Oh my god how did I never realize that the Train Enthusiasts would absolutely crush this game? Thank you for posting! I’m autistic (frog-flavored, not train-flavored) and this will really do a lot of optimization heavy lifting that I want *done* but don’t want to *do* if that makes sense
I don't really use trains in Satisfactory (hell I don't even know if I ever unlocked them) but this is super cool and an invaluable resource for when I start using them. Thank you for taking time to make this video.
in 1.0 trains are going to be unlocked earlier and should be cheaper to build so i would highly recommend using a train system as i myself find it a lot of fun working with trains.
As a guy who spent 230+ on the first save, trains make things real fun. They’re less predictable in terms of throughput than belts, but nothing beats the view of multiple trains on one huge roundabout!
I was already using cross intersection with path... I honestly didn't think it would be in the top performers, you need to get WAY more complicated and take up way more space to get better results, I'm both happy and surprised.
I think it makes sense because it really is the intentional use of path signal from a dev perspective - so it SHOULD perform well as designed. I think I'll continue to use it knowing it's likely the best performant for the work put in. I'd love for every intersection to be a top performing stack interchange.. but I only have so many hours in my life I can put into a game, lol
2 Observations: A: With a single path zone, Double Roundabout is 6 transits better than Roundabout w/ slip lanes. But with split path zones, Roundabout w/ slip lanes is 3 transits better than Double Roundabout. I wonder if that's just due to statistical randomness, or if there's some reason for that. B: Reverse Roundabout w/ slip lanes is clearly better than the both of them though. Against the Roundabout w/ slip lanes, it has a routing advantage (because left turns don't block each other). But against the double roundabout, it just seems to have the advantage of offering twice as many parallel "waiting spots" in the approach lanes (even though nobody actually needs to wait in the slip lanes). I wonder if the double roundabout could be improved if it had separate approach lanes for entering the clockwise and counter-clockwise circles... [edit:] And while we're adding lanes, a cross junction with added slip lanes would be interesting too.
A) There is no randomness in the train operation. All tests start from the same save state with the power off in the test track. The first minute of the test the trains are still spreading out, they all start lined up at their first station. B) adding an additional waiting spot will likely improve performance but it would need to be tested to confirm. The save file I use for the tests is my channel links, if you want to have a go at it. It is setup for stacked interchange but the trains are in the starting positions for a test. You can build any interchange you want there.
After having dabbled in the train world and figuring out decent intersections I wish I had seen this video before, truly a fabulous bit of testing going on here.
For asthetics I love the turbine intersection, but I've never built one. I'm glad to see it is part of the top 3 performer. Thanks for that incredible amount of labour!
i can happily admit i have never even heard of all of the interchanges after roundabout. but, now i want to start building them, just waiting for the update. btw, those builds are fire, thanks for all the work :)
@@BitwiseAssembly thor from PirateSoftware also "only" uses MSPaint for hist graphics, its a great tool for quickly making simple graphics to illustrate something complex
Thank you. The turbine interchange was popular on the Reddit before I even started playing, but it was created in an update before signals and train collisions. So I can only take credit for the signaling choices I made.
You're my hero, I love your dedication to efficient transportation. I'll being taking this with me into 1.0. One thing though. I do think I will make my interchanges into a mound shape so that each train has a tiny boost if they have to stop and start at a merge. Edit: Man, I knew that the gains would be negligible but I never thought the annoyance of implementing it would be so high.
I remade as much of the Stack Exchange and Reverse Roundabout w/ Slip Lanes as I could in a modded blueprint designer and put it up on satisfactory-calculator but it seems I can't link it. I gave up on making the stack into a mound, but maybe I'll revisit later down the road.
@@BitwiseAssembly I've shoved the blueprint in there and it's just another thing to see it running, thanks for providing that! It's a bit thicker since I spaced the roads a bit and a bit taller because I used thicker bases so I was kinda worried if it was gonna fit but I guess I'm bad judge for that. I think I'd like to try making this next [watch?v=EaYaBRcryb8] I'm not sure if I'll be able to fit it into a blueprint though so it might just end up being a centerpiece like you were talking about.
I find it interesting the roundabouts hit failure states causing 4 way blocks with trains blocking each other. I think any intersection that has the potential for a grid lock to not be worth using. The fact that the cross intersection performed higher than I expected is interesting. That is the only 4 way intersection I have on my train line right now, so unless I quadruple the about of trains running through it. I can probably keep using that interchange with no issues.
