EFFICIENT MASS STORAGE - TIMBERBORN Update 6 R&D Session 2
HTML-код
- Опубликовано: 12 ноя 2024
- Join us in Research & Development as we investigate how storage works and how we can create efficient mass storage.
BEAVER MERCHANDISE LINK ►►► www.teepublic....
FULL SERIES PLAYLIST ► bit.ly/46N6wRL
This is part 53 of a challenging new gameplay series in the awesome beaver city building game TIMBERBORN Update 6 on Hard Mode. Can I survive the terrifying droughts and bad tides and engineer the ULTIMATE Beaver Paradise?
Picture Perfect's Artwork Link ► teepublic.com/u...
Support Skye on Patreon ► / skyestorme
Follow Skye on Twitch ► / skyestorme
Follow Skye on Twitter ► / xskyestormex
Skye's Official Merch Store ► bit.ly/3pGE7X7
Skye's Official Game Store ► www.nexus.gg/s...
Skyes PC Spec
------------------------
CPU: Intel i9-9900k @5.0 GHz
RAM: 64GB
GPU: Gigabyte Aorus GeForce RTX 2080 Ti Xtreme Waterforce 11G
End Titles Music:
Tobu - Candyland
Tobu Official RUclips - / tobuofficial
Hear me out, it's another big project, what if you put a tunnel through the dam to give lower access to the storage from the ground level?
ok THAT is an interesting idea! .. gonna use that :)
@@SkyeStorme Have only watched up to 15:49 and to respond to your confusion, the game choose the shortest path and then keeps that path as the primary choice path so long as there is room in the storage to be filled. Once that path choice no longer has storage space available, the game chooses the next closest path (Same goes for the spiral staircase as well). Wash, Rinse, Repeat.
@@SkyeStorme Also for the underground storage, rather than putting the 3x3 spiral to the North of the pillar you could just wrap the 3x3 around the pillar instead and potentially reduce the needed space form a 5x5 down to the 3x3 area only.
That's an amazing idea! There shall be much rejoicing!
@@dracconis69yep that's true, I was going to comment the same, but it still doesn't explain why they filled up the towers in a non-sequential order (first the closest, then the THIRD closest)
Re: warehouse choice
It seems that when picking up debris they are looking for the closest warehouse.
If the closest warehouse is full, then they are looking for next warehouse that is closest.
But NOT closest to the debris, CLOSEST to the warehouse that could not accept the item.
This would also explain why they are choosing the Archie side first then small-one&big-one direction.
Simply their pathfinding algorithm is looking there first on archie side then in then small-one&big-one direction
Agree, this is more obvious in the double helix test where beavers fill the three alternating warehouses on the same single helix. (13:08, btw they won’t prioritize the fourth because beavers returned to the rubble and start calculating the next hauling job, and now they find the first warehouse in the other helix closer. This pattern already starts when filling the third one in the helix, because some beavers returned from the second warehouse and found the first on the other helix is closer)
Along your line of thought, my hypothesis is that the beavers will trigger the rerouting when they arrived at the warehouse and then find it is full. In this situation, they start the finding the path again, which naturally points to the closest warehouse to the current full one. This is definitely a better choice versus triggering reroute at the exact moment when target warehouse is full, since it makes computation easier and covers more edge cases (like when another beaver takes out something immediately). To verify this, we could make the rubble way further from warehouses. Currently they are too close to judge the details.
Also it seems like the calculation is not just by distance, but a score calculated by various conditions with different weight. Conditions definitely involve distance, emptiness and building settings, could involve height or direction, or even random numbers. So if there is no clear winner in one batch of calculations, we could see scattering.
Also seems like they prioritize filling a warehouse before starting to fill an empty warehouse. So if they already started filling a warehouse they will finish filling it before moving on to a closer warehouse.
you also have to take into account how many beavers are moving items, so if enough are allocated to 1 warehouse but are further away, that warehouse might not be filled before the next is started on. only way to know is to just have movers next to where goods are to be moved then run the test
And the side choice (if the distance is the same) is probably on grid level closer to coordinates 0.0 are first.. So Archies side is probably the 0.0 side
if you fill in the center of the helix with small plattforms, you can place the pathes inside and have a 3x3 staircase instead of a 5x5.
