d88402 No.13678048
i started doing this again and got the farthest I've ever gone, one deathworld, pretty much all the shit before the rocket. finally get bots, sweet, my up concerns will be gone, bots are supposed to be better than belts, after carefully managing belts all game long to never drop below 60 ups. spend two hours replacing the shit in the base with bots.
60 ups -> 30ups.
game closed.
what the fuck. is it even possible in this game to finish the game without the game shitting itself on a computer that doesn't have a 5ghz processor?
ca42e7 No.13678129
>replacing literally every belt with robots
>have 1k+ robots flying around at all times
>still have reasonable fps
>make thread to bitch
kys
d88402 No.13678292
>replacing literally every belt with robots
that was clearly a mistake
at what point are bots a better choice than belts? for long distance it's obviously trains. i took a massive ups hit moving plates from bots to a belt, so what, anything high throughput it's better to use belts if your not using a train?
d88402 No.13678298
>>13678292
"moving plates from bots to a belt"
i mean to say that the other way around
3d82e1 No.13678517
Last time I played building a shitload of bots felt like cheating.
d69553 No.13678519
>>13678292
I just use belts, I've never found trains to be worth the tediousness of setting them up.
d88402 No.13679013
everything i've ever read on the forums about ups has said "just use bots, duh" "never use belts" "always go full bots". ya bullshit. my factory was running along smoothly around the 60 ups threshold until i replaced my main copper line, 4 red belts, with bots.
clearly bots are shit for medium range. and for short range? you have stack inserters at that point so i'm failing to see the purpose of bots since they aren't the magic ups cure and perform worse than belts atleast in the situation i used them.
d88402 No.13679032
i think another part of the problem might have been i put a shitton of requester and provider chests down, pretty much at every machine. so alright, next time i'll lump it all together into single chests, but then for what, if it takes 100 bots to match 1 yellow belt fuck i might as well just use underground yellow belts.
this bots meme is a lie
30c766 No.13679138
>>13678048
Honestly I think the best way to utilize bots is by replacing trains with them. The rail system is spectacularly broken and unnecessarily tedious, just connect the outposts that you can to the logistics network and have the bots do the work. Once you have them fully upgraded each bot is like a damn semi-trailer in their own right plus they go at a pretty decent speed. Using them for your production is honestly a pointless waste of time, by the time you're able to reliably manufacture them your base is very well established and needless to say demolishing the entire fucking thing just to rebuild it with bots in mind is completely fucking arduous and just not at all worth it. Unless of course you design your base in the beginning with bots in mind, if you can handle that clusterfuck then sure. Only useful automatons in the game are construction bots, those are the ones that save you a shitload of time. Another idea is just build a personal roboport and have them deliver materials whenever you need them
605b5b No.13679168
If you can't even launch a rocket on Deathworld without lag then you've really fucked up somewhere. I'd need to see your base to see what's wrong. Full belts should be fine for that.
The trick to bots is that you can't make one huge network of shit, you need to make small networks. Large networks causing bots to travel long distances causes way more lag and way more loss of throughput than a bunch of small ones.
Trains are superior to everything else, it's very simple to get them working. Just lay down a 2-way rail system and make a 2 Locomotive/8 Cargo setup.
Ohh, the one thing that CAN fuck up your UPS is if you have tons of pipes. The Fluid calculations are horrendously CPU expensive with expansive pipe networks. Barrel all fluid and send them around that way. It's pretty easy to integrate into a bus, just set one lane as a return for empty barrels.
51091d No.13679356
d88402 No.13679373
>>13679168
this is where i fucked up. it's all one big network. i loose 30 ups as soon as they start unloading plates from a train
a646dc No.13679665
>>13678129
I use belts exclusively and still get 60 fps on a five year old machine launching a rocket every couple of minutes. OP must have some BR tier laptop bogged down with porn dialers or something.
>>13678292
>at what point are bots a better choice than belts?
Takes less planning I guess. It's pretty nice that they can both sort, support multiple sources and destinations AND can throw their full bandwith at whatever is being drained, but I feel like it takes the fun out of it tbh.
605b5b No.13679712
>>13679373
You can definitely compactify that a lot, it would save you lots of bot time just to smush it together. You should really be using beacons. Put a line of beacons, a line of assemblers or furnaces, then another line of beacons, etc etc. Productivity modules go in the assemblers/furnaces and speed modules in the beacons. This way a single assembler can run at about 500% speed. You can also dump a ton of speed modules in and around your oil to give you more when they are running dry.
