Page 1 of 2

Belt vs Bot computing

Posted: Sat May 07, 2016 7:10 am
by CMH
Just trying to build a megabase, and wondering how to optimize it at the end game.

Full belt would be quite disastrous CPU wise. Full bot works much better, that much I know.

Because megabases tend to be, well, big, I was wondering if I had some belts bringing stuff closer (I'm thinking plates) this would reduce the distance bots take to travel to where stuff is needed (but not exactly long enough that a train would be useful).

The question however is how many bots required must this reduce for it to be worth doing? Ie: having a 100 length belt to reduce the number of bots needed by 300, is that worth it? 500?

Given a blue belt transfers about 40 items/sec, one belt could theoretically reduce the need for 10 bots (4 item stacksize per bot). Maybe 20 since the bot doesn't need to double back.

Any thoughts?

ps. I know the best way to optimize is to reduce the distance the bots need to travel and the best way to do that is to move assemblers closer. Well, this is assuming you can't get any closer.

Re: Belt vs Bot computing

Posted: Sat May 07, 2016 7:15 am
by DaveMcW
What you really want to do is minimize the distance between train stations and final assembly.

Don't build giant, distant areas for each product. Instead, build a mini-factory that contains a complete recipe chain and duplicate it many times.

Re: Belt vs Bot computing

Posted: Sat May 07, 2016 7:20 am
by CMH
I did think of that.

But then I thought about the headache which is how to ensure all the different mini factories get their needed plates without starving the rest...

I suppose that's logistics for you, and part of the fun of factorio :D

Re: Belt vs Bot computing

Posted: Fri May 13, 2016 11:06 am
by Mackowatosc
CMH wrote:I did think of that.

But then I thought about the headache which is how to ensure all the different mini factories get their needed plates without starving the rest...

I suppose that's logistics for you, and part of the fun of factorio :D
The answer is, even more smelting, possibly more than one smelting area :)

Re: Belt vs Bot computing

Posted: Mon May 16, 2016 5:15 am
by cjmarsh
I can't give any specific numbers but one thing to keep in mind is that more bots comes with the cost of more roboports to keep them charged. Otherwise they have diminishing returns since they have to wait in line for a recharge near the area they're needed. Personally I would be willing to use a belt over a bot at a 1:1 ratio, 100 belts if it saves 100 bots. That assumes all other things being equal though, that the belt is easily placed and the throughput won't be an issue, so there's a limit to how often belts are better than bots.

Re: Belt vs Bot computing

Posted: Mon May 16, 2016 9:10 am
by Neotix
I used bots once few years ago and decided to not use them ever again. For me they're annoying and looks ridiculous like flying bugs. My entire factory is based on belts. I use only construction bots with personal roboport to handle blueprint and sometimes one roboport with bots to move content from one chest to another when I need to move part of factory with buffer chests and I don't want to do in manually.

Re: Belt vs Bot computing

Posted: Sun May 22, 2016 8:35 am
by CMH
Well, the problems with a belt-based factory is that at a certain point your computer can't handle the calculations and the game slows to a crawl and becomes unplayable. The good news is that the vast majority of people will not build a factory big enough.

This thread is regarding a factory that reaches that point, and hoping to extend the size of the factory further by hopefully being able to (greatly) reduce the number of bots with a (hopefully) minimal number of belts.

Re: Belt vs Bot computing

Posted: Sun May 22, 2016 1:30 pm
by AutoMcD
I think it's a matter of quantity.
Belts for the main grind of ores/plates. For shipping final products around it's bots.

Re: Belt vs Bot computing

Posted: Sun May 22, 2016 7:41 pm
by siggboy
OK, so you're talking about an actual megabase. There are not many items that need to be shipped around in huge quantities, put those on belts. Chain underground belts if necessary (I've heard it helps with UPS).

Everything else goes into bots, all the medium-volume intermediate products included.

For me the rule of thumb is that I don't ferry bulk ware around with robots if the area is larger than an ore outpost or train station. It's incredibly difficult to model, not least due to the dynamic/semi-intelligent manner in which the bots operate.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 6:22 am
by CMH
It actually isn't all that hard to ship huge quantities of stuff, as long as it isn't a particularly long distance. All comes down to how you arrange your base. I'm now thinking it might be better UPS wise to create multiple small bases to create bulk items (copper plates?) just to cut down on the distance required to transport the product.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 6:55 am
by Bart
Neotix wrote:I used bots once few years ago and decided to not use them ever again. For me they're annoying and looks ridiculous like flying bugs. My entire factory is based on belts. I use only construction bots with personal roboport to handle blueprint and sometimes one roboport with bots to move content from one chest to another when I need to move part of factory with buffer chests and I don't want to do in manually.
Wow. Manually repairing walls and turrets is boring as hell.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 7:47 am
by MrDoomah
Bart wrote:
Neotix wrote:I used bots once few years ago and decided to not use them ever again. For me they're annoying and looks ridiculous like flying bugs. My entire factory is based on belts. I use only construction bots with personal roboport to handle blueprint and sometimes one roboport with bots to move content from one chest to another when I need to move part of factory with buffer chests and I don't want to do in manually.
Wow. Manually repairing walls and turrets is boring as hell.
Just build the walls 5 thick and replace the destroyed once in a while? :P

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 12:26 pm
by CMH
Bart wrote:
Neotix wrote:I used bots once few years ago and decided to not use them ever again. For me they're annoying and looks ridiculous like flying bugs. My entire factory is based on belts. I use only construction bots with personal roboport to handle blueprint and sometimes one roboport with bots to move content from one chest to another when I need to move part of factory with buffer chests and I don't want to do in manually.
Wow. Manually repairing walls and turrets is boring as hell.
You got it all wrong, repairing walls and turrets is the ultimate endgame of factorio. That and ferrying plates all over the map so you don't waste precious UPS on belts OR bots.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 12:38 pm
by Shokubai
CMH wrote:It actually isn't all that hard to ship huge quantities of stuff, as long as it isn't a particularly long distance. All comes down to how you arrange your base. I'm now thinking it might be better UPS wise to create multiple small bases to create bulk items (copper plates?) just to cut down on the distance required to transport the product.
I'm working on a non-biter game currently where I modularized my entire base as soon as i got trains.