43:58 The spiral staircases could also go around the pillar. That way the block in the middle isn't "wasted".
I thought the double helix would be around the central column since Skye is such a fan of symmetry
Had to catch up with this episode although it's very late. Would like to say something like "Dear Skye After adapting the "Skye Storme Effect of water dynamics improving flow rates " Which I named a waterfall after in your honour by the way.. Then I realised that your storage solution which needs refining clearly should be called the Skye Rabbit Hole. So have decided just to say.... MIND BLOWN... Far to many giggles for such a serious subject... Love you mate thanks to you and picture for bringing the joy 💚
Hey Skye. Computer programmer here. Routing “problems” like this one are typically handled via a set of rules which would coalesce to form a priority destination.
Obviously, we don’t have access to the source code, but the algorithm might look something like this:
1. Storage is not full. Not paused. Not emptying.
2. Storage is prioritized by haulers.
3. Storage is set to obtain.
4. Accept.
5. ???
#5 could be random (all storages tied), some internal ID (which the game assigns and uses to distinctly identify the building), or other priority such as “elapsed time since last change” (which could be a timestamp of the last moment you checked/unchecked a box on the building).
Due to cuts in recording, we were not able to see your setup process. If there is some sort of tiebreaker, look into the setup steps.
I think it could also be some kind of cache system on path finding. I consistently see that they are taking the same turn on intersections. It seems like the system caches the path for the first hauling trip and then uses that to calculate subsequent trips even after the storage is filled. Only when the subsequent path becomes too long, it flushes the cache re-calculates the full path. That would explain this scenario (13:37) (and subsequently all the others, but that's the best example), where after the third storage it calculates a new path and the closest one turns out to be the first storage on the second pile. Then the occasional anomalies happen because these calculations are made per beaver instead of globally, so there's timing differences when the calculations are made and thus, different amount of available storage (and distance to storage) affecting the shortest path logic and cache.
Granted, I am no programmer by trade, nor have I any kind of education on it. I only understand some very basic programming logic, but have slightly better than average (to be modest) pattern recognition skills. So, I don't really know what I'm talking about.
But If I am correct, it would mean that the most efficient storage system would be one that has only one path. But overall, I don't see randomness anywhere in the "accept" behaviour. "Obtain" seems to have some.
Also, there can be different calculations for short and long distance pathing. It seems that on long distances they decide the closest storage by a straight line and on short distances they use actual walking paths.
Yeah I was thinking its probably based on which he built first. It might take a few cycles for it to be how it should be
What is happening is that the haulers get a final destination when they accept the tasks. So all available haulers on the map start by delivering to the obtain storage, once they arrive at it they either deliver or if it is full they create a new shortest path route. The shortest path for the ones that are routed originally to the obtain is the third one down. Once the obtain is full the few that need to go back to get more then choose the shortest path as being the second one down. The reason so much gets put in the third one down is because of the massive amount of haulers on his map already pathing to the obtain right before it fills.
I'd normally wait to the end of the video to comment but so many questions.
Could it be the order in which the paths were laid (previous comment about districs and lag placing connecting paths without them could suggest complex path tracking), could it be radial distance from the hauling posts (fill of second row in third test area), could it be cardinal directions having an influence on fill direction (what would it look like filling from a path at one end, multiple paths, rather than central and also with storage N,S,E,W from the source pile)?
You really should consider doing complex testing outside of your production environment, is it time for a Matrix style "white void" for these R&D sessions? 😉
i think it could either be paths laid, or order the buildings are built. if none of those match, then there could be a function combining order built and distances
@@semperrabbit05I was thinking it might have something to do with closest to path/closest to district center or middle of map. Meaning they set the targets when they leave the idle area. Target pathing is usually based on closest to something. Just has to find what the something is.
