Factorio Forums

TL;DR

Don't recharge construction/logistics bots when they are mined.

What

When you mine a bot, it forgets how much charge it had. This effectively recharges them to full charge without actually spending any energy (and without any charge time). The suggestion is that bots should always remember how much charge they have, or at least bot charge should work something like how damaged items are handled and put "less than 100% charged" bots in their own stacks and average their charge rather than giving them all 100% charge.

Why

Providing bots with sufficient charge:
* is a logistical challenge (i.e. how do I provide enough power?)
* leads to meaningful trade-offs (e.g. how many batteries do I want in my armor vs other items)
* leads to meaningful tactics/strategy (i.e. how do I best make use of the bot charging power I have - how and where do I deploy my personal construction bots to maximize the work done while minimizing time spent and charge consumed)
Having the cheat (free full charge) in the game short-circuits all of the above. It leads to a situation where some players, instead of properly playing the game (improving power, using enough batteries, being mindful of how much charge a task is going to use, etc.) don't bother with that portion of the game at all (they load up on power-sucking exoskeletons, use little/no batteries, run/drive away from their bots all the time so their bots run out of power chasing them, etc.). I.e. they don't fix the problems with how they are using bots because they don't have any good reason to - they just use the cheat instead.
Now if I were just playing the game in isolation this would hardly be any issue at all - I just don't use the cheat and might not have even ever found it. But pretty much every Factorio youtuber I've seen uses the cheat, and they tend to use the time they spend cheating as an opportunity to complain about Factorio's bot charging. Just check out an example. He goes as far as claiming "There's no way you can keep robots powered."
As far as I can tell, that statement is false. I use personal construction bots all the time (and lots of them - 6+ Mk2 personal roboports is not uncommon for me), and I don't seem to ever run into the bot charging problems that these youtubers run into (once I have reactors - obviously power is quite constrained when running on portable solar panels, though even then batteries do wonders). But then I'm using more batteries, less exoskeletons, not running away from my bots all the time, not sending personal construction bots on unnecessarily long-range tasks (e.g. stand in the middle of what I'm having them do rather than way off to one side), do deconstructs in reasonable chunks (rather than telling them to deconstruct 6 screens full of trees in one go), etc.
If the cheat were removed, these youtubers would basically be forced to learn to play right. The value for me is less annoying videos. The value for the Factorio team is to have youtubers that show the game being played properly instead of constantly cheating and then complaining about the "need" to cheat and about how dumb Factorio's charge mechanics are. If they can be moved to playing right rather than cheating+complaining, that would give prospective players a better impression of the game.
(Anything that can raise "bot charge strategy awareness" might also help - tutorials, FFFs, achievements, statistics (?), better power usage feedback, etc. I believe getting rid of the charging cheat would be the most efficacious means to achieve the desired result, but of course removing the cheat does not preclude also pursuing other approaches.)

I largely disagree. I use modular armor w/ roboports a lot, most the time I rely on PSPs (and the roboport internal batteries) although I'm not above vacuuming up bots. The youtuber you link to is absolutely correct: there is no way to use so many bots with such a high speed research and keep the roboports charge up, this is just basic maths - robots use energy by distance traveled rather than time in flight, so robots with high speed multiplier are sucking down absurd amounts of energy per second. (I believe robots use 5kJ per tile and have a base speed of 3.6 tiles/s, a speed multiplier of 4x would result in 5x3.6x4 = 72kW - a fusion reactor could sustain roughly 10 such robots full time and the maximum of 4 fusion reactors could run 40 bots full time - a 4x speed multiplier is only the non-infinite speed techs (+305%) so with infinite speed research the number of bots which can be sustained is even lower - you could be reasonably looking at only 4 or 5 bots per fusion reactor)

See I think that while there are issues with roboport recharging, taking away the ability to hoover up discharged bots (and in the process recharge them) isn't the way to go about and it would introduce as many issues as it solves. At the moment it provides a useful way to manually intervene when there is *no way* to charge as many bots as you wish to use, what are you going to say? "well, players shouldn't have such ambitious construction projects, or spend more time twiddling their thumbs"? If the ability to hoover-charge bots is taken away, it should only be after resolving the intractable issues with using personal roboport (I am aware that sometimes hoovering is done out of laziness or stinginess, that doesn't take away from the fact there are situations where even 4x personal fusion reactors can't even come close to providing enough power).

