balancers based on various item-to-item ratios, circuit-controlled or otherwise..." etc etc. Such a player will enjoy the game less when they have the bots removed from the game.

Lets pretend for a second that I am that certain type of player. I wanted to create a 1k spm megabase. I don't want to mess around with all the fiddly belts and such, I'm just focused on maximum output. Now, lets remove bots from the game. You haven't just incentivised me to play a certain way - you've actually forced me and removed the alternative as an option. Games are all about choices, players respond to incentives. If you incentivise belts to be more optimal than bots you're just incentivising me to play a way that I don't want to play - but at least I still have the option to play with bots for the ease of use.

My article calls enjoyment derived from the problem solving challenge (you call it the design challenge; same thing) the problem solving node and the creative enjoyment the creative node. The definition is shown in the article and describes creativity as "... being able to create something, not just having to follow a beaten path, or use existing tools for a set preordained purpose. You get to create something your way, your style." you're referring to the problem solving node when you say: "The design challenge is just not there anymore, and if you care for that kind of thing, the game becomes an extremely deprived version of itself." Optimal play never allows for creativity. Discovering the optimal play always allows for problem solving, but once the dominant strategy is obtained, problem solving no longer occurs because the best course of action is known and no more problem solving needs to occur. This is an unsolvable problem in my opinion, and factorio will always suffer from it unless you fundamentally change the type of game Factorio is. Trying to fix up the existence of dominant strategies in games is like trying to fight a never ending fire.

You mention that the existence of late game bots "causes people to feel like the game is setting up an expectation of them abandoning belts for bots, and with them all the fun they were having. This feels hostile". I can't say I've ever experienced that feeling for any prolonged time. Maybe when I first started playing, but that's so long ago now. I don't see it that way - I see bots as a fun option I can choose to include in my games if I am so inclined. I mostly use them for things that I would otherwise be completely unable to do with belts. As such I am thankful for the option, it allows me to be more creative with my solutions, giving me more options to create and problem solve.

Should I merge this topic in the FFF one to get everything at the same place ?

My preference is no, and I agree with the rationale of Zavians and Caine. I think we have a slightly different take on this than the other thread. I'm not so much concerned with visibility as I am with having a solid discussion, and we may lose some of that with a move to the other thread. Thanks

[oracleofepirus]

Factorio is not a competitive game, but a creative and problem solving one which also greatly benefits from a sense of discovery and progression. As such, balance considerations do play a small part in Factorio, but are far less significant than in other games, such as counter-strike where players are highly incentivized to choose the dominant strategy if they want to win. In Factorio there is a dominant strategy, but there is very little incentive to choose it. It is a question of creativity instead of competition. In fact, as long as nothing is absurdly broken or useless, balance should not negatively impact player enjoyment in Factorio.

In this case the breaking point for bots is way way WAY beyond the primary game. Saying that bot bases are OP is like saying that my RPG dream team of max stat lvl 99's is horrendously overpowered. It's certainly true, but it's not a design flaw and it doesn't mean the game's balance is broken. It just means that the player beat the final boss, killed all the super bosses and essentially reached the upper limits of what your game can do.

Balance is always a major point of design. If there is a broken point in a game, players will reach it.

Take the Commander format for Magic: The Gathering. The format rules committee has repeatedly stated that the format rules are designed for ease of understanding and fun. But any reasonably intelligent player will tell you that the banlist is nonsense. It took the committee years to ban a card that starts subgames. It wasn't banned until they figured out that the only players using it were only using it to be assholes (ie, start six million subgames). Literally zero players were using it for the rules committee's definition of fun. At this point, because the banlist is so short, almost every player and game store has additional custom restrictions on which cards can be used, which decks can be played, which players are allowed to play, and even what actions they are allowed to take in-game. What's supposed to be a casual format that can be played by any two people from anywhere is now exceedingly clique-specific. The format has serious platform fragmentation issues that cannot be fixed short of a complete rules overhaul. The difference between bots and belts doesn't have to be fixed, but if it doesn't, I guarantee that the playerbase will fragment, and the losing fragments will quit.

I'm not sure I would agree that bots have a "breaking point". They are simply powerful, they don't fundamentally break the game. They are simply the dominant strategy later on , but ultimately the game still works and nothing is "broken" to the point of it being unplayable. Players are just incentivised to choose bots over belts, but both are still valid depending on your goals.