@@CringeMeHarder there are different types of close. it could be cardinal(( north < south)
20:52
WHAAA??
I think a regular 3x2 spiral stair case is the best solution. sometimes you just don't need fancy, especially for storage, it should be simple and straight to the point.
Been waiting to comment, was sure you were going to spot the 3 spaces you left that would give space for another stack. Removing that spacing allows for 4 more stacks or warehouses.
From a programmer point of view, I think the behaviour is the following: the beaver choose the destination storage which is closest, if there are multiple I suppose N,S,W,E intersections are considered in order to simplify the calculation, in your examples assuming the green area as south and bad water area as north, for example on ground #1 N before S, for example #2 W before E; now it comes the trick, if the chosen storage is full the next storage is calculated from the previous chosen one, this is clear in example on ground #2 because if you have added paths connecting the rows they probably would have filled other storage instead of going full to the end of the row.
Also, for the double helix storage, you could also add a single story storage at the bottom and have some crazy Germans storage solutions.
Oh man! I can't wait to see this implemented! I love watching your brain work.
18:20 it is help. It restricts it to its position (relavtive to everything, but we aren't talking about other things, which is nice)
We only saw you testing distance, but you had 9 different starting locations.
To reduce variables, use a separate district, then control for,
- path laying order
- location of rubble ( compress rubble to one tile )
- left vs right
- number of beavers / only 1 hauling post
- distance from hauling post
- distance from town hall
My best guess with the warehouse situation:
A beaver is assigned to a task of storing materials to the closest warehouse to the debris tile, unless that warehouse is full, then it looks to the closest warehouse to that warehouse.
There is some sort of trigger that lowers the distance of a warehouse that has supplies in it.
Not sure if I'm explaining it well, but it would look something like this:
The first beaver goes a picks a random chance tile and puts it into the closest warehouse based upon that tile.
The next beaver then picks up another random tile, and the distance debuff kicks in due to the materials the first beaver put in there causing a chain effect.
Later on, a beaver gets assigned to store a material in the first warehouse, but when he gets there it's full, so he picks the warehouse closest to that warehouse.
The warehouse they just populated gets the distance debuff once again.
The randomness seen are from beavers being assigned after one warehouse fills up and before another one gets a single item in it.
Those beavers will pick a random tile and go back to moving it to the closest warehouse because there wasn't a debuff on any warehouse at the time of assignment.
At least that's how I would program it.
You could squeeze the storage "stacks" closer together, and then wherever they would face a column make a stairwell. But the stairwell can be more efficient with space if the stairs go up one level alternating left and right in a space of six and use spaces 7, 8, 9 to give symmetrical access.
15:54 it make sense if you think it like this,
beaver picks up package
goes to the closest storage, one to its left but it doesn't know if it has space available
then checks is full and then goes to the closest storage from the first storage it picked which is the one to our upper right.
the beavers in the game don't plan with total knowledge of their surroundings, they just plan the next step, that is why the beavers won't drink water even if they pass through a water tank while doing their errands or the bots won't refuel/recharge when they pass besides a fuel tank or charging station, they can't suspend tasks. So when is at the door of storage one searches for the closest door of storage to complete its currently assigned task.
I suggest you to study how the standard algorithm for simulations works
That storage complex is going to look very interesting. Its also going to be fun seeing it built in an actual lets play.
I did another experiment similar to what you did....but in a test world with only 4 beavers.
Sunflower seeds from a farm will go to the closest warehouse that is supply, accept, or obtain if the farmers are taking them
Only builders will touch supplies on the ground, and they seem to have no preference to supply, accept, or obtain. they will go to warehouses that are further and I have not figured out the pattern. Haulers will steal from other spots to take to obtain so when builders take from the ground and put in supply the haulers will instantly take them to put them in obtain.
You mentioned water falls and now i just want to see Beaver bridge + waterfall. Might be neat to watch beavers emerge from behind a waterfall in the morning and hide back inside them at night.
