Per Aspera: A detailed look at logistics

The “Too Long; Didn’t Read” version of this guide: Build more worker hubs. Make sure that they have drones attached to them. Done — please leave a 5 star review, thanks! 🙂

Note that this guide is broken into two major sections — the first portion is free of story spoilers, and only contains minor tech tree spoilers. The second section is a step-by-step walkthrough applying the lessons of the first section — it, obviously, does contain story-line spoilers.

The bulk of this guide is written from the perspective of someone playing the campaign at the normal difficulty. Some notes will be added, especially to the walk-through, where things change at the Per Aspera (“Hard”) difficulty.

The logistic system in this game is highly similar to the logistics system used in the game Settlers 2[en.wikipedia.org]. If you are interested you can see a Lets Play of this game here. However, in that game haulers (the equivalent to worker drones) cost nothing to produce and gameplay all but forces you to use them in vast numbers. Typically, a Settlers 2 game will involve 5-10 haulers / building. The author always wondered why haulers were free; this game proves why. 🙂

Are your mines stalled due to full outputs? Either you don’t have enough demand (in which case it shouldn’t be an issue) or you need more worker drones.

Are your factories stalled for resources? Either you don’t have enough mines (a legitimate resource shortfall), or you need more worker drones.

Are your factories stalled due to full outputs? Either you don’t have enough demand (in which case it shouldn’t be an issue) or you need more worker drones.

You get the picture — in Per Aspera, the solution to most problems is “Build more worker drones”. I know it seems like you should conserve resources and only build worker drones where they are “really needed” but… If you approach the game in this direction you are going to find yourself extremely frustrated. You can and should expect to build at least one drone hub for every 3 other buildings to ensure the smooth flow of resources in the game.

Each drone hub can host one, and only one, drone. This drone is automatically assigned a territory, based on the road network (also automatically generated) and the presence of other drone hubs that have a worker drone present. Keep in mind that the game will automatically redraw territories as new drones are assigned to drone hubs and existing drones are destroyed (by tick damage, simulating the harsh Martian environment). Adjusting the territories in this way ensures that all nodes can be serviced by at least one drone — it also, unfortunately, means that it is very unlikely that the developers will ever implement the ability to manually specify territories.

Drones can “service” (pickup goods from or deliver goods to) any node that is contained within its territory and the nodes immediately adjacent (connected by a road) to its territory.

The actual process of shipping goods is straightforward:

1. The drone checks to see if there are any goods that it can reach that it can assist with moving (“ship”). If there aren’t any, it returns to the drone hub and idles.
2. If there are multiple goods that it can ship, it first checks the destination of the goods. If one or more of the goods waiting for shipping is destined for a building with the priority flag, it will choose one of those goods to ship.
3. If at this stage there are still multiple options of goods to ship, then it chooses one according to an obscure algorithm that I have no insight into. At a guess, it includes several factors, including minimizing travel time (if there is a good to be shipped in the same node as the drone it will almost always move that good next, for example) and likely how long the good has been waiting for shipping.
4. The worker travels along the roads to the node where the good to be shipped resides.
5. The worker picks up the good and travels to the selected destination, where it drops it off.
6. The process repeats.

There are some things that are still unclear to me (developer feedback would be welcomed here) — in particular, it isn’t clear whether the route that a good will follow is fully determined at the time that the good starts its journey, or the path is recalculated each time a drone considers shipping a good. I suspect that goods recalculate their path each time a worker picks up the good, but maybe not.

In any case, keep in mind that the good can be picked up from outside its territory and delivered to a node outside its territory — this will happen more often than not.

I started a new game (Campaign, Normal difficulty) and proceeded through the game until I got to the point where the game gives you enough resources to choose your own path. At that point I recommend that you do this:

This is the “Worker lens”, accessed by pressing <F1> or selecting the icon immediately to the right of the compass. Its somewhat helpful in troubleshooting logistical problems.

The first thing you’ll notice about the is that all the territories are colored, ranging from yellow (0% “Sector Load”) to red (100% “Sector Load”). You should completely and totally disregard this information — at best, sector load isn’t harmful, but more often than not, sector load is grossly misleading.

You see, what sector load measures is the percentage of all goods shipped that pass through that particular workers territory. So, for example, if 100 goods in total where in transit at a given time, and 30 of them passed through the territory of a particular drone, then that drone will have a “Sector Load” of 30%. This becomes particularly obvious once you build Hyperloops — since the vast majority of goods pass through multiple Hyperloop terminals on the way to their final destinations, all the territories that contain Hyperloop terminals have very high sector loads.

There is another measure provided for drone hubs: When you click on a drone hub to pull up its details, a horizontal bar graph is displayed showing “Load” or “Worker Hub Load” if you view the tooltip. This shows (as absurd as it seems) the amount of time that the worker spent within its territory either moving or traveling to move goods. If a worker hub spends all of its time moving goods outside of its territory (from an adjacent node to another adjacent node, without passing into its territory in between) this number remains low.