I think perhaps the game would be better if there was an efficiency boost to go along with the speed research, so one way or another robots with a higher speed multiplier use less power per tile travelled allowing them to be more productive per unit of energy.

Totally agree with BlakeMW : before fixing the auto-recharge exploit, there should be a rethinking on how construction bots use their energy, and how this energy is provided to them, especially in the case of early modular armor configuration

I don't see what this adds to the gameplay, just adds frustration to those early moments when you only have solar panels

BlakeMW wrote:math

Your math is kind of irrelevant, part 1: His issue is that his batteries are dead. And you can see in the video @7:18 that he kills his batteries by deconstructing trees at the full distance of his roboports (and then some since he runs away from the bots). That's the exact kind of inefficiency I was talking about - zero consideration for how much robot charge he's wasting by doing that.

His claim was that using Mk2 roboports is the problem, not robot speed research. I've used the Mk2's and know that hasn't been a problem for me. I have not got into the infinite science yet so I guess it's possible I may see issues at some point. But still, bot charging before space science seems OK to me. The youtubers are constant offenders - they don't just start doing this after bot speed is into infinite science territory, they do it from the moment they get bots, and it never stops.

I'm not really expecting to see issues even with space science bot speed upgrades, because as you say the energy used is related to distance covered. To build something the same way involves the same distance covered regardless of how fast the bots go, so the same project always requires the same amount of energy. Since I depend on batteries to provide "building bursts" of energy, it doesn't really matter how fast that energy is drained (at least in terms of my energy situation - it could matter in terms of the actual robot charging time so I'll have to see if that becomes an issue). This could be in large part though due to me not being a "speed runner" - by the time I'm ready to do the next build my batteries tend to be fully charged once gain. (But the youtubers I've seen aren't exactly speed runners either - like me they have plenty of downtime between builds.) It may be that you are more of a speed runner and are hitting real issues that I never will, but that's not what I'm seeing in the youtube videos - there I just see a lack of power and an abundance of inefficient behavior.

Your math is kind of irrelevant, part 2: I do not consider running personal bots "full time" as a valid measure. Bots are used in bursts - you start some task involving construction, you do the task, and in some time you're done - end of burst. That may involve robots each doing multiple trips, but there's always stuff to do in between bursts - there's eventually some downtime for the bots. (Even in the portable solar panel phase I tend to find plenty of stuff to do while my batteries recharge, though in that phase I'm also judicious about what I use the bots for - I don't use them for stuff I can do just as fast myself, for example.) This downtime isn't just true for me, it's also true of the youtubers. You should have enough batteries to handle your burst size. If you look @10:04 you'll see he's running with only two batteries - that's kind of inadequate IMO. His inefficient use of bot charge only compounds the problem of not enough batteries.

Another way to debunk "full time" as a valid measure is that it completely negates the need to use bots efficiently. If you can run all the bots your roboports can support continuously, for as long as you want, then there is once again (just like using the cheat) no reason to use bot charge efficiently. I.e., whatever Factorio does, it should definitely *NOT* have allowing full-time use of all your personal bots as a goal as doing so directly DAMAGES gameplay.

Some bot use cases involve extended sequences of bursts -- e.g. building long sections of track and walls, so maybe those would be cases where enough bot speed is able to get it to the point where power generation can't keep up. But even there, when using personal roboports I don't do those tasks with giant blueprints or at full roboport range because that would be really bot-charge-inefficient.

BlakeMW wrote:taking away the ability to hoover up discharged bots (and in the process recharge them) isn't the way to go about and it would introduce as many issues as it solves.

