Curvy Track Beat Skinny Track [Edit: Results Under Review]
-
- Long Handed Inserter
- Posts: 55
- Joined: Tue Nov 25, 2014 11:41 pm
- Contact:
Curvy Track Beat Skinny Track [Edit: Results Under Review]
It's all about that base 'bout that base...
This just in, curvy track has better throughput and compression than skinny track!
Okay, don't get too excited, there's a catch - if you ever flatten out the curvy track into a skinny track everything breaks because items clump together and slow down to a crawl. But, there's some potential here I think. I at least wanted to share the test results with the community in case someone can think of some builds that might take advantage of the triple lane curvy track boosted throughput and compression.
I haven't done tests with like 1000 items or anything yet, just sharing the results from testing I do as I do it. In the video I am testing with 200 items on each track.
Test Video: https://www.youtube.com/watch?v=fOd5op_ee1k
This just in, curvy track has better throughput and compression than skinny track!
Okay, don't get too excited, there's a catch - if you ever flatten out the curvy track into a skinny track everything breaks because items clump together and slow down to a crawl. But, there's some potential here I think. I at least wanted to share the test results with the community in case someone can think of some builds that might take advantage of the triple lane curvy track boosted throughput and compression.
I haven't done tests with like 1000 items or anything yet, just sharing the results from testing I do as I do it. In the video I am testing with 200 items on each track.
Test Video: https://www.youtube.com/watch?v=fOd5op_ee1k
Last edited by spacesloth on Mon Dec 08, 2014 11:12 pm, edited 1 time in total.
Re: Curvy Track Beat Skinny Track in Throughput and Compress
You did not test nor for throughput, nor for compression, just for speed.
Attach your blueprints to forum posts with Foreman or Blueprint string.
Re: Curvy Track Beat Skinny Track in Throughput and Compress
Cool suggestion: Eatable MOUSE-pointers.
Have you used the Advanced Search today?
Need help, question? FAQ - Wiki - Forum help
I still like small signatures...
Have you used the Advanced Search today?
Need help, question? FAQ - Wiki - Forum help
I still like small signatures...
-
- Long Handed Inserter
- Posts: 55
- Joined: Tue Nov 25, 2014 11:41 pm
- Contact:
Re: Curvy Track Beat Skinny Track in Throughput and Compress
Imagine belt that is missing one piece in the middle.
Inserters fill it with items until it's full.
It will look like this:
>xxxxxxxxxxxx >>>>>>>>>>
On the left you have maximum compression, on the right you have empty track. Items are laying as close to each other as they can. They can't be any closer, and you can't put any more items on this track.
This is static compression though. It is useless metric, and when people say "maximum compression", they usually mean moving belt.
Now you switch off inserter and put the missing bit in place. Items start moving on the track.
It will look like this
>>>x>x>x>x>x>x>x>x>xxxx
The distance between items become larger when they started to move.
You can't see it with the naked eye on the yellow track, but the effect is there. On the blue track the difference is most obvious.
Anyway, this is what we call "maximum compression", or "100% compression", or "100% throughput"
The idea is that items can't be any closer to each other on the belt.
What is the difference between compression and throughput?
The compression is related to density of items on the track. The throughput is a characteristic of particular track configuration.
E.g. if classic belt rebalancer have 65% throughput, that means that the flow of items that traveled past this rebalancer now have 65% compression.
Now the important part is that you can't achieve 100% compression on the moving track just by using inserters, no matter how many you use. They can't fit items in the uneven gaps.
To do compression tests properly, you have to start from maximum static compression (preferred), or by joining two well compressed lines with splitter.
Counting items on screenshot is boring, thus compression is usually measured by measuring the throughput of particular track configuration and comparing it to the reference value (straight track).
Looking at the video, it is possible that diagonal track with 3 lines has more throughput than straight 2-line track, but it is impossible to tell without proper tests.
Update: ssilk's explanation above is much better
Inserters fill it with items until it's full.
It will look like this:
>xxxxxxxxxxxx >>>>>>>>>>
On the left you have maximum compression, on the right you have empty track. Items are laying as close to each other as they can. They can't be any closer, and you can't put any more items on this track.
This is static compression though. It is useless metric, and when people say "maximum compression", they usually mean moving belt.
Now you switch off inserter and put the missing bit in place. Items start moving on the track.
It will look like this
>>>x>x>x>x>x>x>x>x>xxxx
The distance between items become larger when they started to move.
You can't see it with the naked eye on the yellow track, but the effect is there. On the blue track the difference is most obvious.
Anyway, this is what we call "maximum compression", or "100% compression", or "100% throughput"
The idea is that items can't be any closer to each other on the belt.
What is the difference between compression and throughput?
The compression is related to density of items on the track. The throughput is a characteristic of particular track configuration.
E.g. if classic belt rebalancer have 65% throughput, that means that the flow of items that traveled past this rebalancer now have 65% compression.
Now the important part is that you can't achieve 100% compression on the moving track just by using inserters, no matter how many you use. They can't fit items in the uneven gaps.