At best, the “Sector Load” metric can be used to find areas where a large percentage of your goods pass through — but it won’t help identify problem spots. And “Worker Hub Load”, while marginally more useful in identifying overloaded drones, still isn’t even close to being definitive — an overloaded worker hub can have a low “Worker Hub Load”, while a worker hub that is performing very well can have a high “Worker Hub Load.” These issues have been reported to the developers, and hopefully we will see some improvements in these metrics in the future.

The next thing to note is the territory claimed by each hub. In the screenshot, all drones except three have exactly two buildings within their territory (including the drone hub itself). Drone hub 1 has four buildings (two solar panels and a chemicals mine), hub 3 has only one building (the drone hub itself), and the central hub has three buildings (an aluminium mine, a maintenance tower, and the landing site).

This is a good thing!

Looking at drone #1 (upper left), the drone can do the following useful things:

1. First, it can ship chemicals from the chemical mine (inside its territory) to the plastics factory (adjacent to its territory).
2. Next, it can ship silicon (adjacent) to the glass factory (also adjacent).
3. It can ship plastic produced by the plastics factory (adjacent) to the glass factory (adjacent), getting it closer to its destination (either the maintenance hub or the landing site)
4. It can ship glass produced by the glass factory (adjacent) to the worker hub just to the south, moving it closer to its final destination (the worker factory).

Looking at the drone just to the south (#13, whose territory only includes the plastics factory), the following options are available:

1. It can ship glass (if delivered by the drone 1, to either its hub or the adjacent solar farm) to deliver to the worker factory (adjacent)
2. It can ship plastic (inside) and deliver it to the landing site (adjacent! Really — there is a road connecting the drone hub directly to the landing site, so its adjacent!)
3. It can ship chemicals from the chemical mine (adjacent — again, there is a direct road between, so its adjacent) and deliver it to the plastics factory

As you should see by this point, a large number of drone hubs produces more potential paths to ship goods than a casual glance at the territory map would indicate. With so many paths available to the game, the odds that a good will remain “stuck” in an output bin (leading to production stalls) is quite low, and the odds that goods will be delivered “on time” to their destinations will help eliminate input stalls.

In the absence of storage centers:

1. Goods only move when they have a specific destination in mind. This destination is determined by a building demanding a particular good due to having empty inputs.
2. The number of goods of a particular type that might be moving on the map is equal to the demand for that good as shown by the “Current Demand” when you hover over a resource in the global stockpile ticker, plus the demand generated by construction sites, and special projects (which you can’t see as a number in the game).

As of Patch 4, with the December 31st Hotfix, there is currently a bug with the global demand accessed from the ticker — demand is only counted if it has successfully identified a good to satisfy that demand which is now in transit. The developers have acknowledged this is a bug and will fix it soon.

Storage Centers change the dynamic somewhat, but likely not as much as you would expect — see the section on storage centers for more discussion on this topic. For now, I’ll assume that there are no storage centers effecting demand, because this is my recommendation for the pre-Hyperloop era.

Be aware that each “Input” slot on a building can be in one of three states:

1. It can be filled with a good, in which case you’ll see that in the building UI (e.g. 2/6).
2. It can be empty, with a good on the way (in which case you’ll see a blue-grey line connecting the building to the good that is being delivered.
3. It can be empty, searching for a good to fill it (in which case you won’t see a blue-grey line connecting the building to the good).

Obviously, it can be difficult to distinguish between case #2 and #3.

So, if you have two food factories (each with 6 slots for water) and one colony (also with 6 slots for water), then you will never see more than 18 units of water (as water is never required at a construction site) moving at the same time in response to demand for factories. Most of the time you’ll see less than that, because some of the input slots will be filled with water either being processed or waiting to be processed, but that’s the maximum.

In my experience, this is very counter-intuitive to many players — they see a building that needs a resource to produce something, and then see sources of the required material sitting idle, and conclude that the game is buggy. It isn’t; well, it might be, but in my experience it isn’t. The most likely reason for this kind of behavior is that the resources are in transit to the building but they have gotten stuck in a building somewhere waiting for transport. See the troubleshooting section to work out what’s going on and how to correct the issue.

As alluded to above, the number of goods moving on the map might be less than the “Current Demand” — this happens when you have a resource shortfall. Only when this is true will adding another building make result in immediate movement of goods. Again, this is confusing to many players; its reasonably to conclude that if demand is high then resources should be snatched out of any building as soon as they are available. This, however, does not take into account the need for goods to transverse the map, and that takes time — indeed, it isn’t uncommon for a good to spend far more time in transit than it did being produced.