Looks like those failed roundabouts were divided up into block zones, whereas in the main test they had path (chain) signals at the entrance to prevent entering and stopping partway.
There is one downside to the turbine interchange, that I would need to fix before I consider using it. There is not enough clearance for the pioneer to ride on top of the train.
@@BitwiseAssembly After people were talking about the vertical clearance needed for the Stack and Turbine intersections, I couldn't help but think "For a flat intersection, the inverse roundabout is pretty impressive!"
It is, but it has a slightly lower priority for straight through traffic. So I would only consider using it on a destination that a lot of trans go to. For all the testing I did here, my main world still only has cross intersections.
@@BitwiseAssembly Ah yes, that's a good point. Maybe for an intersection where you know a lot of the trains will not be going straight through. I'm happy to hear how you use train intersections, because at the end of the day, after all of the min maxing, we still have to make decisions about practicality :D
The Turbine exchange is what I was imagining building when I was running into issues of deadlocking due to having intersections close to my stations that did U turns. Ever since then, I have been removing left turns or crossing rails if I don't need them and relying on long distance U turns. Awful for throughput yes, but the alternative was eventual deadlocks. With my next base I am looking to do the same concept but not rely on loops at the end of rails. The stack interchange looks more doable to create compared to the turbine exchange (Building that seems awful) so I think I will go with that at a 2m slope. One thing I kinda wonder though is how this would all work for T intersections, but Ill check your videos to see if there is one on that.
Appreciate the serious engineering that went into this analysis! In addition to the mesmerizing video footage, i need a 20h loop of every junction! But OP: Where is the scientific paper on the topic? I have seen bachelor thesis with less content and work put into than this video...
I don’t think my potato computer would survive that many hours of video. That initial clip took eight hours to edit, it kept crashing my computer. This is about the average effort I put into analyzing anything I have questions about. One of the reasons why my next big video is taking so long.
Very interesting how the interchanges fair with trains, i appreciate the work you took to make all this. That a block turbine is fatser than path and sub path turbine is interesting. Ive also been playing cities skylines and from what i can gather from that game maybe you can try more interchanges from real life like the speed roundabout or other things.
After seeing the results, I think I'd prefer cross intersections due to verticality and train physics. Idk how much going up/down slopes affects trains, but I know it does SOMETHING, and the cross is surprisingly good. Id consider reverse roundabout, but I'm too lazy to figure out how to make a blueprint of that.
As someone who comes from Factorio and is used to chain signals and smarter trains. i don't really get oath signals. every time i just use block signals instead so that trains only slow down when they must
That was so cool to watch! Taking notes for my 1.0 train network (soon lol). Do you roughly remember how long it took you to build each of them? Would be interesting to see which one has the best build time to efficiency ratio.
The roundabout ones were all built on top of each other as one meta junction. For the versions of it you see I dismantled the tracks not needed. So I can’t tell you the build time. The stack and turbine were about the same build time, which I don’t remember since time flys when you’re having fun.
Question -- I noticed in some of the wide shots that you have the station "arrival" lines all cris-crossing to get into the pink stations, and that on several occations a train had to wait before the cris-cross as an earlier train slowed down to enter the station. Wouldn't this have a ripple affect across the various blocks, which would affect the end results? At the very least it's an extra variable that needs to be accounted for.
There is two reasons a train would stop in station waiting zone. 1) the station the train wants is occupied. 2) another train is traversing the station path zone. Both affect the max possible score an intersection can get and it affects the higher scoring intersections more. When I am scoring the tests I am using the uncropped view. This allows me to see the trains in the waiting zones. I only saw a few instances of a crossover delay that would have reduced the score. The lost point(s) from the station limitations would not change the ranking if I added them. The intersection most affected was the stack interchange. Hopefully that answered your question.
hello ! first of all congratulations for the quality of the video moreover I am impressed by the quality of the stats (similar to the scientific field...) I am a player that never using trains and I wondered if it allowed to be as fluid on the flow of resources used as with conveyors belt ? Again brilliant work of data and stats !
Trains offer more flexibility in the movement of resources than a belt line. If you’re asking about the transfer rate there are a lot of variables you would need to consider, if you are comparing a train line to a belt line.