I would presume that the storage allocation is done at the point the goods become rubble (so the system checks they're valid for collection as storage is available) and then there is a seperate logic controlling which rubble the beavers collect (which will be the same for every beaver, hence the clumping), which isn't linked to the storage they're assigned to. So chances are the goods are allocated in a sensible way, but the intersection with the beaver rubble collection logic makes the overall process seem illogical.
Since you are starting with all mats on the ground and empty containers all haulers start with a path to the closest obtain storage. If they arrive at the obtain after grabbing the materials and it is full, they create a new path to the closest walking distance accept. The closest accept is the third one down because of the double helix. Some beavers finish go back to grab more and at that point the closest path is to the 2nd down accept.
"It was something else I was looking to get over the top of" HEYooooooooooo 🤣🤣🤣🤣 skye NEVER fails to disappoint. The man is comedic genius
OK warehouses, my theory is along the lines of when you spoke about about small water barrels and why beavers don't go to the closest storage coz of pre-assigned tasks and they have to complete that task first. So, each beaver (say 20 of them) will be told to go to left storage A all at the same for example, but because the beavers retain their assigned task as the storage gets full, then they will get reassigned to the closest next storage to the first storage assigned - they will get to left storage A, find out there is no capacity to put their carried item in that storage, so they go to the next closest storage at that moment of the reassigned task, and that maybe not be from the original task source, i.e. the debris pile. II appears the engine doesn't count the number of items, the pre-assign where beavers should go, for example, 4 storages of 5 capacity, 20 beavers will all be told to go to the same storage instead of 5 beavers go to storage A, 5 beavers go to storage B, etc.
At the risk of saying something that you figure out later in the episode. The double spiral is two paths that switching between requires back tracking to the top.
In the first stacked test one and three are on the same branch of the double helix, if a beaver was headed for one and it became full going to two would require going to the top or bottom to get on the other half of the helix, making three closer.
On the second stack they also favored the warehouse on the same spiral as 1 and 2 on this first stack.
However they didn't favor the bottom of the second stack in any way so the redirection idea probably isn't right.
But that spiral, where they use the down stairs furthest from the cliff is clearly preferred.
You can fit the spiral staircase in a 5x3 area, bacause you don't need access to the back. one of the helix's is always on the level you want and the second is just one level higher, so just add a staircase.
I think when the haulers finish a job they pick the nearest thing that needs moving and the nearest destination that can accept it, then they do that. So if they dropped something off on the right they pick the nearest item and warehouse. Its not the closest warehouse to the item but to where the beaver was when he decided to take the job.
So strange comment for you. Ussuing AI (Co-Pilot) I questioned the storage mechanics, this was the basic response I got, it referenced release notes and fandom wiki
In Timberborn, beavers prioritize filling storage based on several factors:
Proximity: Beavers tend to fill the closest available storage first. This minimizes travel time and ensures efficiency in resource management1.
Resource Type: If a specific type of resource is assigned to a storage building, beavers will prioritize filling that storage with the designated resource. This helps in organizing resources and ensuring that critical supplies are readily available1.
Equalization: Beavers also engage in a process called “equalizing,” where they distribute resources evenly across multiple storage buildings. This can sometimes appear inefficient, but it helps in maintaining a balanced supply of resources throughout the colony2.
Priority Settings: Players can set priorities for different tasks, including storage. By adjusting these settings, you can influence which storage buildings beavers will fill first3.
These mechanics ensure that your beaver colony remains well-organized and prepared for various challenges. Have you noticed any particular patterns or issues with how your beavers manage storage in your game?
1: Timberborn Early Access Update 3: Timberborn Guide for Beginners 2: Reddit Discussion on Storage Equalization
So the equalization would be the confusing randomization of filling random ones, as well as the proximity of resources to the warehouses. So. .. may the odds be ever in your favour . . .
The only question is if the underground storage goes in at the bottom layer.
Going to be fun seeing you build this system several times in the next episodes because following a blueprint is hard.
Especially one you made yourself.
20 minutes in and my theory is... The beavers go to the "closest" storage facility, So A (the location of the item) to B (the assigned storage).