It would not *introduce* any issues. At best it might better reveal existing issues, and revealing issues is the first step in addressing them. If you never reveal them, they're unlikely to be addressed.

BlakeMW wrote:At the moment it provides a useful way to manually intervene when there is *no way* to charge as many bots as you wish to use

Are you talking roboport charging pads and charging rate, or input energy (reactors + batteries)? If it's the latter, then you are essentially just reiterating your assertion that it should be possible to run all your personal robots "full time", which I've already rejected.

BlakeMW wrote:what are you going to say? "well, players shouldn't have such ambitious construction projects, or spend more time twiddling their thumbs"?

There are full-size roboports. You can use many of them at once. Ambitious construction projects are fine, but players should use the right tool for the job.

Next you'll be telling us how miners should mine all the ore under them instantly and belts should move 1e9 items per second and assemblers should have crafting speed one million because otherwise players are going to have to "spend more time twiddling their thumbs". Players should just learn how to play the game, learn how to scale up to make progress go faster. That goes for robot stuff too. And if you're trying to scale up bot-based construction and you're using personal roboports for that, you're doing it wrong.

BlakeMW wrote:If the ability to hoover-charge bots is taken away, it should only be after resolving the intractable issues with using personal roboport

What "intractable issues"? You haven't really made the case that there are any.
Yes, you can't run thousands of bots 24/7 over the whole map using just your personal roboports. None of those are "intractable issues". They are gameplay limitations placed on portable/wearable devices, gameplay limitations that provide the challenges/trade-offs I listed in my post, and that help keep full-sized roboports relevant for construction tasks.

BlakeMW wrote:I think perhaps the game would be better if there was an efficiency boost to go along with the speed research, so one way or another robots with a higher speed multiplier use less power per tile travelled allowing them to be more productive per unit of energy.

While I am not entirely opposed to the idea of having some bot efficiency tech, at *best* it should only slow down a bit the rate at which bot power usage (energy/time) goes up for each speed bump - not stop it. Too much of this (making this new tech overpowered) would just be another way of removing any need for efficient use of bot charge, so it would have to be applied judiciously to avoid destroying gameplay.

If you were looking at a new research related to bots and charging and dealing with high bot speed, it would seem to me to make more sense to improve bot recharge rate (assuming recharge rate actually does become an issue once you're sufficiently far into space science bot speed). Even an infinite recharge rate wouldn't remove the need to use bot charge efficiently so it doesn't destroy gameplay.

Koub wrote:there should be a rethinking on how construction bots use their energy, and how this energy is provided to them, especially in the case of early modular armor configuration

Why? What problem are you trying to solve?
Early bots are limited. You have to use them judiciously. That's a *good* thing. (Maybe they should come a bit earlier in the game, but they should still be limited.)

Klonan wrote:I don't see what this adds to the gameplay

Then you didn't read/understand my post. There are gameplay mechanics (challenges/trade-offs) present that people are completely missing out on because rather than playing the game they are cheating their way through it.

If bot charge cheating is good, why not just build in cheats for everything else too? Then people can just skip the whole game in its entirety.

Klonan wrote:just adds frustration to those early moments when you only have solar panels

Only if you're doing it wrong. I have no such frustrations. I don't use the cheat. Solar + batteries is useful, it gets lots of use, and it's actually the most interesting part of the whole game in terms of personal roboports because it requires the most consideration with respect to doing things efficiently and not wasting bot charge on tasks that don't really need it.

I definitely agree that bots should only be charged via an actual charger.

I think a better solution would be that bots lose all battery when mined, so that there's no technical challenge of adding the data to the bots in inventory and you don't have to worry about stacking issues.

If you want it there's a Mod called Picker Extended that does this amongst other things. personally I find not being able to mine robots out of the air annoying in about 8 different ways but if you want it there's mods for it

JohnyDL wrote:If you want it there's a Mod called Picker Extended that does this amongst other things. personally I find not being able to mine robots out of the air annoying in about 8 different ways but if you want it there's mods for it

You'd still be able to mine them, it just wouldn't magically refill their batteries.