There are four types of logistics failures that you may run into:

1. Insufficient demand
2. Insufficient supply
3. Logistics bottlenecks
4. Bugs

Insufficient demand exists when a source of supply:

1. Has 6/6 in its output
2. When selected, has no silvery grey lines leading from it

Note: This is a common source for the complaint “My factories don’t have any resources, but my drones are doing anything — its a bug!”

Example:
The reason that this Chemical Plant is suffering from an output stall (6/6 in its “Output” stopping production) is that all existing demand is being fulfilled from other (closer to the demand source in this case) sources.

To fix this problem you must create additional demand — building a new food factory, for example. Note that doing so will likely reduce your production rate (certainly in the short-term, and perhaps in the long term), but may increase your gross production.

But probably not.

When players are actually looking at the output of a stalled building like this it is generally because they have a need for the resources that this building produces elsewhere (for example, you might look at a water mine when a food factory isn’t producing food because of lack of water). In this case, the full source building indicates that there is a logistics bottleneck elsewhere, and focus needs to move to finding and fixing that bottleneck.

While building an additional consumer for the good being requested will cause additional goods to be shipped of the type that the building being inspected produces (although not necessarily from that specific building), its highly likely that the goods will find their way to the same bottleneck as the first shipment of goods did, and the actual problem (lack of production) won’t be fixed. You can limit the likelihood of this happening by building the new consumer building in a very different area than the existing consuming buildings (so that the path that the new goods take is very different from the path that goods to the existing buildings), although distributing your production like this make its more difficult to manage your production buildings.

If, however, you don’t have a building that is stalled due to lack of input, then you are “demand limited” — and that’s a good situation to be in! That means that all of the buildings that consume this resource are full and producing with a high (close to 100% efficiency) and you have reserve production capacity to expand your base further. This is particularly common for mines, especially silicon mines.

Insufficient supply exists when

1. A demand source doesn’t have a full set of inputs
2. When clicked on, there are no silvery grey lines extending from the demand source (alternatively, there are silvery grey lines — but the total goods that they point to doesn’t add up to the expected demand)

Example:

This Spaceport (“Spaceport #2”) is demanding parts (gear icon), but no parts are available. Note that there are parts available on the map (10, to be exact), but some combination of the following are true:

1. The parts are at another job site
2. The parts are traveling to another job site

In this case, 2 parts are at worker factories (both are prioritized, so their demand trumps the demand from the non-prioritized spaceport), and the remainder are at various Hyperloop construction sites (one each). Note that, barring the use of priority, the AI will generally distribute resources evenly across all sources of demand. When supply is very limited, this will result in a large number of sites that have some of what they need, but not all, and progress is very slow.

To fix this problem:

1. Ensure that your existing sources of supply are working at high efficiency. If they aren’t you need to troubleshoot that first.
2. Build additional supply, if that is possible (if your shortages is of, say, water, then this may not be an option).
3. Eliminate some of the sources of demand — you can use the power button to turn off anything, including construction sites. This will cause any resources at that location to be released for redistribution as well, which may also help.
4. Designate one of the sources of demand with the priority flag — this will ensure that the limited supply is concentrated in a single location, where it can actually do some good.

As of Patch #3, there is a bug that makes limited supply more serious — see the bug section for more details.

Note

When the game identifies that insufficient supply exists demand generation works differently — in particular, even a prioritized factory that consumes the good won’t generate demand for the good beyond what it requires to fulfill start production (if it isn’t running) and one additional production cycle. So, if you are short of aluminium, a prioritized Electronics factory will only generate demand for at most two units of aluminium at a time, and that only if it is currently not producing due to lack of inputs.

One one level, this makes sense — it ensures that at least some of the scarce good gets to other consumers. On the other hand… This practically guarantees lots of flutter, and flutter is bad (see the section with that name in this guide). In my opinion, this should be changed, and prioritized buildings should generate their full demand even during a resource shortage. If this stalls the player’s economy, then so be it, but the players priority flag should be respected.

The symptoms of a logistics bottleneck are:

1. When clicking on the resource consumer, there is one silvery grey line for each missing input. See below for an example.
2. One (or more) of the nodes pointed to by the silvery grey lines contains more than 1 or 2 resources.

Practical example:
The steel factory in question should be generating demand for 16 items — 12 iron, and 4 carbon.

Checking these lines out one at a time reveals:

2 iron and 1 carbon in this node.

1 Iron in this node

7 iron and 6 carbon in this node

3 carbon in this node

1 iron in this node

1 iron in this node

Adding this all together, we have 12 iron and 4 carbon — exactly as expected. Sometimes you will find extra resources (e.g. you expected 12 iron and 4 carbon, but found 16 iron and 8 carbon). In this case, you need to check other sources of demand (another steel factory in this example), as resources demanded by the other demand sources may be coincidentally be occupying the same node.