But if the storage fills up, prior to the beaver getting there, it is still traveling on that trajectory, so it keeps going to the next/nearest storage on that line, WITHOUT turning around. Which in this case would be C.
And maybe the beavers choose the left versus right side, depending on where they originally came from. If they came from the left side of the map, to the central cross road, the closest storage to them would be the first one one the left side.
Maybe there is so much lag between the calculation and that is why the beavers continue going in the one direction, rather than altering the course to go to the nearest storage on the opposite side.
so here I am at 11:30 watching top and 3rd filling up, and watching the stairs being used - double helix staircase, first and third are same staircase. You have to switch stairs to get to 2 and 4. To try and influence which stairs are used, maybe path length is involved, but path on top of 1st storage unit to shorten path to second stair spiral.
you could put impermeable floors on top of the leves of the pipes to cover them up and make it look better.
OH my, I've just restarted this map on my Steam deck, and I found a hidden mine on the hill where the two waters come together before coming down. There used to be a little hole. They have changed that to 14 natural overhands with a mine deep under it,
Ignore me if you have all already found it, I have no friends to get all excited with, so I thought I would say it here lol,
Probably something to do with distance from the district center. The filling order matches the paths (first two tests)
Maybe a bias to which turn to make when distance is equal
now at 15:30, beavers pick up boxes and head for 1st unit, but unit is full before they arrive (check for destination at pick up when it was not yet full, no accumulation of data in transit, and check again at unit entrance when full), since they are at the unit door, which is next closest unit to go to in same path direction first, then reverse path direction.
My thoughts.... On the Helix staircase, Beavers fill up level 1 on side A. then when it is full, they still try to take the produce to the first storage, but it is full, and then the next closest storage is level 2 on side B.
I would say that the beavers are coded to continue walking in the same direction along a path rather than doubling back, even if the warehouses are the same distance away.
29:40 You could make that center storage work if you would use platforms instead of levees at entrances. An even make path across all storages and move that second row one tile closer to first row of storages. Except if you want to make pillars of ground tiles :)
I think the way it works is the worker gets an order to pick up a thing, it then calculates the path to the pick up and adds the distance to the closest storage. The path finding to the pick up begins from the hauling post. If there is a tie in distance their algorithm prefers to the right because that is the positive/positive movement in a grid quadrant (the math is easier (done in fewer steps)).
This could also be a result of coping delivery companies' algorithm which prefers right turns as you don't have to wait for as many lights.
Have you thought of doing 2 wide platforms to hide the stairs going up into the towers on each side of the middle. Hide the stair channel that goes up and only have a small spot open where the stair comes up to the top.
On my last save I made a waterwheel generator with a large underground space underneath. I then placed all the industry and storage inside stacked in that underground space. I found out afterwards that I didn’t make the generator quite large enough, so it was a little underpowered. But finding how to make space for the industry for 6 bot assemblies plus its industry was lots of fun!
Can't wait to watch you actual put in the new storage
Dunno if you figured the tower scenario out yet but the path to get to the 2nd floor down has to go up first then down to get to the 3rd level from the floor. In the double towers they are going down first and every down is the odd floor. If that makes sense
My theory is that they calculate the next one when they get out of the warehouse that got full, meaning that the closest one (starting from the north) is the one that gets filled next, but this doesn't seem to reflect what we see on the vertical one...
Perhaps the order of building also has an effect when they are in the same place (on top of each other)
Dear Skye,
I don't understand how they made their thing for the distribution on the storages. Official wiki just says it's the nearest storage, and some comment on their Discord point that it's the nearest storage when the beaver starts the tasks (way before actually picking the resource).
Well.. that's what they do.. they are picking an item for storaging. Probably knowing the closest storage. When picked the beaver willn controlling if the storage is still available. If not, search one the next storage in the area of the full storage. When 2 storages are even close together. 1 storage will have a lower coordinates number (like 0.0) and be the next storages. When there are to much beavers on there way to the same storage. The beaver will search for the next warehouse that is the closest to the actual location. That can be an other locatio, depending on the distances between storages
i think the way they are storing it is in priority order. once the second one is consumed, the third one is the next closest to them in terms of pathing.