Everything has multiple breaking points because everything can be mathematically defined. However, you are correct that aren't bots fundamentally overpowered. In this case, it's the implementation. Everything in Factorio has collision detection, except for one entity. Can you guess which one that is? I dare you to turn off collision detection in any game and still call that good design and say it isn't broken. In fact, I dare you to turn off Factorio's train collision detection.

Optimal play never allows for creativity. Discovering the optimal play always allows for problem solving, but once the dominant strategy is obtained, problem solving no longer occurs because the best course of action is known and no more problem solving needs to occur. This is an unsolvable problem in my opinion, and factorio will always suffer from it unless you fundamentally change the type of game Factorio is. Trying to fix up the existence of dominant strategies in games is like trying to fight a never ending fire.

This observation is mostly correct. Optimization is the polar opposite of creativity. However, optimization can be reduced to a negligible point. Optimization requires three things, a fixed starting point, fixed mechanics, and a fixed endpoint. If, at any time, those conditions change, then optimization must either restart, or is no longer possible. While the starting point and mechanics cannot change in Factorio, it is entirely possible to have mechanically opposing endpoints. However, since Factorio is still incomplete, any arguments for endpoints must wait.

[BlakeMW]

The more I think about it, the more it's obvious that as long as using bots roughly doubles the utilization of Prod3 and Speed3 modules, then it doesn't really matter how good bots are relative to belts, bots will be the prefered solution for min/maxing players - unless like bots become so expensive the investment in bots becomes comparable to the investment in the modules. It can be summarized that for various reasons the Prod+Speed combination is too good, too good to pass up.

I've often thought that speed bonuses and negatives should be applied separately, that is when you have a -60% speed penalty and a +100% speed bonus, instead of getting a final speed modifier of (1 - 0.6 + 1) = 1.4, instead you get (1 - 0.6) * (1 + 1) = 0.8, this would make the speed penalty and consumption penalty of productivity modules much less negatable and a setup where you have twice as many assemblers with productivity modules more comparable to one where speed beacons are used to boost the assembler speed, in other words it would less favor dense packing (dense packing would still be favored, not just as massively so). Productivity is basically too good to pass up and it doesn't need to have massively powerful synergy with speed modules to be worth using, these days we even have productivity modules being used in the world record speed runs - and without beacons at all.

So I think that massively nerfying the synergy of prod+speed would also go some way to making bots less of an equilibrium strategy. (Note that even if the synergy is nerfed, just having your productivity assemblers run twice as fast from a +100% speed bonus would be very worthwhile, there's no need for it to be an effective +250% speed boost to be worthwhile from a Return on Investment perspective)

Now this suggestion of nerfing the synergy is by no means a complete solution, but I think it could be part of the solution, particularly since as long as Prod+Speed has really OP synergy any logistic method which best supports it will be most favored.

[Drury]

Optimal play never allows for creativity. Discovering the optimal play always allows for problem solving, but once the dominant strategy is obtained, problem solving no longer occurs because the best course of action is known and no more problem solving needs to occur. This is an unsolvable problem in my opinion, and factorio will always suffer from it unless you fundamentally change the type of game Factorio is. Trying to fix up the existence of dominant strategies in games is like trying to fight a never ending fire.

You mention that the existence of late game bots "causes people to feel like the game is setting up an expectation of them abandoning belts for bots, and with them all the fun they were having. This feels hostile". I can't say I've ever experienced that feeling for any prolonged time. Maybe when I first started playing, but that's so long ago now. I don't see it that way - I see bots as a fun option I can choose to include in my games if I am so inclined. I mostly use them for things that I would otherwise be completely unable to do with belts. As such I am thankful for the option, it allows me to be more creative with my solutions, giving me more options to create and problem solve.

Again with the "bots can do things belts can't"

Belts can do anything bots can, and in a less braindead way. I have train stations that use circuit balancers to unload items onto sushi belts depending on what items are the most needed at the time, then a couple of sets of filter inserters (also circuit-balanced) filtering items off the sushi belt and feeding them onto dedicated belts. I could rip all that circuitry apart, replace the chests with provider chests, put requester chests at each assembler and call it a day. What I don't get is why you call that the more creative option. You say there is problem solving involved that way - where? When you need to expand your logistics to accommodate extra manufacturing, you just plop down a provider chest and a requester chest. When you need more throughput, you just pump bots into the system. That's about as much problem solving as plopping down more solar panels and accumulators.