@@BitwiseAssembly flexibility and being able to group together an entire production run and dispatch it to different plants more easily than with belt conveyors? To have the same fluidity as belt conveyors, do I need to take into account the train's travel time and the quantity it carries? And possibly add several trains on the same track/or more wagon per train? thank you for taking the time to answer my beginner's questions :D
You can find plenty of answers to belts vs trucks vs trains elsewhere, but the simple answer is belts are perfect for short distance. Trucks and trains can both be good for high throughput medium to long distance. Belts are your only option for short distances, but can be as long as you want. You could have a belt bridge network that spans the map. Belts get harder to manage the longer they the network. You have to upgrade each section with new tiers and if you have to add a new belt line if you require more capacity. In contrast, roads and rails can be reused by new routes. If you need a higher capacity on the same route, you can add another truck or train on the same path. If you need a new route, you only need to build a road or track from existing infrastructure to the new location. Trucks vs trains is interesting. Trucks need fuel which can be more difficult to get started. Trucks can go pretty far round trip with just coal making a good strategy to have a single truck bring coal to a central place to fuel other truck stations. There is no such thing as a road in Satisfactory. You build foundations that trucks drive on. This means there's no pathing information. To automate a route, you must record yourself driving the route in a loop. This makes trucks simple for small routes. Once a route is recorded, you can only delete "nodes". You cannot move them. It is easy in the beginning to mess up a route you record. The easiest way I've found is use the vehicle path mode (shows connected lines instead of just arrows), then if you make a mistake, you can back up and try to resume the route and delete nodes later. You can add more trucks to the same route without recording a new route. You must drive and record new routes. Trains are a new level of complexity. Tracks have a set of rules and nuances that make them difficult to get started. I struggled at the beginning until I watched a few UA-cam videos on how they functioned and how to build sections. It took a bit of practice. Now building tracks isn't difficult. There's a pattern to it. Tracks and stations take up a lot of space. In terms of the whole map, not much, but if you're used to belting everything, a simple U-turn takes at least 7 foundations wide and train stations are huge. It takes a bit to get used to the scale. Trains are electric and work off your grid rather than needing fuel. In some ways this is easier. Train networks can be expanded. A new route only needs to connect new areas to existing tracks. Trains routes can be managed by scheduling between train stations. To create a new route, you create a schedule from one train station to another. The game figures out how to route the train on your track network. You don't need to record yourself driving the train first. This makes it easier to make new routes, especially longer ones. You can play the game however you want. Personally, I like seeing trains and truck driving everywhere. Roads and tracks add to the aesthetic of the game. If I needed to create a new long distance route, a belt bridge would be more daunting than adding more roads or tracks. Roads are simpler with blueprints. Beautiful curves are easier with the blueprint plus mod that gives a larger blueprint area.
After careful examination the problem is caused by placement of path and block signals. Generally replacing all block signals should always increase performance, but the block turbine has additional block signals in the design which makes it even more subdivided.
Big takeaway for me is that trains will use slip ways - I've never built them because I sort of thought the pathing wouldn't use them properly, but it looks like it used them fine in this test.
@@BitwiseAssembly Steeper ramps would mean that heavier trains lose more speed going up them, so this might be ideal for the space tradeoff. Not sure how much that effects trains when they're already at full tilt, though if it's very close to a start/stop point, that could be a factor.
Could you add some blueprints to your share Google Drive directory ? I do not know which are too big to blueprint, but visually, I think the Cross Intersection Path would fit a blueprint designer
Even in the 5x5 I don’t think the cross intersection would fit. You need 1.5 foundation length track to connect the track to when building the intersection. And you can’t having anything crossing the blueprint designer border.
I think for interchanges with lanes for every direction, like turbine or stack interchange, block signals are better, because the trains don't need to "request" their right of way (like they do for path signals) and instead just fly through the interchange at max speed. If you use (subdivided) paths, they need to enter the block in front of the signal first, to request green and if that block is to short, they slow down for the still red path signal, through which they slow down the whole interchange. I guess if there was a subdivided path stack interchange, it would be slower than the block signal version too. For normal intersections or roundabouts tho it's way easier to just use (subdivided) paths.
Combination of two factors. 1) Trains will not enter a path zone unless there is a block zone it can reserve on the other side. 2) trains can wait in a block zone in the interchange which could allow the next train to go around the waiting train. I did not test the stack interchange with path or subdivided path because it did not meet the criteria to require a path signals.
It did not meet the either criteria to require path signals. None of the zones had 2 entrances and 2 exits where two trains could traverse without hitting or a train stopping in any of the zones could cause gridlock.