Dear Skye,
So the double helix storage... would it be doable if you made a "fat" double helix? The inner core the 3 long like normal and then 5 longs on the outer? A Helix within a Helix. That way you would have plenty of path space AND the storage could be centered on the helix.
4:00 The Storage just needs to be 1 higher to accommodate the spiral, so you could have storages on all 4 sides, just it would be 2 at even heights step and 2 at odd heights.
I think there's 2 things: shortest path and current storage. If there's already stuff in one they might prefer this one to an empty one.
Other thing is they might calculate path when they pick it up so they see space somewhere and go there to realise there is no space anymore and recalculate from there. That would explain why they would continue on one side.
Another thing, The left and right might be due to where they picked it from. If you had only one tile of path leading to the pile that would remove that parameter.
just re-watching this while I brainstorm your double helix idea with large warehouses. only way I have been able to get the helix to work was using a x4 overhang across front of warehouses the a x3 overhang to link them across. After an hour building rebuilding this these were only over hangs that worked giving smallest space used. but basically is the same area as your previous storage video using platforms
If you want some more information, look up the A* ("A Star") pathing algorithm. I warn you, it's a rabbit hole. Computer pathing logic is a PITA, and I don't mean the unleavened bread. There's another game called Screeps which has a whole thing about pathing logic with it similar to this problem.
What I suspect is going on, is that it's basically "trying" one direction first from the District Center, and if it's really close in distance then it will be fairly random as you're seeing. It's possible it tries directions in a NESW priority direction attempt, or it tries one at random. Either way, it then "caches" that result, which then produces the same priority fill each time, until a "refresh" is requested (possibly each new day, or when a new beaver is born, or when a new path is added to the network).
21:48
Time to send this video to the Dev team and tell them that they have some explaining to do!!
Sorry for missing the Premier last night got Stuck in football traffic.great video!
Thanks Phillip .. we missed you :)
@@SkyeStorme I know
My guess is at first they choose by distance from the item to the warehouse, if distance is the same and all are empty its RNG, once the first item has gone in, the next beaver prioritises that one because it has items in. when one warehouse fills the beavers that were en route to that warehouse, with items in hand, then remap to the nearest warehouse to them at when they are stood at the door of the full warehouse, with items in hand. Then the next beaver decides to go to the nearest warehouse that has already got items in, so they don't end up scattering the items into multiple warehouses
I think the only advantage to a pile in the central area... would be Dirt, because it's used in the mud bath.
Dear Skye and Picture,
It is based on patching from the district it seems like. Whichever is closes via the shortest path gets started first. so when you were on the mountain the left path may have been shorter than the right. When you were on the ground the left path may have been shorter than the right. Watch the haulers and which way they run off to. It is based on either hauling posts or the city district and patching and how close it is to one or the other it seems like. The haulers take the shortest path first.
Are you going to store water and other liquids? Because you may want to think about putting a layer of large drums under ground for misc liquids and have something on top of them to then out the water.
But on coordinates level one is allways closer to 0.0
@@PauxloE yes on the ground it was the same from the pile of stuff to the warehouses. However it isn't the same from the hauler posts to the pile of stuff on the ground. Some beavers may take one path and others may take another path. I think the path distances dictates which one they go to. If beaver A has 100 feet to go from its hauler posts to warehouse C and 120 feet to warehouse A it will go to C. While beaver B has 150 feet to go to Warehouse A but 160 feet to Warehouse B it will go to A. Sorry for the confusion. But it is about the paths from each warehouse and hauler post and patching and all that.
Dear Skye, Beautiful... Just Beautiful.
But the priest of the holy spud miss their temple. The holy spudonomicon needs a resting place.
Signed a concerned in awe potatostian
8:30 basically you want tree access with the property of access in the order of n log n distance.
12:21 my guess is 1 and 3 and 2 and 4
15:00 i think your beavers will fill the storage in order you assigned them.