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.

[rasmusbergpalm]

I'll skip the more philosophical "what is fun" debate and assume we wish to incentivise players to use more belts and less bots (note: less bots, not none)

The main reason I see that bots are OP is that they have unlimited throughput. With enough bots you can empty a million plates out of a chest in a single tick. Belts and inserters just can't do that.

So the solution is simple: limit the throughput of bots to x% of that of an optimal leveled inserter + belt setup.

One simple way is that only 1 bot at a time can insert or remove from the same chest at a time, and that they insert/remove comparable to that of a leveled inserter.

That's one part of the equation the next is speed. Since bots go in straight lines and belts in manhattan, that means the belts have to go (a+b)/sqrt(a^2 + b^2) <= sqrt(2) longer, so make the max bot speed 1/sqrt(2) * leveled belt-speed. Factoring in factory layout + terrain constraints is harder but multiply by a fudge factor until it seems rights.

This way one inserter + belt is as ~effective as a conga-line of bots throughput wise.

Now you can tune the different variables up and down to more or less incentivise players to use belts or bots.

If you set the variables such that bots have ~50% the throughput of belts, then it'll be like everything else in factorio:
There's an easy way (using bots) for that "aah I don't care i just have to fix it" situations, and then later you'll have to refactor it to belts to increase throughput 2x.

I think updating the default map mode and leaving the un-nerfed bots in a different map mode (e.g. powerful bots), similar to how we e.g. have rail-worlds, would be a nice middle ground.

[ActofTreason]

Take the Commander format for Magic: The Gathering. The format rules committee has repeatedly stated that the format rules are designed for ease of understanding and fun. But any reasonably intelligent player will tell you that the banlist is nonsense. It took the committee years to ban a card that starts subgames. It wasn't banned until they figured out that the only players using it were only using it to be assholes (ie, start six million subgames). Literally zero players were using it for the rules committee's definition of fun. At this point, because the banlist is so short, almost every player and game store has additional custom restrictions on which cards can be used, which decks can be played, which players are allowed to play, and even what actions they are allowed to take in-game. What's supposed to be a casual format that can be played by any two people from anywhere is now exceedingly clique-specific. The format has serious platform fragmentation issues that cannot be fixed short of a complete rules overhaul. The difference between bots and belts doesn't have to be fixed, but if it doesn't, I guarantee that the playerbase will fragment, and the losing fragments will quit.

This is comparing apples and oranges. Factorio is a single player creative problem solving game in which one constructs factories. It's not a competitive game. Balance is important as I have already said, but it is less so than a competitive game. In Factorio choices need to be valid at some point, but that's about it. Magic the Gathering is a highly competitive game in which the balance and rules are of paramount importance - without it the game breaks and barely becomes playable - it stops being fun.

I highly doubt the playerbase will split over this. It would be like splitting over the use of bobs mods. We all don't have to agree to play the game one way.

Everything has multiple breaking points because everything can be mathematically defined. However, you are correct that aren't bots fundamentally overpowered. In this case, it's the implementation. Everything in Factorio has collision detection, except for one entity. Can you guess which one that is? I dare you to turn off collision detection in any game and still call that good design and say it isn't broken. In fact, I dare you to turn off Factorio's train collision detection.

I'm not sure we're working with the same definition of breaking point, but it would be interesting to hear you elaborate on what you mean. Regarding collision, I don't think it would matter for bots on or off. It's just a convenience thing. All you're doing if you add unit collision to bots would be making them more annoying to use, and they would likely need re balancing. Trains on the other hand are obviously supposed to collide. That's just a game immersion issue more than anything. If you could build hover trains and they were supposed to glide over one another then that would be balanced too.

This observation is mostly correct. Optimization is the polar opposite of creativity. However, optimization can be reduced to a negligible point. Optimization requires three things, a fixed starting point, fixed mechanics, and a fixed endpoint. If, at any time, those conditions change, then optimization must either restart, or is no longer possible. While the starting point and mechanics cannot change in Factorio, it is entirely possible to have mechanically opposing endpoints. However, since Factorio is still incomplete, any arguments for endpoints must wait.

