Factorio Forums

dood wrote:Blue belts are nowhere near that level.
You secured a big chunk of iron ore and some oil, you're good to go and you should take the plunge earlier rather than later.
When you start getting t3 assemblers sounds about right. Nowhere near endgame.
Fun fact: changing the t2 assemblers and belts of a build to t3 makes the entire thing 50% faster without starving the belts.

Why? What do blue belts offer you in the midgame? Instead of running 4 blue belts anywhere, run 6 red belts. 6x11.5 == 69 iron per tile of 6xRed belts. While 4x31.5 == 126 iron per tile of 4xblue. And when you see the price of 12x1.5 == 18 iron per tile of 12x Yellow Belts, you can start to see the serious cost-savings advantage of staying on earlier game tech for longer periods of time.

Its not like you spend any player-time placing belts in the midgame. Construction bots lay out the designs. So it doesn't matter if you're using yellow, red, or blue. They'll pick up the correct belts from the mall and place the proper belts in the correct locations.

The only time you should use blue-belts in the midgame is when you CANT use "more red belts". IE: your original early-game mainbus (which was likely built too small and weak to scale). "New designs" where you expand your base out a bit should generally use red-belts. Early game main-bus designs SHOULD be small and compact, so its a natural fit to upgrade those mainbuses to blue. But moving forward in the midgame??

I generally stick to red. EDIT: Mostly because red-underground size of 6 is way better than yellow-underground size of 4, while blue-underground size of 8 feels superflouous to me. (Useful for sure in endgame builds, but 6-size of red-undergrounds are a great balance between costs and practicality).

Prod mods in smelters is some new level of desperate.

Well, its either Prod3 Modules in all smelters, or grow your mines by 20%.

And yeah... making 20% more outposts is way harder. Welcome to endgame thinking.

I personally work with mods like factorissimo, which makes faster belts a must as soon as possible, as smaller buildings have limited input/output. coupled with bobs mods for mark 3 smelters and mini machines to be able to fit 104 smelters into a small factory building means i can saturate 4 blue belts easily and usually upgrade to purples before using modules.

I wondered about the blue speed too. Then when I was reading about furnaces I had an "aha" moment. I hope my memory is correct, but here it goes...

Using a good layout you can upgrade your smelters and on the same footprint by replacing your belts, furnaces (and probably inserters, I can't remember this detail) - all the in/out connections remain the same:
- Yellow belt fully saturated with ore input ==> Stone furnace array ==> Yellow belt fully saturated output. 1x speed.
- Red belt fully saturated with ore input ==> Steel or Electric furnace array ==> Red belt fully saturated with ore input. 2x yellow speed.
- Blue belt fully saturated with ore input ==> (Steel?) Electric furnace maxed with modules ==> Blue belt fully saturated with ore input. 3x yellow speed.

The gist of this should be correct, but maybe someone can correct any errors or vague/missing info.

I work with belts all the time. There is rarely need for 1:2 ratios on belts. As you upgrade your layouts the items/second doesn't change in neat 1:3 nor 1:2. For example when you put in productivity modules in you go from 1.25 to 0.7, or a ratio of 25:14. When you add beacons it totally depends on how many beacons you have per assembler.

In fact having the same increase in absolute numbers means if you have a blueprint for a yellow you can just add another one of those as you upgrade your belts. Eg you have one of them when your supply is yellow. You can add another when you get reds, and one more when you get blues. This is arguably simpler than doubling each time.

There's also the fact that blues have an even 40/second. I see blue as the standard, reds as 2/3 and yellows as 1/3. If anything should change it would make more sense to nerf yellows to 10 and reds to 20.

Hedning1390 wrote:There's also the fact that blues have an even 40/second. I see blue as the standard, reds as 2/3 and yellows as 1/3. If anything should change it would make more sense to nerf yellows to 10 and reds to 20.

Unfortunately, bots are better and belts need a buff. Under no circumstances should belts get nerfed under the current balance. If you want round numbers, 15/sec, 30/sec, and 45/sec seems to be a better series.