Also train tips: You can give all of your train stations the same name (i.e. Ore Fill) rather than numbered and they'll figure out which one to go to. I don't really recommend 2-headed trains since it's way less efficient (locos weight 3x that of cargo sections and are deadweight in reverse), just lay two tracks parallel to each other.
a646dc No.13679761
>>13679712
>Also train tips: You can give all of your train stations the same name (i.e. Ore Fill) rather than numbered and they'll figure out which one to go to.
Doesn't this scale like shit though? Wont they all head for the nearest?
605b5b No.13679819
>>13679761
They won't head for one that another train is already at/heading towards.
d88402 No.13679825
>>13679712
thanks, instead of rage quitting this map i really should just pull just grab up all these bots and massively rebuild this base, with beacons, like you said. what i was trying to do was incrementally do this and ended up with both a main bus and this 1000 bots in all directions lagging everything with 50 requester/provider bot chests everywhere.
a61231 No.13679980
According to official information, belts actually hit your UPS far more than bots (something they're continuously reworking to try to mitigate) because it has to simulate the movement of every single item on the belt, instead of just a bot that carries several items. The items on a belt are actually affected by physics (they are really OBJECTS, not ENTITIES) while bots are not, as they fly over everything.
The other things that cause massive UPS drop are biter spawns and pollution updates. So if you have a massive sprawling base putting out tons of pollution that is spreading hither and yon, that may be your problem. However, since you mentioned "deathworld", I suspect it's the biters. When you leave an area so the chunks get un-loaded, the biters that were chasing you in that area freeze in place, but get their movement data stored. This can add up to getting single-digit UPS if you aggro them enough but don't actually kill them.
As for throughput, I've never found a use for bots to solve that issue. There isn't much that can't be solved by just liberal applications of MORE BELTS.
914956 No.13679990
>>13679980
Belts have been greatly optimized in 0.15 so your info may be somewhat out of date.
342b5f No.13680054
>>13679373
That network is literally nothing, your problem is somewhere else.
a646dc No.13680058
>>13679990
I thought the improvements were delayed until 0.16?
c49ef5 No.13680062
When will this game ever so on sale? I keep waiting for a sale for almost two years now and nothing.
a61231 No.13680073
>>13679990
It's possible. I haven't gotten around to playing 0.15 yet, because I'm a bit scared that I'll have to re-learn everything in regards to how to balance different types of production, because all the science packs were changed.
>>13680062
Anon, plz don't. "Discussion" over this destroyed the last Factorio thread.
a646dc No.13680079
>>13680062
They've said it wont go on sale until it's at least done. Just give them the measly 20 bucks, goy.
914956 No.13680093
>>13680073
>>13680058
Can't find any patch notes about belts being optimized in 0.15 so I must have misremembered, sorry.
>>13680062
They've pretty explicitly stated that there'll be no sale exactly because otherwise everyone would just keep waiting for a sale. Pirate it if you don't want to put the 20 bucks down upfront.
a646dc No.13680094
>>13680073
All the science packs beyond green are different now. None of them are as complicated as blue, though yellow comes close. It's just the burden of setting all that shit up
Also holy shit military science eats a ton of iron
914956 No.13680125
>>13680094
You just remember blue as complicated because it was such a large leap in complexity from green, producing yellow in good quantities is definitely harder than the old blue and space science is technically the most complex science package by magnitudes. In general you need much, much more raw materials to keep science running now, which means larger factories, which means either spaghetti lines or stepping up your scalability planning, and that's both great from a factory tutorial perspective and harder than the old science used to be.
342b5f No.13680141
>>13680094
>>13680125
It's much better now IMO, previously you could just set up your research and do everything without thinking about it, but now you may need to actually make some decisions.
a646dc No.13680187
>>13680125
To be honest, it was probably the need to set up both oil production followed by both plastic-> red circuit and sulfur->batteries chains, all at once)
d88402 No.13680221
>>13680054
>>13679980
i think your right. i tried to benchmark this again, and it's not nearly as bad, but i'm doing it on another save i made a minutes afterward in the base. the previous save that was totally shitting the bed was in the middle of a biter field. i think something about that save and a ton of biters being active fucked it all up.
d88402 No.13680229
still this base obviously isn't optimized at all, i need to be smarter with bots and i was in the middle of trying to convert to smelt at the base to smelt on site at the ore field
605b5b No.13680321
>>13680229
Smelting on site is a waste, you end up needing to tear down and rebuild every time a field runs out. You also run into problems when your field is 1/2 run out and lots of your smelting capacity is unused because the miners aren't filling a full belt to feed the smelters.
Better off designating a central smelting area and dumping all of your ore there.
46f71f No.13680407
>>13680229
The one rule of this game is always prepare for expansion.