A couple rules of my game i set for myself.
Bots for personal use and construction only.
Steam Power Only.
Mining is done directly into train
Smelting is done directly from train
Each major product is built at a separate facility. Supplies and products are transported by train.
I'm using an oil transport mod to transport Lubricant and such to other facilities. Cracking is done at the main oil plant.
Insane resources mod makes this game type a bit simpler since I don't need as many raw resource stations.
The only exception left in my modular station game is what i call my basic supplies base. It's where all of my construction supplies are built. Basically everything but science with a station for a "Construction" Train as a base for all my crazy buildng.

Currently I'm building with the idea of a 4x(4/6/11) base. So I want to fully supply 16 Red Sci, 24 Green Sci, and 44 Blue Sci.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 1:23 pm
by siggboy
@Shokubai

Do you use any sort of mod that makes research more expensive? If not, you should try it (I found the "Expensive Research" mod to be quite balanced).

Actually I'm in a no-biter game as well, and I'm planning to do something very much like your game. I like the rules that you've given yourself; it's not easy to come up with an interesting scenario that way without making rules that are very forced and arbitrary (such as "have two radar ranges of space between factories").

One big reason why I want to play it that way is to test and show off my train scheduler (it's kind of a "logistic train network").

How do you define "major product"? Are those only rocket parts or anything more complex than green circuit? Anything with more than two ingredients?

How do you define "separate facility"? Did you make any rule for how "separate" they have to be or does that not matter as long as they are not connected with belts to other "facilities"?

I'm not really ready to go "belt only", it's possible to have an interesting game with a mixture of belts and bots. Bots can also be very useful in mining outposts, so you wont have to create a belt system to get the ore into the train. Building mining outposts is a chore, I don't want to make it even more painful by restricting myself to belts.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 1:34 pm
by Shokubai
siggboy wrote:Do you use any sort of mod that makes research more expensive? If not, you should try it (I found the "Expensive Research" mod to be quite balanced).
No but i use victory research which adds an insane amount of additional research to the game.
siggboy wrote: How do you define "major product"? Are those only rocket parts or anything more complex than green circuit? Anything with more than two ingredients?
It's probably easier to exclude things than tell you everything I think of as major.
Things like Iron Sticks I think of as minor items that i make at a facility where they are used directly.
Major products are anything you might bus. Red/Green Circuits, Gears(debatable), Steel, Iron/Copper, plastic, Sulfur...
Major really means "used by lots of stuff" (or in large quantity) rather than "complicated product".

Because of my rule set, I am not making say...Steel where I'm mining Iron. I have to bring in Iron Ore. I have to plan my Drop off and pickup stations and lines accordingly.
siggboy wrote:How do you define "separate facility"? Did you make any rule for how "separate" they have to be or does that not matter as long as they are not connected with belts to other "facilities"?
Geographically disconnected. Only connected to other bases by power and rail. Since there are no biters I pick locations that are a good distance away from each-other arbitrarily.

siggboy wrote: I'm not really ready to go "belt only", it's possible to have an interesting game with a mixture of belts and bots. Bots can also be very useful in mining outposts, so you wont have to create a belt system to get the ore into the train. Building mining outposts is a chore, I don't want to make it even more painful by restricting myself to belts.
My idea still works with bots and I gave it a lot of thought that i could do the same thing either way. Honestly with the distance between bases and how modular everything is...the belt limitation is really minor. It isn't the same as the same idea in a large Bus base.

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 1:49 pm
by CMH
I play no biters as well. Clearing them kinda gets boring since nothing really changes. So I removed them.

Instead I went the way of Bob's mods to add complexity. But I found that finishing it doesn't really require bob's mods, so I made a mod to require the top tier stuff from bob's mods to finish.

Can still do a megabase with it, just requires a heck lot more :D

For giggles, here's where I uploaded it (in this forum)

p.s. Although I'd appreciate feedback, my modding skills probably aren't good enough to deal with making it compatible with other mods. So this works with Bob's Mods, and a few others, but I can't guarantee anything.

viewtopic.php?f=51&t=16814&start=20

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 5:01 pm
by AutoMcD
how do you deal with needing artifacts?

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 5:25 pm
by Shokubai
AutoMcD wrote:how do you deal with needing artifacts?
Alien Science

Re: Belt vs Bot computing

Posted: Thu May 26, 2016 6:25 pm
by siggboy
CMH wrote:Instead I went the way of Bob's mods to add complexity. But I found that finishing it doesn't really require bob's mods, so I made a mod to require the top tier stuff from bob's mods to finish.
Yes, this is great, it's like the "Harder Endgame" mod for Bob's Mods that somebody else made. I've used that during my Bob's Mods campaign, but never got around to finish that playthrough. Maybe I'll have a look at yours for my next game. Yours does require God Modules though, and they're a no-go for me, so I would have to change that.

The problem I have with Bob's Mods is that while it adds a lot of desired complexity, it's very unbalanced in some areas, and it adds way too many things that simply are not needed and add nothing to the game except being overpowered.

I think we need a complete overhaul of Bob's Mods -- like a "Diet" version of Bob's Mods -- that consolidates the many good ideas and removes all the cruft.