Chances are however, that 13.333/s is a round number in engine terms: the engine runs nominally at 60-ticks per second, and roughly 7-items fit on a belt tile.

Only blue belts matter when comparing bots to belts. I am not advocating a belt nerf though. I was just replying to the notion that each upgrade should be double speed and the suggestion I made then was to treat blue as the base rather than the yellow.

About the even numbers; it's 1/32 tiles per tick (yellow) and 32/9 items per lane, or 1/9 items per tick per lane. Is that even numbers? I have no idea why each item takes 9 spaces instead of some other number. Probably what they thought looked the best, but from a purely mechanical standpoint one could easily imagine each item taking 8 units instead of 9 which would make much easier calculations (as 8 is a power of 2, just as 32) and end up with your 15/30/45, or 12 for my 10/20/40.

Hedning1390 wrote:Only blue belts matter when comparing bots to belts. I am not advocating a belt nerf though. I was just replying to the notion that each upgrade should be double speed and the suggestion I made then was to treat blue as the base rather than the yellow.

About the even numbers; it's 1/32 tiles per tick (yellow) and 32/9 items per lane, or 1/9 items per tick per lane. Is that even numbers? I have no idea why each item takes 9 spaces instead of some other number. Probably what they thought looked the best, but from a purely mechanical standpoint one could easily imagine each item taking 8 units instead of 9 which would make much easier calculations (as 8 is a power of 2, just as 32) and end up with your 15/30/45, or 12 for my 10/20/40.

frankly, i think the base game can buff belts without a negative impact to gameplay.
I would buff blue belts up until a fast inserter or red inserter has issues picking up a single item on the outside of a corner.

belts + inserter pickup speed could use a buff vs bots.

I think you guys are way off, blue belts so you can run from one end to other quicker.

.... yeah .... my family doesn't let me out much ....

And now add Bobs Logistics and Deadlocks Stackers and you get belt speeds of 10, 20, 30, 40, 50, 60 and stacks of 5 items. Many more combiantion of odd ratios to get there.

If everything where simply 1:2 ratio then things would be boring.

Is there any specific reason why blue belt speed aren't just double the red belt speed?

Here is my thought process

Red belts are 2x faster than yellow belts => 2 yellow belts can be merged into a fully compressed red belt
Blue belts are 1.5x faster than red belts => 3 red belts can be merged into 2 fully compressed blue belts

Issues with blue belt speed:
1) This makes your base obsolete real fast:
Because the ratios are different, you can't do in place upgrades if you build planning to make maximum use of yellow / red belts at the start of the game. Since you have to rebuild a lot of the main bus, most people just straight up switch to bots at this point and i think blue belt not being double speed is a huge part of that.
For example, you have a red belt main bus and you take 1 full red belt off of it and build the section according to the 26.666 items per second.
Once you upgrade the main bus to blue belt, that section just doesn't work because you would only get 20 items per second with 1 splitter (old splitter)

2) It is awkward to work with multiples of 3 belts in factorio (or any odd number greater than 1 number of belts)
Splitters by nature work with only 2 belts at a time. Trying to handle 3 belts means we don't use half of one splitter. This makes things non symmetrical and i personally like builds that have symmetry in them. If you want to have symmetry, you are essentially locked into working with pairs of blue belts at a time, which sometimes is just too much or too expensive, especially during the transition time from red belt to blue belt.

---

Point 1 doesn't really matter as much now since we have priority splitters but point number 2 is still an issue, working with 3 belts at a time just doesn't look very good so my question again:

Is there any specific reason why blue belts are at 40 items per second instead of just straight up double the speed of red belts or more? Are speeds higher than 40 /s inefficient in terms of code / memory performance?

basically each tick, 60 per second,
yellow belt moves items 1 position per tick, which is 13.33333 per second, 800 per minute
red belt moves items 2 positions per tick, which is 26.66667 per second, 1600 per minute
blue belt moves items 3 positions per tick, which is 40 per second, 2400 per minute

you can get belts that are faster in mods so twice as fast as reds are possible its just the way the mechanics of the game work