I'm not convinced we are talking about the same thing. For any given choice you will have a "best" option, or two or more "Equally best options" depending on your desired outcome. You can change all the parameters you want but this will always be the case surely? Since Factorio is a sandbox game the player objective is more user defined than game defined.

I have no idea how you plan to reduce optimization to a negligible point in a game like Factorio. Doing this while maintaining a sense of progression within the game also seems like an impossible task. Are belts as good as bots? yes? Then why build bots when there is an investment cost associated with them?

Again with the "bots can do things belts can't"

Belts can do anything bots can, and in a less braindead way. I have train stations that use circuit balancers to unload items onto sushi belts depending on what items are the most needed at the time, then a couple of sets of filter inserters (also circuit-balanced) filtering items off the sushi belt and feeding them onto dedicated belts. I could rip all that circuitry apart, replace the chests with provider chests, put requester chests at each assembler and call it a day. What I don't get is why you call that the more creative option. You say there is problem solving involved that way - where? When you need to expand your logistics to accommodate extra manufacturing, you just plop down a provider chest and a requester chest. When you need more throughput, you just pump bots into the system. That's about as much problem solving as plopping down more solar panels and accumulators.

This is one issue when talking about game design. One must necessarily get very specific with their language or else miscommunication will occur.

I understand that belts + circuit are capable of replicating what bots do. However, belts alone cannot do this task.

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?

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.

Factorio is one of the few games where it doesn't matter which direction the game is "pushing" you in. You don't have to choose it. imo, it in no way diminishes your accomplishment to have created that sushi belt solution. Just because there was an easier option and you made the choice not to use it doesn't mean it isn't impressive in its own right. If anything it makes it more impressive because you were never forced to do that solution. I'm surprised you find it disheartening at all.

[Daid]

I've been reading up on this. And as a non-bot player, I'm confused by 1 thing.

I'm reading that bots are basically "unlimited". However, when I do use construction bots, I noticed them waiting at roboports to be charged, in line, because a single port can only charge X bots at the same time. Doesn't this limited the maximum throughput of bots? For every X amount of traveling, they need Y changing time, and for every Z*Y continues charging time you need a roboport? So won't your whole base just end up as roboports? To have enough charge-time for robot travel? Or is the charge time per fly time so low that it really doesn't matter that robots need to be charged?

[Drury]

I understand that belts + circuit are capable of replicating what bots do. However, belts alone cannot do this task.

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?

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. You need a more versatile solution, something that can deliver all those items where they need to go - and the bots are it. The problem is that they completely lack the depth that the belts have in the midgame. You end up having to choose between frustration and boredom. That's what I mean by the game pushing you.

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.

[BlakeMW]

I've been reading up on this. And as a non-bot player, I'm confused by 1 thing.

I'm reading that bots are basically "unlimited". However, when I do use construction bots, I noticed them waiting at roboports to be charged, in line, because a single port can only charge X bots at the same time. Doesn't this limited the maximum throughput of bots? For every X amount of traveling, they need Y changing time, and for every Z*Y continues charging time you need a roboport? So won't your whole base just end up as roboports? To have enough charge-time for robot travel? Or is the charge time per fly time so low that it really doesn't matter that robots need to be charged?

The energy usage vs charging speed is balanced such that you tend to need maybe 1/5th of your factory to be roboports if you go with mega-bottage, they will also use a substantial percentage of the factory energy usage. It's not a trifling investment in roboports, but it's still pretty cheap compared with the beaconized setups the roboports are supporting.

[rcp27]

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. On the other hand the game is structured so that launching a rocket is the notional objective of the game, and the tech tree is structured so that all core mechanics become available by the time you are in a position to send up your first rocket. The game is designed so that getting the rocket launched is within the reach of a reasonably casual player to keep the broad appeal of the game. The point at which supporting production without a game breaking loss of performance without bots is necessarily well beyond the rocket launch, but equally the “easy mode” bots become available well before that point. To my mind it is this disconnect that gives rise to the conflict about whether bots are a necessary step in game progression, or an unfair “iwin button”. An attempt to mitigate this was the gating of requester chests behind high tech science.