The primary problem is clearly the third node on the list (“Worker Hub #1”) — almost half of the iron and carbon destined for this steel factory is “stuck” in this node. Switching to the worker lens and allowing time to proceed reveals…
Part of the problem is that this node (“Food Factory II #0”) is considered adjacent to the node with the bottleneck (it has a direct road connection) but is quite distant — the worker is making long trips between these two nodes, slowing throughput.

Next, this node is adjacent to the hyperloop. While a hyperloop doesn’t demand resources in the sense that a factory does, it serves as a destination and source for a large amount of resources.

Finally, the fact that this node is adjacent to two different worker factories, both marked with priority flags (one in territory, one adjacent to the east) is another issue — goods destined to a prioritized destination always get moved first, regardless of how long other goods have been waiting.

My recommendation here is twofold:

• First, add a new drone hub to block the long connection to the food factory. If the drone never takes this long trip, things will be much better.
• Second, add a drone hub to the south, adjacent to the solar farm. The new node will be adjacent to this node but not adjacent to the worker factory (make sure there is no road between this node and the worker factory!), so it will be free to pickup and move goods towards the steel factory

The new nodes are scheduled for construction. Note that adding the node to block adjacency for the food factory immediately improves matters somewhat, as it blocks access to the food factory.

Immediately after the two new worker hubs are completed and staffed

Fixed — the bottleneck is cleared!

Production is up dramatically at the factory.

A bit later and the impact on overall production is obvious.

Yes, this is a time consuming process — but this is how the game is supposed to work. You can limit the number of times you have to go through this involved troubleshooting process by… Building lots of worker hubs. 🙂

Seriously, the more worker hubs you have the less likely problems like this are to occur, simply by virtue of having a large amount of drones available.

As of patch 4, with the December 31st hotfix, I am aware of one bug that directly impacts logistics:

1. Resources, almost always travelling via hyperloop, will start “bouncing” between two nodes, making no progress.

To be crystal clear: There are many other bugs, some of which are catastrophic, that impact other areas of the game. As of patch #4, this game is not production quality and if it were released at all it should have been tagged Early Access. The scope of this guide, however, is logistics, and that’s why this section is so short.

A good bouncing between two Hyperloop terminals counts as “in transit” for the demand source, yet never arrives. With large scale projects, especially late game space missions, the odds of at least one good getting caught in a cycle like this is nearly 100%, and it stalls progress indefinitely. To fix this problem:

1. About 90% of the time, turning off both the hyperloop terminals that are involved in the “bouncing” will solve the problem — in ~2-3 in game months.
2. Turning off, then back on, the source of the demand will always fix the problem immediately.

Because of this first bug, the developers released patch #3 which… Automatically toggles (in the background) long-running sources of demand, mostly impacting construction sites and spaceports.

While this doesn’t eliminate the problem (goods still bounce between Hyperloop terminals) it does limit its impact — and patch #4 reportedly reduces the frequency of the bouncing, although it certainly still happens.

There are a couple of additional bugs that indirectly impact logistics:

1. Storage facilities, including landing sites, will request goods from outside of their road network — e.g. goods that cannot be delivered to the storage center. This is mostly a cosmetic bug, as it only seems to effect goods from scrapped sites, and my recommendation is to not use landing sites (which is, by far, the most common source of this bug). The demand doesn’t seem to cause any issues other than the fact that the scrapped site never clears (because the goods are never delivered) — the storage center continues to demand other goods and release goods as required.
2. The behavior of storage centers, while working as designed (so, not a bug by definition) is odd enough to justify listing it in this section. See the storage center section for more details.
3. The “Current Demand” shown in the global stockpile (along the bottom of the screen) doesn’t include unsatisfied demand (demand where no good has been identified for transport to the building generating the demand). This is an acknowledged bug, but it doesn’t actually effect the movement of goods on the map.

Flutter occurs when a good switches continuously between destinations, making no progress towards either. Such a bug is common in RTS games, and is always catastrophic — both a good and a worker are “stuck”, and under the very best of cases it slows production dramatically. In most cases, though, if flutter occurs at all it eventually spreads and all workers and goods are stuck fluttering.

Hyperloop “bouncing” acts in most respects like flutter — that’s why it is so bad. I don’t think it is flutter, because if you focus on one of the two Hyperloop terminals involved you can see that the good is always trying to go to the same destination (it just isn’t able to make progress), but this is at some level irrelevant — it has the same effect.

This game tries to avoid flutter by only moving goods with a specific destination in mind and, with very few exceptions, that destination is engraved in stone. So if a good from a far distant carbon mind is routed to a steel factory then that carbon is going to that steel factory, and if that means that production stalls… Well, you can’t make an omelet without breaking eggs, right?

It is very tempting to say, “Oh, what would be the harm of having the steel factory re-search for resources, and if it finds one closer, grab that instead,” and perhaps this really would work better. Or, perhaps, it would result in flutter occurring 1 time out of 1000. If its the later, then making this change would be catastrophic, because its likely that hundred of thousands of goods will be produced in the course of the game.