Instead of smelting on site just make a smelting area that all your ore gets sent to that is larger than what you need.
https://dddgamer.github.io/factorio-cheat-sheet/
Find out how many iron belts you need and then make more.
If you can, make beacons, or at least prepare for beacons when making your large smelting area.
914956 No.13680511
>>13680229
>>13680407
Just automate absolutely everything you may ever need, or at least build an all-purpose assembly area that you can configure to produce like 100 assembling machines or fast splitters or something at a time. I also like to keep a stock of commonly used intermediate products like green circuits, gears and such just to speed up handcrafting or temporarily speed up some assembler short on materials.
0afc93 No.13681406
I have 16 full express belts of copper just for green circuit production in two different facilities being fed by a constant rotation of trains and it is not nearly enough.
I have only about half the green circuits I need, and I am struggling to expand fast enough to feed my refineries. I've already stuffed production modules in everything I can. I don't know what else to do.
SEND HELP (and copper)
0afc93 No.13681429
>>13678519
In vanilla that's definitely the case- most people I know play with RSO, which makes resources richer, spreads them out more, and makes resources further away richer, I'm not sure what of that you can achieve by using a railworld now.
a646dc No.13681453
>>13681406
Deconstruct entire base and drive four hours in one direction. Patches should be bigger out there.
4096cd No.13681483
>>13679990
Belts will likely never be as fast as bots. Bots only need to deliver their item to the location at a time, so other than just for visuals, don't really have to do anything at all. They could be implemented by a timer.
605b5b No.13681634
>>13681406
How big are your trains? You can always make trains bigger to increase throughput. My largest unloading station is 3 locomotives and 12 cargo wagons.
>>13681483
IIRC the devs changed belts so that a max compressed belt of items is treated and moved as a single item. So if you have 1000 ore on a long uninterrupted belt it requires no more CPU time than if it was 1 ore on the belt.
914956 No.13681727
>>13681429
Default railworld preset more or less works like that, but with less large starting patches compared to RSO, and biters won't build new bases at all. Might want to disable the last one if you think it'll be too easy, kind of tends to remove the challenge.
a646dc No.13681750
>>13681727
It's not like biters add serious challenge without mods anyway. Once you build wall you can basically forget about them until you run out of space.
4a6438 No.13681791
>>13681750
Speaking of mods for the biters, any of you guys got any experience with the Rampant mod? Does it make for more interesting gameplay?
605b5b No.13682164
Anyone tried cargo wagons as a bus?
Was fucking around with bobs/angels, I calculate around 400 items/s or 24,000 items/m throughput. And you can set filters so that multiple items can coexist evenly, theoretically 80 separate good types can be bussed on one lane.
0afc93 No.13683288
>>13681634
How many belts do you unload that into?
d88402 No.13684824
>>13681750
i think i'm going to abandon deathworld for this very reason. it isn't really more of a challenge now that i've expanded and gone this far. it only makes it a pain in the ass to expand, it just takes longer, and now it apparantly causes a bunch more lag. i've heard the lag goes down if you expand out well past your pollution cloud, which in deathworld would take forever, without any real risk or challenge, just time. that's cool i guess but my processor can't handle it, so i'm starting over on marathon to stay used to expensive recipies but hopefully be able to push walls outside of the cloud without saying fuck my life every time.
d88402 No.13684844
>>13682164
i wonder if the ups hit is for the inserts will be better than belt/splitters? looks interesting though. and you could definitely stick more items on the same bus. wouldn't the throughput be limited to 2 stack inserters?
bb9211 No.13684853
a646dc No.13685111
>>13684853
Never ever. They want to make it to 0.16 before they launch. Apparently there'll be an artillery wagon and some other minor goodies plus more optimizations, but it's basically done.
d88402 No.13685145
>>13685111
i've been planning on buying it, but i keep procrastinating. I'm going to be pissed if I don't buy it before 0.16 and they up it from $20->$40 on release
605b5b No.13685845
>>13683288
12. Don't have a working example to show you since I lost all my saves when I bought the game, but here's what my blueprint looks like. I find it quite aesthetic.
>>13684844
Due to the fact that the cargo wagon can extend over inserters you can get 4 stack inserters in vanilla and 6 fast stack inserters in bobs mods (though it looks weird having them at 3-length). Vanilla throughput is about 100 items/s. Slightly better than two blue belts, but it would be best as a combined bus for dozens of items that you only need a few of. e.g. you could barrel every fluid and send them down a single wagon bus line, and if you used filter inserters you could send empty barrels back up the line by having one inserter facing the opposite direction.
I think inserters are supposed to be really well optimized, especially when going from container to container.