@@BitwiseAssembly so for that type of intersection it is best to go with only block signals? I do also have a stack interchange (slightly different) but i used path signals
@@BitwiseAssembly with "for that type" i meant the stack because i already built one as i said. It currently opperates with a normal path signal not subdevided or anything
Ok, sorry for the confusion. so path vs block,I believe that any interchange that can support a Block only signaling solution will perform better than a solution including path signal. But the difference may be so minor you will not notice the difference unless you are stress testing the interchange.
The levels for the stack interchange are all 10m apart. If the entry hight is 0 the levels are at +15, +5, -5, -15. As to if I will make a tutorial, I am not sure if I will. I have been avoiding talking in videos, so I need to figure out a method for the video to be easily understood without talking.
You shouldn't change test conditions during the test. The systems need to be 100% identical EXCEPT in the junction itself or there's no real test between the systems, only between the two that exist in the identical network..
? I did not change the test conditions. All of these are from the same save state and only the junctions and signaling changes. The start is powering the train network. Are you seeing something different?
You should watch my video train signals testing for more insight. But the answer is depends the longer the entry block length the higher speed the train will maintain. But the relationship is not linear leading to some very long entry blocks if speed is the only consideration. Train will start requesting a reservation when two things are true. The train is in the entry block and braking distance calculation to the path signal says to start braking.
Do you think we could get a world download for this save file? I'd like to play around with some different intersections, but I'm far too lazy to build one of these test setups myself 🤣
I will think about it. I have been avoiding talking because my setup not ideal for audio, it’s barely functional for playing and capturing the videos. But I will think about it.
It’s mainly a noisy environment issue. My setup is weird. My good computer is so loud i only use it remotely from a 7 year old laptop in the family room. So when I am making these videos/playing I am sitting on the couch with the laptop stream casting into the good computer. I am normally covered in my kids as they are watching TV. I have a lot less time to work with if I exclude family time.
IMO your roundabouts are too small. Need to be at least double the diameter, they need to be able to hold a full vehicle in each quadrant otherwise is pointless.
I appreciate time spend on collecting all data tested, you are actual pioneer of this game
Thank you, I am hoping this helps fellow pioneers.
@@BitwiseAssembly man you kidding ?? You didnt help.. you just REDEFINE what Train efficiency in Satisfactory IS ! this should be the in-game tutorial
Oh my god how did I never realize that the Train Enthusiasts would absolutely crush this game? Thank you for posting! I’m autistic (frog-flavored, not train-flavored) and this will really do a lot of optimization heavy lifting that I want *done* but don’t want to *do* if that makes sense
I don't really use trains in Satisfactory (hell I don't even know if I ever unlocked them) but this is super cool and an invaluable resource for when I start using them. Thank you for taking time to make this video.
If it helps a fellow pioneer, it was worth all the hours it took make.
in 1.0 trains are going to be unlocked earlier and should be cheaper to build so i would highly recommend using a train system as i myself find it a lot of fun working with trains.
yess, and there are a couple of really good series for track building if you do get in to trains.
As a guy who spent 230+ on the first save, trains make things real fun. They’re less predictable in terms of throughput than belts, but nothing beats the view of multiple trains on one huge roundabout!
@@BitwiseAssemblyi love the altruism in the Satisfactory community, it really shows that there is good left in people
I was already using cross intersection with path... I honestly didn't think it would be in the top performers, you need to get WAY more complicated and take up way more space to get better results, I'm both happy and surprised.
I think it makes sense because it really is the intentional use of path signal from a dev perspective - so it SHOULD perform well as designed. I think I'll continue to use it knowing it's likely the best performant for the work put in. I'd love for every intersection to be a top performing stack interchange.. but I only have so many hours in my life I can put into a game, lol
The round about with sliplanes is simply mesmerizing. I could watch for hours
This is amazing. The amount of labor and data here. Truly wonderful.
Saving this for 1.0 thank you! Amazing video
I'm absolutely gonna use the turbine interchange a lot. It's SO beautiful - and also thank you, very helpful video!
2 Observations:
A: With a single path zone, Double Roundabout is 6 transits better than Roundabout w/ slip lanes. But with split path zones, Roundabout w/ slip lanes is 3 transits better than Double Roundabout. I wonder if that's just due to statistical randomness, or if there's some reason for that.