To do compression tests properly, you have to start from maximum static compression (preferred), or by joining two well compressed lines with splitter.
Counting items on screenshot is boring, thus compression is usually measured by measuring the throughput of particular track configuration and comparing it to the reference value (straight track).
Looking at the video, it is possible that diagonal track with 3 lines has more throughput than straight 2-line track, but it is impossible to tell without proper tests.
Update: ssilk's explanation above is much better
Attach your blueprints to forum posts with Foreman or Blueprint string.
-
- Long Handed Inserter
- Posts: 55
- Joined: Tue Nov 25, 2014 11:41 pm
- Contact:
Re: Curvy Track Beat Skinny Track in Throughput and Compress
I will keep doing tests and hopefully have some more videos to follow up on this but below are some quick calculations I've done.
The compression results were measured beginning with a 100% compressed track as suggested but what bothers me I guess is there will be two different measures of throughput. One starting with a fully compressed belt and one if placing items on a belt directly with inserters. This is because with the diagonal belt, placing items on the belt directly with inserters gets double the amount of items on the middle track (fast lane) which would lead to higher speed and throughput I believe.
So, more to come but my knee jerk calculations are below. I will continue to record more information as I get time to test more.
Thanks for taking the time to help me learn how to properly test this stuff.
Compression Test Results
-----------------------------
Vertical Track: 64 items per 10 tracks = 6.4 items per track
Curvy Track: 36 items per 10 tracks = 3.6 items per track
Speed Test Results
----------------------
Curvy Track 32/32/36 tracks in 10 seconds (outside and inside lanes)
Average of 3.33 tiles per sec
Vertical Track: 1.875 tracks per sec according to wiki
Throughput Calculations
-----------------------------
Curvy Track 3.6 items per track * 3.33 tracks per sec = 11.988 items per sec
Vertical Track 6.4 items per track * 1.875 tracks per sec = 12 items per sec
The compression results were measured beginning with a 100% compressed track as suggested but what bothers me I guess is there will be two different measures of throughput. One starting with a fully compressed belt and one if placing items on a belt directly with inserters. This is because with the diagonal belt, placing items on the belt directly with inserters gets double the amount of items on the middle track (fast lane) which would lead to higher speed and throughput I believe.
So, more to come but my knee jerk calculations are below. I will continue to record more information as I get time to test more.
Thanks for taking the time to help me learn how to properly test this stuff.
Compression Test Results
-----------------------------
Vertical Track: 64 items per 10 tracks = 6.4 items per track
Curvy Track: 36 items per 10 tracks = 3.6 items per track
Speed Test Results
----------------------
Curvy Track 32/32/36 tracks in 10 seconds (outside and inside lanes)
Average of 3.33 tiles per sec
Vertical Track: 1.875 tracks per sec according to wiki
Throughput Calculations
-----------------------------
Curvy Track 3.6 items per track * 3.33 tracks per sec = 11.988 items per sec
Vertical Track 6.4 items per track * 1.875 tracks per sec = 12 items per sec
Re: Curvy Track Beat Skinny Track [Edit: Results Under Revi
this is cool as.
Has anyone actually got a useful application for it yet though?
Has anyone actually got a useful application for it yet though?
Re: Curvy Track Beat Skinny Track [Edit: Results Under Revi
I won't put too much afford into it, cause the behavior might change completely with v0.12. See https://forums.factorio.com/forum/vie ... 801#p53250
Cool suggestion: Eatable MOUSE-pointers.
Have you used the Advanced Search today?
Need help, question? FAQ - Wiki - Forum help
I still like small signatures...
Have you used the Advanced Search today?
Need help, question? FAQ - Wiki - Forum help
I still like small signatures...
Re: Curvy Track Beat Skinny Track [Edit: Results Under Revi
Yeah, not a fan. I like the wrinkle that the corner slowdowns introduce to teh game. Its the little foibles that make a game.ssilk wrote:I won't put too much afford into it, cause the behavior might change completely with v0.12. See https://forums.factorio.com/forum/vie ... 801#p53250
Eg. random blocks in the sky in minecraft. the occasional island spawn in factorio.
having slowdowns on a belt is a good thing to have, it makes you design in a certain way and introduces a good limitation on the player.
Re: Curvy Track Beat Skinny Track [Edit: Results Under Revi
Yeah, they give the belts in some range a chaotic behaviour.
Cool suggestion: Eatable MOUSE-pointers.
Have you used the Advanced Search today?
Need help, question? FAQ - Wiki - Forum help
I still like small signatures...
Have you used the Advanced Search today?
Need help, question? FAQ - Wiki - Forum help
I still like small signatures...
Re: Curvy Track Beat Skinny Track [Edit: Results Under Revi
More relevant post: https://forums.factorio.com/forum/vie ... 272#p42937ssilk wrote:I won't put too much afford into it, cause the behavior might change completely with v0.12. See https://forums.factorio.com/forum/vie ... 801#p53250