-1 from me.

As already said it doesn't add anything interesting to the gameplay. It just makes attempting to use personal constructions bots before you get fusion power even more frustrating for the player.

For me the current behaviour is adds to Factorio's Quality of Life.

It would be nice if one could recharge from the power grid so personal roboports can still be used pre fusion generator. Also this would make it even more important to have personal roboports handle ghosts in order of distance. Because with the inability to recharge by cheat the longer flight path would costs again in longer recharge time.

Zavian wrote:-1 from me.

As already said it doesn't add anything interesting to the gameplay. It just makes attempting to use personal constructions bots before you get fusion power even more frustrating for the player.

For me the current behaviour is adds to Factorio's Quality of Life.

So then allow charging from the grid or buff solar panels or some other mechanic to make construction bots prior to fusion reactor feasible. Also not everything needs to be easy early, it's ok for some content to not be easy until further research/resources are available.

Recharging bots by mining them seems like a pretty clear break of mechanics.

mrvn wrote:It would be nice if one could recharge from the power grid so personal roboports can still be used pre fusion generator."

After a long hiatus from the game, I've just been playing an 0.16.16 game, and I have to reiterate how eminently usable personal robots are with modular armor. Besides some (most?) players using bot charge very inefficiently, I think a common mistake is not using enough portable solar panels. For my current game I started with 2 MK1 batteries and filled the entire rest of the armor slots with solar panels. Once the MK1s were full (takes less than one game day) I popped out 4 solar panels and put in a roboport. (The robotport will suck a lot of battery at that point to fill its internal charging banks.) I used it like that for a while before popping out 4 more panels and putting in night-vision. And that's it - don't try cramming any more stuff into modular armor and you'll have 130kW of power (less any night-vision power) to keep robots charged.

My experience this game? I used my robots as much as I wanted, 24/7. My batteries stayed in the green the whole time except once. (It dipped down to about 50% charge while building a 40 assembler green-circuit build, built in 10-assembler chunks (could have used smaller 5-assembler chunks to be more efficient) but in rapid succession so there was no recharge time in-between chunks. I of course moved my character from chunk to chunk as the build progressed to minimize the average distance the bots had to travel. A lot of the time my builds were material-constrained so I was building chunks as materials came in which gives batteries more time to recover, but in this case I had saved up materials for the whole build.)

Trying to use even one exoskeleton at this point in the game is a huge mistake - it not only sucks down 200kW when you use it, but it takes up 6 slots meaning you give up about *half* your solar power 100% of the time. At this early stage of the game, use a car to get around fast. (Even late game the car tends to be faster, and trains when applicable.)

For early game, don't use shields with a roboport. (Shields cost a constant 40kW each due to displacing solar panels.) And don't swap roboports with shields to switch between "build mode" and "combat mode" - they both lose all their charge when you do that, and the roboport especially consumes a *lot* of charge when you put it back in which means a *long* recovery time. Use two sets of armor instead.

Once you get power armor, you can put in even more solar panels, and also at that point you can afford to have more roboports and some shields if you want (e.g. 2 each). You still shouldn't use exoskeletons at this point (because they're still too power and space hungry - wait until fusion power). You can be somewhat more "care-free" with respect to robot charge, e.g. regularly constructing and destructing larger chunks at a time (e.g. chunks in the 24x24 tiles range rather than 12x12 tiles). Of course, as always, you can manage your charge vs laziness and do larger chunks when you're at full battery power.

One thing to keep in mind: While at first thought it may seem that constructing/deconstructing in smaller chunks would be a "useless annoyance", it's anything but. It doesn't just save bot charge - it saves player time as well. That's because you're not spending nearly as much time waiting for bots (very, very slow bots at the early modular armor stage) to travel long distances.