B: Reverse Roundabout w/ slip lanes is clearly better than the both of them though. Against the Roundabout w/ slip lanes, it has a routing advantage (because left turns don't block each other). But against the double roundabout, it just seems to have the advantage of offering twice as many parallel "waiting spots" in the approach lanes (even though nobody actually needs to wait in the slip lanes).
I wonder if the double roundabout could be improved if it had separate approach lanes for entering the clockwise and counter-clockwise circles...
[edit:] And while we're adding lanes, a cross junction with added slip lanes would be interesting too.
A) There is no randomness in the train operation. All tests start from the same save state with the power off in the test track. The first minute of the test the trains are still spreading out, they all start lined up at their first station.
B) adding an additional waiting spot will likely improve performance but it would need to be tested to confirm.
The save file I use for the tests is my channel links, if you want to have a go at it. It is setup for stacked interchange but the trains are in the starting positions for a test. You can build any interchange you want there.
After having dabbled in the train world and figuring out decent intersections I wish I had seen this video before, truly a fabulous bit of testing going on here.
For asthetics I love the turbine intersection, but I've never built one. I'm glad to see it is part of the top 3 performer.
Thanks for that incredible amount of labour!
i can happily admit i have never even heard of all of the interchanges after roundabout.
but, now i want to start building them, just waiting for the update.
btw, those builds are fire, thanks for all the work :)
It is a travesty you only have 219 subs. Keep the content coming. This is really good stuff.
I'm stealing pretty much everything.
That’s expected, I don’t talk and my graphics are MS paint. This is just a hobby, I do not expect that to change.
@@BitwiseAssembly thor from PirateSoftware also "only" uses MSPaint for hist graphics, its a great tool for quickly making simple graphics to illustrate something complex
the stack interchange takes so much space vertically but man is it ever smooth!
Prior to this video, I changed my intersections from turbines to simple roundabouts, but now I will change them to stack interchanges
This video is amazing. Thank you for all the work
I’m mind boggled with how much like intersections/interchanges there are here and the data is really 👍🏽.
this so simply stellar, i just wish trains in satisfactory weren't such a massive pain to work with.
I am really grateful with you for sharing this titanic data collecting work, you truly deserve the golden statue pioneer
Incredible presentation here. Nice work. And I really like the Turbine Interchange.
Thank you. The turbine interchange was popular on the Reddit before I even started playing, but it was created in an update before signals and train collisions.
So I can only take credit for the signaling choices I made.
Thank you for taking the time. It is an interesting traffic study!
You're my hero, I love your dedication to efficient transportation. I'll being taking this with me into 1.0.
One thing though. I do think I will make my interchanges into a mound shape so that each train has a tiny boost if they have to stop and start at a merge.
Edit: Man, I knew that the gains would be negligible but I never thought the annoyance of implementing it would be so high.
I only used the cross intersection in my game. I might consider using a more complicated one as a centerpiece.
I remade as much of the Stack Exchange and Reverse Roundabout w/ Slip Lanes as I could in a modded blueprint designer and put it up on satisfactory-calculator but it seems I can't link it. I gave up on making the stack into a mound, but maybe I'll revisit later down the road.
my testing world download is available in my channel links. It is currently setup with the stack interchange.
@@BitwiseAssembly I've shoved the blueprint in there and it's just another thing to see it running, thanks for providing that! It's a bit thicker since I spaced the roads a bit and a bit taller because I used thicker bases so I was kinda worried if it was gonna fit but I guess I'm bad judge for that.
I think I'd like to try making this next [watch?v=EaYaBRcryb8] I'm not sure if I'll be able to fit it into a blueprint though so it might just end up being a centerpiece like you were talking about.
I find it interesting the roundabouts hit failure states causing 4 way blocks with trains blocking each other. I think any intersection that has the potential for a grid lock to not be worth using. The fact that the cross intersection performed higher than I expected is interesting. That is the only 4 way intersection I have on my train line right now, so unless I quadruple the about of trains running through it. I can probably keep using that interchange with no issues.
Looks like those failed roundabouts were divided up into block zones, whereas in the main test they had path (chain) signals at the entrance to prevent entering and stopping partway.
I know it's not the most efficient but there's something majestic about that slip lane roundabout, definitely stealing that
getting flashbacks to the train nerds arguing in factorio
gonna stick with my smoothbrained roundabouts
Even though the turbine interchange is the second best, I feel it’s worth it for the most visually interesting one
There is one downside to the turbine interchange, that I would need to fix before I consider using it. There is not enough clearance for the pioneer to ride on top of the train.