Loved the store idea it was great.
11:00 if there is a queueing limit like with getting water/biofuel, these experiments are no good because too much has to be moved at the same time and it is more affected by queueing than anything else.
Could potentially save space on the side staircases, since they won't have to be double Helixes. They only need to go one direction.
4:52 I think the beavers need to figure out a way to make elevators
You forgot the impermeable floor tiles 😂😂😂
1. The builders are the ones picking things up from the ground. So Could the behaviour of prioritize what storage to go to differ between builders and haulers? Maybe they obtain and accept choices only effects the hauler?
2. Would it work with a 3x3 helix staircase and then a 5x5 helix on the outside of the 3x3 to access the buildings/storages? Probably really inefficient and a waste of space but it might look cool.
I wonder if part of it is due to the direction the beavers arrive from. Arrive from the right right side first, arrive from the left, left side first
For a while, I thought the order they filled a warehouse might have to do with how many tiles it was from a city center, but that doesn't seem to be the case after a few tests in the video.
looks like they choose the closest storage, and if there are 2 it randomises, while the next storage looks to be calculated from the previously used storage? maybe?
We need "If you can understand the beaver" slogan merchandise
Maybe it has to with the moment of completion. Codewise they are numbered, maybe they just go to the next integer in the database. I know that the stacked ones don’t track in that case because you can only complete them in order, but maybe it has to do with path connections - they could be stored as integers as well.
In fact this makes some sense, as a completed building only comes available when it gets a path hooked up, so maybe there is another database with path connections, and those get addressed logically by first come first serve.
There is sone merit to that if you see the behaviour between accept/request doesn’t differ anything.
@13:30 seems like it is going to the first and then in directional order on that path.
I guess three tiles per level counting the staircase as two tiles. That is actually the minimum if you want a single staircase path with access to each floor. Ofc with large warehouses you only need access to every other floor. Easiest modification to the double helix would be to put a 2-overhang on the 3-overhang before turning the corner, which makes a 4x4 double helix with straight double staircases. Can't get more efficient with a single path. Most efficient with a double helix would be a 6x6 one with 4 staircases on each side, but you would have to start the second spiral in the middle of a side rather than at the corner, which looks odd.
Dear: Skye
Beaver logic is not Human logic.
The Storage idea is an interesting one for sure. looking forward to the next video.
Seems that when the warehouse becomes full they are already in that direction and now closer to the one next to it.
19:46 I bet it’s a something due to x y cords. Probably the way it’s coded and then the z probably code changes on each cord.
If you move all your industry ontop of Archie’s back then you can put a lot of storage on there
That was my thought as well. x,y,z coords and I think the distance is taken from the district center and not the hauling posts as well. Would need to test though
Just a thought...Are the beavers filling the storage on the basis of distance from the district center? Or a combination of distance, fill type request (accept or obtain), and proximity to source?
I think maybe you may have the space to add a 9th tower of storage in the middle of the front row? That'll give you 17 in total.
13:30 so they are all using one side of the spiral stairs
the way those are built is it will only be on the same side every 4 layers, which means every other large warehouse
so as they go down the stairs to reach storage 1, when it fills, the beavers will logic from their current location to storage 2, which is the closest from where they intended to move them
and then again, continue to aim to fill it up. Once 2 is full, the logic over to find 3 which again is only two stairs further away
the few beavers which (and this is random) had their logic take them to the one opposite 1, or below 1, or just... not 1. They are the ones you see as "randomly getting a few" and their brains do the same, filling the odd warehouses despite the bulk of beavers going elsewhere
maybe there's a limit on how many beavers can be assigned to fill a storage at any one time
but i think once a beaver has selected a warehouse to fill, it won't swap until it's full, then it'll look for the closest based on it's previous target warehouse, and not just the new TRUE closest
the beaver's need to swap what they are hauling, or just have a pause all together, for the logic to reset
The reason for the pattern in which they fill the storage seems simple.
Well, of course there must be some simple internal rules that make them chose the storage to the right first.
But after that it makes sense for the beavers to fill the second storage on the right next.