Avoiding flutter is a top priority in any RTS-style game — its just much more important in this game, because flutter tends to be harder to detect.

As mentioned previously, flutter will occur in this game if you have a prioritized factory (which, during a resource shortage, can only generate demand for enough goods to fulfill the current production cycle + one more) and other, non-prioritized consumer of the same good. Due to the cap on demand from the prioritized factory, goods will start to travel towards the non-prioritized consumers, but before it arrives the prioritized consumer will have consumed the limited quantities of goods that it can demand and re-tag the good. This may occur in transit, or it may occur after the good has arrived at the non-prioritized consumer — either is bad.

Storage centers are… Implemented oddly at the current time (Advanced version as of 1/15/21). I’m tempted to just omit this section altogether, because I think big changes are coming, but…

At the current time, I strongly recommend against constructing storage centers at all. If you must construct them, at least wait until you have a Hyperloop system up and running, as they make the “quirks” of storage centers easier to live with.

Per a developer, storage centers generate demand for at most one good of each type they have space for. Once that good arrives, they generate a new request (if not at maximum), and so forth. Furthermore, goods destined for storage are only shipped when no other goods are available for shipping (e.g. they are “low” priority) and goods destined to storage centers are, unlike other goods that are in-transit, may be “stolen” (have their destination changed) by other buildings to fulfill their demand. The reason that storage centers work this way is (again, developer verified) to avoid congestion in your drone / hyperloop networks.

However, as of patch #4 with the December 31st hotfix, storage centers sometimes generate far, far more demand than this.

This storage center has no minimum set (for any good), yet is requesting large quantities of water (all the lines pointing to the southwest are for water). In fact, it is requesting far more water than it has capacity to store, and actually already contains more silicon that it should be able to store.

Per patch notes, this bug is fixed as of the Advanced (beta) release as of 1/14/21. I’m leaving this here just in case it isn’t — if you observe this behavior, its a bug and should be reported as such.

Storage centers which have a minimum level set for a good will generate sufficient demand all at once to achieve the minimum level (an empty storage center whose minimum is set to 100 will generate 100 separate demand requests). This demand isn’t ultra-low priority demand — it is of high enough priority to pre-empt demand from non-prioritized factories. Thus, if you set a minimum in a storage center it will monopolize production of that good until its demand is satisfied. Once it hits the minimum, it will act like a normal storage center (demanding no more than a single good of each type) until it drops below the minimum, at which point it becomes “greedy” again.

Once goods enter a storage center they are very reluctant to let them go. Based on my experience, the following are true when there is no resource shortage:

1. If a storage factory has no minimum set, or the storage center currently has goods in excess of its minimum, goods will be released:
   1. If it is the closest source of a good required to allow a non-prioritized factory to start producing goods it will provide the goods. In addition, it will provide enough goods to satisfy the next production cycle — but no more. That is to say, an empty Steel factory can pull up to 4 iron and 2 carbon from a storage center. Demand past this, however, will never be met with goods pulled from the storage center.
   2. If it is the closest source of a good required to for a prioritized factory, it will provide goods as requested.
   3. If it is the closest source, it will provide food and water as required for colonies and research posts, even if the colony or research outpost is not prioritized. For example, an empty Research Outpost can pull up to 6 water and 6 food from a nearby storage center.
2. If a storage center has a minimum set, and it hasn’t yet achieved that minimum, the storage center will release goods as follows:
   1. It will never release goods to a non-prioritized factory.
   2. A prioritized factory will be provided goods as required, without regard to the minimum.
   3. It will provide enough goods to a non-prioritized colony or research outpost to satisfy the current production cycle, plus one more. This is the exact same behavior that occurs with storage centers interacting with non-prioritized factories above the minimum.
   4. A prioritized colony or research outpost will pull goods as required, without regard to the minimum.

I’m unsure of the behavior when there is a resource shortage — in that situation, I’m pretty certain that storage centers will release goods sufficient to satisfy demand regardless of whether or not a minimum is set — but only after all goods that currently exist on the map have been tagged for delivery to the destination. This certainly seems to be the behavior of space ports, for example, when you schedule start 11 simultaneous space elevator launches — carbon first comes from the mines or is stolen from goods that are currently in transit, and only after all the carbon on the map has been tagged for delivery to one of the space centers does it start pulling from nearby storage.

The above isn’t a bug (verified in discussions with the developer) — it is working as designed. The developer conceded that this design… Perhaps less than optimal for the user experience, and will be re-examining how storage centers work.

Given the above, my recommendation is to avoid storage centers unless you are willing to prioritize all of your factories and leave them prioritized indefinitely. This comes at a pretty dramatic cost of flexibility, but that’s the only way to get goods out of storage centers reliably at the current time.