@@BitwiseAssembly After people were talking about the vertical clearance needed for the Stack and Turbine intersections, I couldn't help but think "For a flat intersection, the inverse roundabout is pretty impressive!"
It is, but it has a slightly lower priority for straight through traffic. So I would only consider using it on a destination that a lot of trans go to.
For all the testing I did here, my main world still only has cross intersections.
@@BitwiseAssembly Ah yes, that's a good point. Maybe for an intersection where you know a lot of the trains will not be going straight through.
I'm happy to hear how you use train intersections, because at the end of the day, after all of the min maxing, we still have to make decisions about practicality :D
Wow, awesome work. I'am looking for a long time, for content like yours. There are so many ideas to try for me. It helps a lot :)
The Turbine exchange is what I was imagining building when I was running into issues of deadlocking due to having intersections close to my stations that did U turns. Ever since then, I have been removing left turns or crossing rails if I don't need them and relying on long distance U turns. Awful for throughput yes, but the alternative was eventual deadlocks. With my next base I am looking to do the same concept but not rely on loops at the end of rails. The stack interchange looks more doable to create compared to the turbine exchange (Building that seems awful) so I think I will go with that at a 2m slope.
One thing I kinda wonder though is how this would all work for T intersections, but Ill check your videos to see if there is one on that.
Appreciate the serious engineering that went into this analysis!
In addition to the mesmerizing video footage, i need a 20h loop of every junction!
But OP: Where is the scientific paper on the topic? I have seen bachelor thesis with less content and work put into than this video...
I don’t think my potato computer would survive that many hours of video.
That initial clip took eight hours to edit, it kept crashing my computer.
This is about the average effort I put into analyzing anything I have questions about. One of the reasons why my next big video is taking so long.
Very interesting how the interchanges fair with trains, i appreciate the work you took to make all this. That a block turbine is fatser than path and sub path turbine is interesting. Ive also been playing cities skylines and from what i can gather from that game maybe you can try more interchanges from real life like the speed roundabout or other things.
After seeing the results, I think I'd prefer cross intersections due to verticality and train physics. Idk how much going up/down slopes affects trains, but I know it does SOMETHING, and the cross is surprisingly good. Id consider reverse roundabout, but I'm too lazy to figure out how to make a blueprint of that.
wow, as a no-trains satisfactory player, I now need another video explaining this video...
I would suggest the previous two videos to this one should help you.
They go into block and path signal behaviors.
Crazy work here ❤❤❤
✨✨✨QUALITY CONTENT✨✨✨
awesome!!!
Complete et usefull video, thanks 👍
As someone who comes from Factorio and is used to chain signals and smarter trains. i don't really get oath signals. every time i just use block signals instead so that trains only slow down when they must
That was so cool to watch! Taking notes for my 1.0 train network (soon lol).
Do you roughly remember how long it took you to build each of them? Would be interesting to see which one has the best build time to efficiency ratio.
The roundabout ones were all built on top of each other as one meta junction.
For the versions of it you see I dismantled the tracks not needed. So I can’t tell you the build time. The stack and turbine were about the same build time, which I don’t remember since time flys when you’re having fun.
But I would just use the normal cross intersection if you care about build time.
It scored well for the effort.
@@BitwiseAssembly Thank you!
Question -- I noticed in some of the wide shots that you have the station "arrival" lines all cris-crossing to get into the pink stations, and that on several occations a train had to wait before the cris-cross as an earlier train slowed down to enter the station. Wouldn't this have a ripple affect across the various blocks, which would affect the end results? At the very least it's an extra variable that needs to be accounted for.
There is two reasons a train would stop in station waiting zone.
1) the station the train wants is occupied.
2) another train is traversing the station path zone.
Both affect the max possible score an intersection can get and it affects the higher scoring intersections more.
When I am scoring the tests I am using the uncropped view. This allows me to see the trains in the waiting zones.
I only saw a few instances of a crossover delay that would have reduced the score.
The lost point(s) from the station limitations would not change the ranking if I added them. The intersection most affected was the stack interchange.
Hopefully that answered your question.
this needs top be available in the wiki or game, holy fuck
It is. satisfactory.wiki.gg/wiki/Railway#External_links
hello ! first of all congratulations for the quality of the video moreover I am impressed by the quality of the stats (similar to the scientific field...) I am a player that never using trains and I wondered if it allowed to be as fluid on the flow of resources used as with conveyors belt ?
Again brilliant work of data and stats !