Sir3n wrote:

darkfrei wrote:https://mods.factorio.com/mod/DoubleSpeedBelts

I noticed your mod. Only thing i'm wondering is whether it will actually affect performance

Nope, the same thing as vanilla. But you must separate material flow to the red belt for better working with inserters. So, blue belt is for transfer, not for inserters.

dragontamer5788 wrote:Why? What do blue belts offer you in the midgame? Instead of running 4 blue belts anywhere, run 6 red belts. 6x11.5 == 69 iron per tile of 6xRed belts. While 4x31.5 == 126 iron per tile of 4xblue. And when you see the price of 12x1.5 == 18 iron per tile of 12x Yellow Belts, you can start to see the serious cost-savings advantage of staying on earlier game tech for longer periods of time.

Its not like you spend any player-time placing belts in the midgame. Construction bots lay out the designs. So it doesn't matter if you're using yellow, red, or blue. They'll pick up the correct belts from the mall and place the proper belts in the correct locations.

4 blue belts transport 9600 iron plates per minute. You can save 168 tiles during that time. It takes much more time from bots to put them on ground. And in any case difference in costs are totally negligible in the phase when you consume 10000 iron plates per hour. I prefer shorter construction time instead of some mathematical savings of nothing.

In my opinion it is flaw in game balancing that infrastructure costs are negligible (except maybe modules). You can easily begin with 100 furnaces for iron and copper and expansion speed depends only how fast you can build. There is no challenge and no decisions what is best order to research and build. Under current conditions there is no technical reason to think costs except player time and possible CPU limits if you intend to build a huge base or have a weak computer.

I agree that speeds of belts does not fit in megabase scheme which seems to be the most important in developing (or is UPS saving only easy way to decline all laborious suggestions). Despite recent buffing bots are clearly overwhelming solution if you do not explicitly decide to make a belt base. Steeper increasing of speeds or additional tiers would be nice and easily programmable way to make belts better in technical sense. On the other hand, I do not like simple 1:2 or 2:3 ratios, so I suggest something like 1:2:5 or 1:3:8. In such case blue belts could be more expensive and need steel in addition to iron gears.

You must be playing different settings than me. For me after the initial start the limiting factor is clearing out aliens. Which means researching and building weapons and defenses. When playing with Bobs + Angles, RSO and Natural Evolution this is far from simple or fast and you have to balance resource usage. If you spend all your starter ores on furnaces and belts then you have nothing left to get advanced enough weapons to take on the aliens blocking more ores.

dood wrote:

dragontamer5788 wrote:Prod mods in smelters is some new level of desperate.

They are last on the wishlist, yes, but prod/speedmodded smelters make sense in a 10k sci megabase. Not so much in a "gonna launch my rocket and be done" game though. Needing to mine/train in 20% fewer minerals can make a big difference in a massive megabase. And requires you to rebuild/expand your mines less often.

What's actually truly desperate - is prodmodding your mines to squeeze that extra 30% of yield from them ('though once you get a bunch of mining production infinite research stacked, by comparison that bonus won't be all that useful anymore, and it becomes smarter to just use one speed, 2 efficiency - or no modules at all).

Everyone who wants each higher belt tier feature double the speed of the preceeding tier, should try Double Speed Belts. Works fine with the three vanilla belt tiers and promises to adjust mod-added belts too.

mrvn wrote:You must be playing different settings than me. For me after the initial start the limiting factor is clearing out aliens. Which means researching and building weapons and defenses. When playing with Bobs + Angles, RSO and Natural Evolution this is far from simple or fast and you have to balance resource usage. If you spend all your starter ores on furnaces and belts then you have nothing left to get advanced enough weapons to take on the aliens blocking more ores.

Surely if you take much biters you have different problems. I choose almost always low frequency and large size for all resources, because I want to find them instead of spam small patches to everywhere. I use always commands to clear ores when I begin to build actual base (when I get bots). And I know that fighting against biters at beginning is annoying but eventually I always win. So I take very large starting so that I do not have to bother biters before late midgame but they give varying resource sink at late game.