Once you gain the Hyperloop technology, build them, and connect them together most of your logistics problems disappear — unless you are trying to do it, the its hard to break your economy once you have Hyperloops, so I’m not going to spend a good deal of time on this section.

That being said, there is one thing new players should know about Hyperloops — they “leak”! 🙂

Ok, they don’t really leak, but each Hyperloop conduit has a maximum transport capacity per Sol (in-game day) and when this capacity is exceeded, goods will start queuing up for transport in the source Hyperloop terminal. This, and of itself, isn’t a problem.

The issue is that the logistics AI is clever, and it runs a check whenever it considers using a Hyperloop terminal to see if the conduit that it intends to use will be overloaded when it gets there. If the check determines that the conduit will be overloaded, and no alternate conduit is available that moves it closer to its destination the good won’t even travel to the Hyperloop terminal — instead, it will travel along the roads (via the drone hub network) to its ultimate destination. This comes in three forms:

1. Leaky source Hyperloop: The Hyperloop where has an overloaded conduit and some goods never enter the Hyperloop. This is most common with storage centers, which can dump vast amounts of goods into the logistic network in a very short period of time.
2. Leaky destination Hyperloop: The obvious Hyperloop for the goods to travel to (the one closest to the destination) is only accessible via an overloaded conduit, so goods exit the Hyperloop “near by” and travel by drone hub for the last increment of their journey. This is most common near Spaceports / Space Elevators.
3. Leaky intermediate Hyperloops: Good enter the Hyperloop network, travel for a while, but then “hop off” (potentially in the middle of nowhere), travel along roads for awhile, then hop back into the Hyperloop network.

The actual behavior isn’t a problem — a developer has verified that playtesting proves that leaky Hyperloops provides minimizes travel time for goods given the resources that the player has provided.

But….

Since the goods never actually arrive at the Hyperloop, it appears to the player the Hyperloop conduit never actually gets congested. From the point-of-view of the player, it looks like a bug — goods are avoiding the Hyperloop system, despite the fact that the Hyperloop system has spare capacity. Furthermore, without the goods actually sitting in the Hyperloop terminal queuing for transport, it is rather hard to determine if you’ve actually fixed the problem.

The solution is… Delete drone hubs

Shocking, I know, but in this situation the best solution is to figure out where the goods are trying to get to, and delete all the drone hubs that connect the two areas. Once there is no longer a drone hub / road route to their destination, they won’t leak any more, and you’ll be able to see the overloaded Hyperloop in all its glory.

To actually fix the problem the solution is… Build more Hyperloop conduits and / or Hyperloop terminals!

Specifically, more goods / Sol can travel over a shorter conduit, so adding Hyperloop terminals in between two terminals can improve throughput somewhat. Also, while goods prefer to travel on the path that takes them closest to their final destination, they will consider any route that at least moves them closer to their final destination, so simply interconnecting the Hyperloop terminals you have may significantly improve the throughput of your Hyperloop network.

Ultimately, you’ll likely want to have a “Northern Hyperloop”, “Central Hyperloop” and “Southern Hyperloop” that are all interconnected with one another to ensure that goods always have a variety of different conduits to travel down.

Unmarked spoilers after this point — do not proceed if you don’t want to be spoiled!

The playthrough was started with Patch #2 and finished with Patch #3. I’m, for obvious reasons, not going to redo the entire walkthrough each time the game is patched, so… Your mileage may vary.

The walkthrough that follows is not the only way to defeat the first enemy base — it isn’t even the best by my own standards. The strategy I employed is designed to show that the “build lots of drone hubs” strategy is effective, and along the way provide a guide to overcoming one (of two) major challenges in the game.

Some obvious optimizations to the strategy shown include:

1. Build slightly fewer drone hubs, especially in the long spurs leading off to resources. Really, you could get by with about half as many. That’s still far more than most players build, in my experience.
2. Perhaps build some storage? I deliberately didn’t build storage (didn’t even research it) because storage makes logistics problems harder to troubleshoot. The rules for when storage generates demand (it seems to be distance based) and how the AI decides which storage will get an item is unclear to me, so I just skipped it altogether. It would be very, very helpful to have some storage with electronics in it before the first sneak attack, though.
3. Using nuclear weapons to free up the CO2 in the ice caps is, by far, the slowest way to finish the first terraforming stage. Anything else is faster, if that’s a priority.

I also didn’t deploy a second landing site to explore the nuclear silo (in SA-9). I’m dubious as to whether deploying a landing site like this is a net time savings on any level — to actually explore the nuclear silo, you’ll need to setup a fully functioning economy using the resources patches that spawn in when you drop the lander, then build the research post. By the time that I unlocked SA-9 (the earliest you could deploy the launch site) my spur was fairly close to the site anyway, so… Beyond that, you’ll eventually want to connect the two bases (to share resources and power), so you’ll have to build the spur anyway.