Trains offer more flexibility in the movement of resources than a belt line.
If you’re asking about the transfer rate there are a lot of variables you would need to consider, if you are comparing a train line to a belt line.
@@BitwiseAssembly flexibility and being able to group together an entire production run and dispatch it to different plants more easily than with belt conveyors?
To have the same fluidity as belt conveyors, do I need to take into account the train's travel time and the quantity it carries? And possibly add several trains on the same track/or more wagon per train?
thank you for taking the time to answer my beginner's questions :D
You can find plenty of answers to belts vs trucks vs trains elsewhere, but the simple answer is belts are perfect for short distance. Trucks and trains can both be good for high throughput medium to long distance.
Belts are your only option for short distances, but can be as long as you want. You could have a belt bridge network that spans the map. Belts get harder to manage the longer they the network. You have to upgrade each section with new tiers and if you have to add a new belt line if you require more capacity.
In contrast, roads and rails can be reused by new routes. If you need a higher capacity on the same route, you can add another truck or train on the same path. If you need a new route, you only need to build a road or track from existing infrastructure to the new location.
Trucks vs trains is interesting. Trucks need fuel which can be more difficult to get started. Trucks can go pretty far round trip with just coal making a good strategy to have a single truck bring coal to a central place to fuel other truck stations.
There is no such thing as a road in Satisfactory. You build foundations that trucks drive on. This means there's no pathing information. To automate a route, you must record yourself driving the route in a loop. This makes trucks simple for small routes. Once a route is recorded, you can only delete "nodes". You cannot move them. It is easy in the beginning to mess up a route you record. The easiest way I've found is use the vehicle path mode (shows connected lines instead of just arrows), then if you make a mistake, you can back up and try to resume the route and delete nodes later. You can add more trucks to the same route without recording a new route. You must drive and record new routes.
Trains are a new level of complexity. Tracks have a set of rules and nuances that make them difficult to get started. I struggled at the beginning until I watched a few UA-cam videos on how they functioned and how to build sections. It took a bit of practice. Now building tracks isn't difficult. There's a pattern to it.
Tracks and stations take up a lot of space. In terms of the whole map, not much, but if you're used to belting everything, a simple U-turn takes at least 7 foundations wide and train stations are huge. It takes a bit to get used to the scale. Trains are electric and work off your grid rather than needing fuel. In some ways this is easier.
Train networks can be expanded. A new route only needs to connect new areas to existing tracks. Trains routes can be managed by scheduling between train stations. To create a new route, you create a schedule from one train station to another. The game figures out how to route the train on your track network. You don't need to record yourself driving the train first. This makes it easier to make new routes, especially longer ones.
You can play the game however you want. Personally, I like seeing trains and truck driving everywhere. Roads and tracks add to the aesthetic of the game. If I needed to create a new long distance route, a belt bridge would be more daunting than adding more roads or tracks. Roads are simpler with blueprints. Beautiful curves are easier with the blueprint plus mod that gives a larger blueprint area.
How the hell was the turbine block better than the turbine subdivided path?
After careful examination the problem is caused by placement of path and block signals. Generally replacing all block signals should always increase performance, but the block turbine has additional block signals in the design which makes it even more subdivided.
I just use Roundabout with subdivided path on 12 segments
Great, how show us how to build a stacked interchange!
Big takeaway for me is that trains will use slip ways - I've never built them because I sort of thought the pathing wouldn't use them properly, but it looks like it used them fine in this test.
The slipway is a slightly shorter track to the destination, and trains will always take the shortest path.
How small, elegant and efficient turbine interchange is.
Wouldn't it be possible to make an even more compact Stack Interchange ?
Sure just use the steeper ramps.
The use of only 1 m ramps is just a preference.
@@BitwiseAssembly Steeper ramps would mean that heavier trains lose more speed going up them, so this might be ideal for the space tradeoff. Not sure how much that effects trains when they're already at full tilt, though if it's very close to a start/stop point, that could be a factor.
Could you add some blueprints to your share Google Drive directory ?
I do not know which are too big to blueprint, but visually, I think the Cross Intersection Path would fit a blueprint designer
Even in the 5x5 I don’t think the cross intersection would fit. You need 1.5 foundation length track to connect the track to when building the intersection.
And you can’t having anything crossing the blueprint designer border.
dude you're crazy :D
So, subdivided paths are always better, which makes sense. Except for the turbine interchange, where the block version is better. Any idea why?