You're only looking at the relative positions of the storages and the pile of items ... but there are beavers already on their way to the first storage on the right when it reaches max capacity.
Those beavers are closer to the second warehouse on the right than the one to the left.
And I assume the other beavers follow them and fill up one storage at a time to reduce the number of variable item stacks and simplify item management in the game.
The thing with the weird hotbox might have been the regular pile & not the underground one
Dear Shye
Thanks for sharing 👍
18:48
Maybe the orientation has something to do with it?
Like is it setup north to south or east to west......?
Is their pathing different when cleaning rubble (builers) compared to transporting (haulers)?
Yah it's very cool
once an OBTAIN storage unit has been loaded Haulers cannot take from it.
Just set the "deep storage" to supply. They wont fill a supply warehouse until everything else is full.
seem to recall someone in a video saying the small storages only accepting 1 beaver/bot to queue up to bring or take from it, wonder what the numbers are for larger storages, if thats even a thing, and if so, that's playing some part along with pathing somehow in which storages are being selected
Dear Skye, thumb, thump, finger on the screen of me phone;
As an observer of this game and a few select players on the tube and twitch, and not being able to play as i have 5 broken PC's from a decade ago.
The game seems to run in a clockwise fashion, like tree planting, crops, the harvesting the same way, so storage may be the same but from the orientation from the District Center, then path, and /or where it comes from, floor, building etc, still in a clock wise orientation. Could be the same for helix stairs and other pathing structures?
Maybe build a clock face of storage, with your mass of pine resin in the middle, multi-path to the times of clock, destroy and see if 1 is before noon etc?
Also in the storage demonstration/r&d, you used levies/dirt (at future proposed building site)but using wooden columns/pillars, you can use them as part of the structure and walk through them?
I enjoy your work, sort of the Tony Hart of Timberborn. ;).
I was going to say Rolf...
But he turned out to be a bad man. Sorry tmi.
Hits return... Click
Now back to my hole.
Runs back,... Hi, Picture.😊
Now back to hole.
This is a good suggestion - I was also thinking it might be partially based on direction - North South East West as that was one change between the two setups so doing a clock style layout and seeing what happened would be very interesting! Failing that do 8 storage all aligned North, South, East, West (so total 32) and see what they do when you destroy a pile of stuff in the middle.
I think the building you were thinking of was the centrifuge..
So from your flat tests it looks like a slightly naive path finding algorithm but i would guess it selects random directions at intersections when looking for storage. Then it looks like that is probably assigned to the item, then a worker goes to the nearest item and i guess a few came from a different direction? Or pathed at a different time (busy on another task when you deconstruct) and go to a different pile which explains the random small amounts but it is very strange.
What you really need though is to constrain your variables a bit better and test changing one at a time but I would guess that could take some time in a dev world or seeing the actual algorithms in use.
we got a beaver on snow white to give the lesson 🤣🤣 Jk love the thumbnail
Are you dead set against using any mods? There is a very nice “ladder” available in mods that’d solve the double helix storage issue.
Put the helix staircase around the column
Dee'uh @skye,
Oi, yuh tests had a whee bit of a flaw in 'em mate. Ya be needing to give them another go to verify consistent inconsistency.
Test one example: if upon reset the storage 2 was filled before storage 3, then you'd know that there truly is no consistency.
So 'ave another go at it Gov! Blimey, I wasn't taking tha piss, just trying to give a hand. Cheerio!
Would the double helix staircase going around the pillar still give you access to all the levels? Might look a bit better.
Is the distance between their hauling post / workplace and the closest storage probably factored in?
Warehouse choices, I wonder if this is some kind of Depth First Search vs Breadth First Search kind of thing?
Skye - If you put single high platforms (with paths) in the center of your double helix, wouldn't they be able to crossover much easier and reduce pathing?
Dear skye, what sort of seagul defence/pummeling systems does have the 633 squadron have available?
Isn't it the same way for the planting pattern?
Look at how they approach planting out a new area with crops, pretty sure that is the same way...