Is it worthwhile? My feeling is no, and I don’t like the free resource patches that spawn with a new landing site, so I don’t use them. Your mileage may vary, and I certainly understand their value when the Random Number Gods decide they hate you with random resource spawns.

With all that said, on to the walkthrough.

The starting position, with the tutorial not yet completed.

I built a couple of Areological Scanners (henceforth simply “scanners”), exploited the resources that they discovered, and waited patiently for colonists to arrive to increase my building limit:

While waiting for the colonists to arrive it is time to decide where to expand to next:

Given that this isn’t my first game, I recognized that:

1. First, I’d want to expand to the southwest, because there is a mission objective there and
2. Second, I hadn’t located any other deposits of silicon, which is required for to produce “Electrical Parts”, which is critical to the production of maintenance drones (and many other things) which is a critical need in the future.

Made the choice easy — I needed to expand toward the silicon patch located to the southwest.

At the time of this screenshot, I had already researched the first two “basic” engineering techs, simply to increase my building limit, and was working on researching the first “basic” space tech, to allow me to build two new spaceports (I hate turning off the “Import Colonists” mission, so I wasn’t interested in launching the SA-9 satellite until I had a second spaceport built).

Two years later, and new construction was getting close to SA-9,

Boy, that’s a lot of drones, isn’t it?

In year 10, I unlocked SA-9, built the the scanner to search the silicon patch, and planned further construction into SA-9 to both reach the chemical patch and get access to the POI for the nuclear weapons.

A close up of my base — note that the only additions to my base are some mines, two spaceports, and one basic colony. In particular, I haven’t built any additional factories at this time. Why? Because it is much easier to keep a single factory running at 100% efficiency, and as long as one factory is enough… Why build two?

If you were wondering: Researching the abandoned silo unlocks the ability to find and mine uranium. It does not guarantee that you will find uranium to mine, and without uranium you can’t use the microsuns. Thankfully, the Random Number Gods were merciful, and a Uranium patch is within reach.

A year later, and the research outpost has been created (nuclear weapons!) and spurs have been built off of my major highway to reach aluminum and uranium.

In year 13 my food factories started exhibiting production problems (production efficiency less than 90%), so I investigated the problem.

An output stall. Note that the silvery line that leads off to the upper right is critical to troubleshooting this problem — the line indicates that at least one of the goods at this location has been assigned to be shipped to another node (the other node can be found at the end of the silvery line — in this case, it leads to a food factory). If the line wasn’t present, then the issue would be lack of demand, in which case I’d need to look elsewhere for the problem.

In this case, however, demand was present, so the problem was logistical. A quick switch to the “Worker” lens revealed the problem.

Drone hubs alone aren’t enough — you actually have to build drones for them to help. As it is, a single drone was trying to service an enormous territory (including two chemical mines and an aluminium mine) and failing. Note that the territory color is not red — which is why the “worker efficiency” rating is so useless.

Many players in this situation would build a second food factory. This is a terrible idea, for two reasons:

1. With two factories, chemicals and water will be divided evenly between them. This creates the possibility that one of the factories has water but no chemicals while the other has chemicals and no water. In that situation, nothing is produced, and efficiency goes down.
2. Secondly, it doesn’t matter how much demand (by building more factories) you generate — if the resources are bottlenecked by logistics production isn’t going to occur.

In this specific scenario, the correct solution does turn out to be “build a new factory” — a worker factory. I would have preferred to wait until the Factories II research had completed (again, one factory is much easier to manage than two), but… Needs must.

I also noted at this point that I needed to up my steel production, and started a new spur towards the iron patch. Helpfully, the iron patch also had some nearby carbon, which is also required for steel manufacture.

Year 14 brought the start of uranium mining (finally) and exploitation of the aluminium patch. I also finally finished researching “Factories 2” and upgraded most of my factories to level 2 (one at a time — a factory being upgraded can’t produce anything).

Research moved to Engineering 3 — I want my hyperloops. 🙂

At this point I finally used the microsuns and nuked Mars. Got the achievement too. 🙂

Year 15 started a new terraforming stage, so I figured it was a good time to take inventory — and, in particular, to prove that yes, it is possible to have 90% efficiency without using hyperloops.

Note that food production is demand limited (I only have one Colony and a couple of Research Posts, and all of them were full of food, and the storage in the landing site was full as well — so the food factories were sitting with full outputs, which isn’t a problem).

The steel production stands out, obviously — thus the work on the northwest spur, to further exploit the iron field that resides in that direction.

Oh, and I decided I needed some more research outposts (for both population and the R&D bonus) — if my food factories are stalling due to lack of demand, then I might as well expand my population, right?

And I aborted my research of Hyperloops, which was taking forever, and moved to researching “Import greenhouse gases” — which I’ll use to complete the next phase of terraforming.