OK, now that I've said all that... I do think it would be OK if personal robots could smartly choose to make use of full-sized roboports for recharging when they happen to be nearby, but I don't know how useful that would really be (since you could then just be using the full-sized roboports instead since they're right there). I don't think it makes sense for robots to charge directly from the grid itself without some kind of robot charging device (e.g. either a roboport, or a dedicated robot charging device like Bob's mods has). What there could be is an "armor charging station" that is attached to the power grid - the player would walk up to the station and "jacks in" and everything in his armor at that point (roboports, batteries, shields) gets fully charged at its max rate (limited by what the power grid can supply of course). As much sense as that makes in terms of "game world reality", though, it really doesn't do much in a gameplay sense. That's because if you're going to have power and special armor charging entities available, why not just have power and full-sized roboports available (and use them)? I guess it would affect gameplay more if the "jacking in" was a built-in feature of the armor (maybe after a tech upgrade?) rather than requiring an entity (armor charging station). Note that it would be somewhat self-defeating if the charging ability were provided via an in-armor module (especially a large in-armor module) because then it too would be displacing portable solar panels.

NotABiter wrote:

mrvn wrote:It would be nice if one could recharge from the power grid so personal roboports can still be used pre fusion generator."

My experience this game? I used my robots as much as I wanted, 24/7.

...snip...

Agree wholeheartedly, modular armor is perfectly capable of running roboports. You just need a proper loadout of internal gear to support what you want to do.

I have considered a third set of armor with exoskeletons for speed. But I agree that a car is almost always a better option. And I’m already juggling two sets of armor, don’t really want to juggle a third.

I even use two sets of armor at Mk2. I’m always running around with a utility frame and a war frame.

I have to admit I’m a little confused why only certain parts of the game seem to need a “balancing treatment”. Either the whole game should be balanced, or balance doesn’t really matter overall.

mrvn wrote:Also this would make it even more important to have personal roboports handle ghosts in order of distance. Because with the inability to recharge by cheat the longer flight path would costs again in longer recharge time.

I agree some supporting feature(s) would be nice, though the form I would prefer is an easy-to-change in-game setting for "reduced active roboport range" for personal robots. I.e., let me set the range beyond which the bots won't respond to ghosts at all. (The devs could actually implement both, both user-defined range and sorting for an even better solution. But reduced range is a complete solution by itself while sorting is not.)

For many things (long walls, train tracks, any large builds) this would significantly reduce the amount of "twitch" involved with efficiently using robot charge. Instead of placing wall blueprints one small chunk at a time as I walk along side, I could plop down one giant blueprint and then just "walk the wall" as the robots do their thing, all while achieving the same (or possibly even better) bot charge efficiency.

The reason why sorting by itself doesn't solve the problem is that, once bots are assigned to any close stuff, other bots will still run out to maximum range, wasting lots of charge in the process. That problem might be avoided in some cases (if you can just stay exactly with the bots as they work - note that in this case you have to follow the bots rather than in my solution where it's more the bots following you). But bad things happen if, for example, you need to go do something else (and all your bots then run off trying to complete stuff in the other direction while you are now running away from them). It can also be an issue if water/trees/buildings/pipes/etc. are blocking your path so you have to walk around something and can't keep up with the robots which, once again are flying away from you at unnecessarily long range. It's also an issue when you run out of supplies, go and get some more, and then head back - your bots will leave you to go construct things while you are still too far away from the target - and with early bots it's especially sad because they will quickly fall behind you because even without exoskeletons you can run faster than they can. There's also a problem if the build density is "too low" - e.g. a series of large electric poles with a couple lasers next to each one will result in robots heading off for the *next* pole as soon as it's in range, every single time. There's probably other cases I'm not thinking of at the moment -- I believe that being able to limit the range would do a much more predictable and reliable job of putting the player in charge of the bots, likely resulting in less player frustration/tedium and fewer inefficient bot charge usage cases.

Bots should not leave when you move faster than them. Or maybe even more than half their speed. It's beyond stupid for bots to leave you while you are racing around in a train. It's still annoying when walking somewhere. It's deadly (for the bots) while running away from aliens.