Perhaps the solution is to produce some mechanic that limits the usefulness of logistics bots in a way that is gated behind a significant production of space science. Maybe a finite bot-chest interaction time that cripples logistics bot utility for large scale production at first, but is reduced via infinite research such that the point at which bot throughput outpaces belts comes several levels down the line. This would allow bots to be useful for the low output hard to make items like satellites for new players, and give them the chance to get to know bot mechanics, but prevents them from becoming the go-to solution for every task. For the mega base builders, getting the science needed to cross the threshold for bots to take over will still be possible, preserving the gameplay they enjoy. Essentially the idea would be to push the transition much later in the game.

[Zavian]

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. On the other hand the game is structured so that launching a rocket is the notional objective of the game, and the tech tree is structured so that all core mechanics become available by the time you are in a position to send up your first rocket. The game is designed so that getting the rocket launched is within the reach of a reasonably casual player to keep the broad appeal of the game. The point at which supporting production without a game breaking loss of performance without bots is necessarily well beyond the rocket launch, but equally the “easy mode” bots become available well before that point. To my mind it is this disconnect that gives rise to the conflict about whether bots are a necessary step in game progression, or an unfair “iwin button”. An attempt to mitigate this was the gating of requester chests behind high tech science.

Perhaps the solution is to produce some mechanic that limits the usefulness of logistics bots in a way that is gated behind a significant production of space science. Maybe a finite bot-chest interaction time that cripples logistics bot utility for large scale production at first, but is reduced via infinite research such that the point at which bot throughput outpaces belts comes several levels down the line. This would allow bots to be useful for the low output hard to make items like satellites for new players, and give them the chance to get to know bot mechanics, but prevents them from becoming the go-to solution for every task. For the mega base builders, getting the science needed to cross the threshold for bots to take over will still be possible, preserving the gameplay they enjoy. Essentially the idea would be to push the transition much later in the game.

I think if the devs wanted to go that route, they could just move requestor/buffer chest to later in the tech tree (ie make them require space science. Player logistics requests would still be available. It's not that hard to request the solar panels + everything else and just handcraft your first satellite. If they wanted a replacement for logisitic bots being used for low volume production they could also implement ropeway conveyors viewtopic.php?f=6&t=7977 . The conveyors could be relatively low volume, and relatively slow, and hence not really suitable for bot based production of anything that is needed in any volume, but still able to handle things like satelite production, and supply construction trains/walls etc.

[oracleofepirus]

This is comparing apples and oranges. Factorio is a single player creative problem solving game in which one constructs factories. It's not a competitive game. Balance is important as I have already said, but it is less so than a competitive game. In Factorio choices need to be valid at some point, but that's about it. Magic the Gathering is a highly competitive game in which the balance and rules are of paramount importance - without it the game breaks and barely becomes playable - it stops being fun.

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

I highly doubt the playerbase will split over this. It would be like splitting over the use of bobs mods. We all don't have to agree to play the game one way.

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.

I'm not sure we're working with the same definition of breaking point, but it would be interesting to hear you elaborate on what you mean. Regarding collision, I don't think it would matter for bots on or off. It's just a convenience thing. All you're doing if you add unit collision to bots would be making them more annoying to use, and they would likely need re balancing. Trains on the other hand are obviously supposed to collide. That's just a game immersion issue more than anything. If you could build hover trains and they were supposed to glide over one another then that would be balanced too.

You have a fatal understanding of design if you think that convenience does not play a major part of it. Nothing in a game is supposed to function any certain way. It does so only by a designer's choice. 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. Train players would be thrilled to not have to fuck with millions of chain signals while questions about train intersections would just cease to exist.

Everything has major effects given enough time and even cosmetic changes are no exception. The removal of black magic splitter sorting in today's update not only changes the appearance of a splitter's output, it also affects lane balancer designs.

Every choice you make, whether you perceive it or not has a major effect down the line. Do you know why atm card pins are 4 digits long? It's because the inventor decided that the security of two extra digits wasn't worth hassling his wife.

All things have points at which they no longer function or render other parts useless. Most of those points have to do with economy of scale. Some have to do with automation. Take a look at how many things break when a Wal-Mart comes to town. If you are unable to apply game theory to literally everything, then I cannot believe that your success in your board game wasn't luck.