The additional iron and carbon signficantly improved my production of steel.

Year 16 triggered the deadly sneak attack. As always, damage was widespread, but the maintenance drones were on the job — and dropping like flies, of course.

The solution for this particular crisis is straightforward:

1. Make sure you have a large quantity of “Electronics” and “Plastics” on hand. Generally speaking, you’ll have a full set of plastics by the time you get to this point — other than maintenance drones, nothing else requires it, so it tends to accumulate in large quantities in maintenance towers naturally. Electronics, on the other hand, tends to be a problem. To address this:
2. Suspend (turn off) any ongoing construction and spaceports with reoccurring missions (in my case, two spaceports running the “Import Greenhouse Gases” repeatable mission). Most construction requires electronics to complete, and you want to make sure the only thing consuming electronics is the maintenance towers.
3. Put priority on all of your electronics factories. Remove priority from anything that consumes electronics — specifically, the worker factory. Again, you want to make sure the maximum number of electronics are available to the maintenance towers.
4. Wait

The combination of slower production and higher environmental (“tick”) damage this far more tedious to address on hard. The above suggestions work, but you really need to make sure you aren’t doing anything other than making electronics and wait. You may need to prioritize your maintenance towers.

At this point the number of maintenance drones had started to rebound. Still not time to celebrate, but once the number of maintenance drones stops going down, you are likely out of the woods.

Once the military research tree is unlocked, you should stop whatever you are researching and switch to researching drone hives. Once you have drone hive researched, build a row of them across the northern most extent of your base. In my case this was the northern spur — not desirable, but that’s where the iron was. Yes, this costs electronics: hopefully, your maintenance drone situation has improved from “dire” (25-45%) to “worriesome” (45%-65%). If you are still in “panic” state (<25% of your maximum capacity) and you’ve gotten to this point you are, more than likely, doomed.

Once you’ve researched it, build the assault drone factory. Yet another thing to consume electronics — make sure you are still making progress towards restoring your maintenance drones, and don’t start anything new.

Most of a year later, sitting at 192/210 maintenance drones I cautiously restarted construction — in this case, new construction, because now that my plastic factory was actually producing plastic (to replace the plastic that the maintenance drones consumed) my food production was down due to lack of supply. So expanding and building some new chemical mines seemed like a good idea.

Extending the picket line of drone hives, production of drones is underway. Worker drone levels are dropping pretty significantly here, so I restored the priority to one of the Worker factories and removed it from the assault drone factory, knowing that I had enough assault drones to defeat the first attack.

Finally I’m satisfied with my maintenance drone population (200 /210) and remove priority from the electronics factories and restore it to both of the worker drone factories.

Restarting the spaceports (importing greenhouse gases, still).

The first attack (well, the first one with drones). Note that you need to manually vector your drones (left click on the number above the drones to select them, then left click on the map to specify a direction) as the drones won’t compute efficient intercepts on their own.

Also, note that drone hives themselves function as defensive towers — when powered, they attack enemy drones within the shaded area. They are very powerful attackers, but proper positioning is difficult.

Settling in for a long tail chase. If you can’t get ahead of the enemy drones, you might as well as target the enemies directly (left click on the number above your drone, then left click on the number above the enemy drone). This will set your drones to follow the enemy wherever they go.

All enemies dead (with some damage to my base), returning to base.

Starting the process to unlock SA-1

Found a bottleneck when I clicked on this node by accident.

Fixed the bottleneck by adding a third strip of worker drones along that segment. I also started building additional spaceports (up to my limit) and finally bit the bullet and built a second parts factory.

The second attack. Sorry about the poor screenshot clarity — there was a dust storm in progress and I paused the game, so the picture is pretty noisy. The attack was easily repelled.

The enemy base is revealed (once the unlock for SA-1 completes).

The enemy base is destroyed — victory is mine! And lots of resources to exploit is good too.

Unless there is enormous amounts of demand, I’m not going to continue the walkthrough after this. I may post a separate guide for running a hyperloop based economy, but being efficient isn’t critical after you destroy the first base.

The only real remaining threat is oxygen levels, and addressing that is definitely beyond the scope of this guide.

No.

All screenshots are from Per Aspera (hard) difficulty, in story mode. While the game technically isn’t over yet, I’ve deployed all the components required for my ending and I’m just waiting for the CO2 level to drop low enough to win.

The starting area — the vast majority of the production is here (everything except for steel).
The spaceport complex. Based on my experience with in this playthrough, I think it will likely work better if the storage centers are separate (but nearby) the space center complex. With this layout, there aren’t enough worker drones to keep goods flowing into the storage centers or the space centers, when they are active.
When I expanded steel production, I moved it over here, because that’s where the iron was. Note, however, the original steel factories are still present and still active!
The space elevators.

Population over time
Worker drone population