Since bots are overpowered anyway, i would vote for a "if mined while not charged" == "charg on next start"

I realized that a simple sorting strategy (as in "closest first") even as an "add on" to my idea of limiting the range, is not optimal. What you really want (in addition to range limiting) is for the bots to take the average motion of the player character over time into account. You then want to compute estimated paths in 3 dimensions (2 spatial dimensions and one temporal dimension) to determine total estimated travel distance for each path. Finally, you want to compute this not only for "if the bot were to take off now" but also "if the bot were to take off some time in the future" (e.g. 10 ticks from now) so you can see how each path's estimated travel distance is changing over time. Once you have that information, you want to assign bots ASAP to paths that will be longer in the future before they become so (fastest growing paths first), you want to deprioritize paths that will be shorter in the future (and ideally within that set prioritize paths closest to flipping from shrinking to growing), and you want to completely defer paths that will not only be shorter in the future but which will become so much shorter so fast that that rate exceeds the bot's speed (or, as you say, maybe even defer if that rate is only 50% or 75% or something of the bot's speed - maybe that would be another user setting).

For example, if the player is walking along a long ghost wall (which, depending on how thick the wall is and how many turrets/tracks/etc. it has, may be "walk and stop, walk and stop" to allow the robots to keep up, thus average rather than instantaneous motion being key), bots should make sure the trailing ghosts are handled before the upcoming ghosts even if some of the upcoming ghosts are closer, because the player character is getting closer to the upcoming ghosts over time but is getting ever farther away from the trailing ones.

Finally, the range limiting should be dynamic rather than static - i.e. you should also defer if the total path is just too long. That should (at least mostly) eliminate the problem of bots jumping out of high speed trains (because even if the ghost is close now the estimation will show that the bot's return path will be massively long [infinite actually]).

On the other hand... I think some people just think it's funny when bots do stupid stuff.

mrvn wrote:Bots should not leave when you move faster than them. Or maybe even more than half their speed. It's beyond stupid for bots to leave you while you are racing around in a train. It's still annoying when walking somewhere. It's deadly (for the bots) while running away from aliens.

Absolutely. I would not object to removing the charging cheat so long as bots (1) charged while in storage, not floating around you, (2) did not leave your inventory to do work if you are moving faster than they can move.

Simple, effective.

Unfortunately, from what I've seen recently I suspect that devs think that robots taking off even when you are going at max train speed is a "fine behavior". Aaargh, frustrating. I am so tired of forgetting to remove my roboport and losing all my bots on the way to the worksite, and worse... since they will try to get to you in a straight line (and my rails tend to loop around a perimeter) I have no way to find where they are, so I can't even go pick them up.

zOldBulldog wrote:

mrvn wrote:Bots should not leave when you move faster than them. Or maybe even more than half their speed. It's beyond stupid for bots to leave you while you are racing around in a train. It's still annoying when walking somewhere. It's deadly (for the bots) while running away from aliens.

Absolutely. I would not object to removing the charging cheat so long as bots (1) charged while in storage, not floating around you, (2) did not leave your inventory to do work if you are moving faster than they can move.

Simple, effective.

Unfortunately, from what I've seen recently I suspect that devs think that robots taking off even when you are going at max train speed is a "fine behavior". Aaargh, frustrating. I am so tired of forgetting to remove my roboport and losing all my bots on the way to the worksite, and worse... since they will try to get to you in a straight line (and my rails tend to loop around a perimeter) I have no way to find where they are, so I can't even go pick them up.

I wouldn't mind removing the roboports, that could be scripted, if they would retain their charge. Early on when you have solar cells it takes forever to charge up the roboport and you loose all that charge when you remove it.

The only way to go is to remove the bots while traveling. E.g. dump them in the cars trunk while driving. Not so easy to do on a train because they get unloaded easily when pausing at a station.

BTW, I really appreciate this thread. It clued me to the existence of the cheat and I have been abusing it since.

It removes a lot of the annoying drudgework that would otherwise make a 2nd or 5th play-through annoying. Long live the cheat!!!