I'm not convinced we are talking about the same thing. For any given choice you will have a "best" option, or two or more "Equally best options" depending on your desired outcome. You can change all the parameters you want but this will always be the case surely? Since Factorio is a sandbox game the player objective is more user defined than game defined.

I have no idea how you plan to reduce optimization to a negligible point in a game like Factorio. Doing this while maintaining a sense of progression within the game also seems like an impossible task. Are belts as good as bots? yes? Then why build bots when there is an investment cost associated with them?

Optimal requires a fixed starting point, a fixed set of game mechanics, and a fixed ending point. You want to talk apples to apples? There are games with fixed starting points, fixed mechanics and noending points.

What's the optimal way to play Minecraft?

[Daid]

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.

The player base is also split over mods or not. And (in less extent) over biters or not.

Which doesn't matter. A split player base doesn't matter. Unlike magic, where you REQUIRE a lot of other players to make it fun, in factorio, they are optional, and you only need a few that want to play like you do, if you care about multiplayer that is, and another split right there. Which all doesn't matter...

UNLESS, the developers start tailoring for one of the splits that has no possibility to play like the other option. For example, greatly nerfing bot factories. But they clearly stated in the FFF that they have no such desire. They won't remove bot factories, they won't remove mods, they won't remove single-player. As those are well established features within the game.

[Yinan]

I always find it strange that people who want to incentivise something try to do that by impeding the concurrent thing instead of actually making the thing that they want to incentivise better.

Both in this thread and in the FFF thread I see so many solutions about "how to limit bots in certain ways" and I just don't get why that is necessary.
You wan't to incentivise players to use belts? Then buff belts, make them better or add more options to it.
What do you get from limiting/nerfing bots other than the "satisfaction" of that other players that enjoy using bots the way they are can't enjoy them that much anymore.
You're not actually bettering anything for you, you're just making it worse for others.

Or basically:
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.

[BlakeMW]

I always find it strange that people who want to incentivise something try to do that by impeding the concurrent thing instead of actually making the thing that they want to incentivise better.

So far in this thread I've focused on the fact that Bots allow for better utilization of prod3 assemblers and speed3 beacons, by allowing more of them to support each other and ramping up the speed multiplier.

I previous suggested nerfing beacons so they only have 1 radius coverage.

But part of the problem with the beacons is that because they have a 3 tile radius, you can have a 2 tile gap between rows of beacons and rows of assemblers: this is sufficient space to make good use of requester and provider chests, but not enough space to make good use of belts.

Instead of nerfifying beacons, there is in fact a buffification solution: Increase the beacon radius from 3 to 4 tiles! This wouldn't do terribly much to make bots better, but by allowing 3 tiles gaps between beacon rows and assembler rows it would allow players to fit in two belts - conveniently double the belt throughput!

[Caine]

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.

[Koub]

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.

But you'd have to choose between fitting 2 beacon rows and not have the room to provide enough throughput via belt (so bot use) or just one row, and allow yourself less convoluted contraptions via belts, due to more space available.

[Caine]

That sounds like a nail in the coffin for belts to me. Currently you need to rely on belt weaving to be able to have the same amount of beacons and the choice clearly goes to bots for many megabase players. With this setup, it would effectively half the amount of beacons you can have when using belts instead of bots. That is a massive reduction in factory output.

[Zavian]

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.

But you'd have to choose between fitting 2 beacon rows and not have the room to provide enough throughput via belt (so bot use) or just one row, and allow yourself less convoluted contraptions via belts, due to more space available.

But that would just encourage bot based builds where people try to get as high a craft speed as possible. I saw somebody about a week ago who insisted be wanted his bots carrying copper wire, rather than doing direct insertion, because he wanted a craft speed of 8. If you could add more beacons then he'd probably an even higher craft speed. If beacon range was 4, and there wasn't also some cap to the number of beacons you could use (or a scaling mechanic that made more beacons less effective), then I expect the standard 8x8 build with a row of beacons, then a row of assemblers, then another row of beacons to become 2 rows of beacon, then the row of assemblers (with no gap for belts or inserters or chests), then the row of inserters + chests/belts, then another row of beacons. I think such a design would have a craft speed of 8, (rather than the current 5.5 for an 8.8 design), and wouldn't have any more room for belts (indeed it has less room for belts, because you can tunnel belts under inserters on both the left and right of the assembly machines in the current belt based 8x8 builds).