I think for interchanges with lanes for every direction, like turbine or stack interchange, block signals are better, because the trains don't need to "request" their right of way (like they do for path signals) and instead just fly through the interchange at max speed.
If you use (subdivided) paths, they need to enter the block in front of the signal first, to request green and if that block is to short, they slow down for the still red path signal, through which they slow down the whole interchange.
I guess if there was a subdivided path stack interchange, it would be slower than the block signal version too.
For normal intersections or roundabouts tho it's way easier to just use (subdivided) paths.
Combination of two factors.
1) Trains will not enter a path zone unless there is a block zone it can reserve on the other side.
2) trains can wait in a block zone in the interchange which could allow the next train to go around the waiting train.
I did not test the stack interchange with path or subdivided path because it did not meet the criteria to require a path signals.
Screw max efficiency I am making turbine interchange any day of the week. Looks too good.
Why was the stack not tested with path signals? I suppose it would be similar to turbine and blocks are better than path?
It did not meet the either criteria to require path signals.
None of the zones had 2 entrances and 2 exits where two trains could traverse without hitting or a train stopping in any of the zones could cause gridlock.
@@BitwiseAssembly so for that type of intersection it is best to go with only block signals? I do also have a stack interchange (slightly different) but i used path signals
Depends on which best you are aiming for. Highest throughput would be the stack interchange, easiest for train beginners the cross intersection.
@@BitwiseAssembly with "for that type" i meant the stack because i already built one as i said. It currently opperates with a normal path signal not subdevided or anything
Ok, sorry for the confusion.
so path vs block,I believe that any interchange that can support a Block only signaling solution will perform better than a solution including path signal.
But the difference may be so minor you will not notice the difference unless you are stress testing the interchange.
Do you think you can make a tutorial on how to build the stack interchange?
The levels for the stack interchange are all 10m apart.
If the entry hight is 0 the levels are at +15, +5, -5, -15.
As to if I will make a tutorial, I am not sure if I will. I have been avoiding talking in videos, so I need to figure out a method for the video to be easily understood without talking.
any failures means not worth building. you lose any speed advantage to the hours it will be sitting still until you find it and fix it
You shouldn't change test conditions during the test. The systems need to be 100% identical EXCEPT in the junction itself or there's no real test between the systems, only between the two that exist in the identical network..
? I did not change the test conditions.
All of these are from the same save state and only the junctions and signaling changes.
The start is powering the train network.
Are you seeing something different?
What's the optimal length for the block-segment, before a Path-Signal?
Like how early should a train reserve it's path through an intersection?
You should watch my video train signals testing for more insight.
But the answer is depends the longer the entry block length the higher speed the train will maintain. But the relationship is not linear leading to some very long entry blocks if speed is the only consideration.
Train will start requesting a reservation when two things are true.
The train is in the entry block and braking distance calculation to the path signal says to start braking.
@@BitwiseAssembly Oh I didn't know about the second condition with the breaking distance.
That makes so much more sense now.
Do you think we could get a world download for this save file?
I'd like to play around with some different intersections, but I'm far too lazy to build one of these test setups myself 🤣
There is one in my channel links.
When the cities skylines nerds play satisfactory 😂
30h of City skylines hardly makes me a nerd. I do have 3000h in Satisfactory though…. so maybe that’s why.
Cities skyline interchange ? Naaaah
Satisfactory interchange ? Hell yeah
Can you find these as Blueprints somewhere?
Sorry. I don’t have these as blueprints.
So much time spent collection and sorting data and way to less people can appreciate. Just missing a voiceover
Yup, it’s definitely missing the voiceover.
I am 43 videos on this channel and I have not talked once.
@@BitwiseAssembly You should change that IMO. Not everyone is as data affine as us ^^
I will think about it.
I have been avoiding talking because my setup not ideal for audio, it’s barely functional for playing and capturing the videos. But I will think about it.
@@BitwiseAssembly You could use AI to take over the talking bits, or get someone else maybe to do the talking for you.
It’s mainly a noisy environment issue.
My setup is weird. My good computer is so loud i only use it remotely from a 7 year old laptop in the family room. So when I am making these videos/playing I am sitting on the couch with the laptop stream casting into the good computer.
I am normally covered in my kids as they are watching TV.
I have a lot less time to work with if I exclude family time.
IMO your roundabouts are too small. Need to be at least double the diameter, they need to be able to hold a full vehicle in each quadrant otherwise is pointless.
amazing !