Stellaris Dev Diary #380 - Defenders in the Stars

[Critical Ethics]

To reduce my rambling to a minimum — this is basically the only complete solution to the problem.

• Adding system caps (soft or not) with penalties for overstacking would require players to spread fleets across more locations.
• Stacking anyway should be a costly solution, de-emphasized by exposing other parts of the war front.
• Dispersion of force — like a realistic, believable star empire would do — should be the go-to.
• A complete, solid deployment and automation system for fleets would be a must-have to avoid excessive micromanagement.

Following this military rework, a redesign of the gameplay loop and economy is necessary, as you will eventually reach a point where additional fleets are no longer needed.
So additional game goals — beyond growth and endless militarization — must be introduced.
Different power mechanics like politics, utopian society systems, espionage, technology, and other avenues must be designed.

I’m just repeating myself — and what’s written in my signature here on the forum — but an adaptation of the Hearts of Iron IV land army and combat system for Stellaris would be my ideal solution.

Deep Space Citadels — and defenses in general — should have an extensive role in providing border protection, and the game should be balanced around that.

You're not repeating yourself. There's a pretty big difference between saying that static defenses having caps means fleets should have caps vs saying that the entire game needs to be scrapped and started from scratch except for keeping some of the names the same. The latter is new information, and is not so much a magic bullet as it is a magic supernova.

I'm going to back up a bit to what seems to be our core disagreement. I want people to not send all their ships to attack one thing. So do you. Therefore I want "send all my ships to attack one smaller thing" to be less effective (e: and more expensive, alongside other mitigations already in the game) . So do you. I consider penalising out-of-combat fleet intersections to be undesirable in its own right, for reasons I've already articulated, and unnecessary. You disagree.

If you consider penalising out-of-combat fleet intersections to be desirable or necessary, what, exactly, is the balance or gameplay issue with fleets docking at the same berth in peacetime, or travelling together in peacetime, or docking together or temporarily interacting en-route to different targets during wartime? Why does this need to be prevented out of combat?

If you consider it simply unimportant, and my concerns unfounded, what is the con of limiting any new mechanical penalties to in-combat that outweighs the cons of having to rebuild everything else about the game up to and including the entire map system?

 

[Imp0815]

You're not repeating yourself. There's a pretty big difference between saying that static defenses having caps means fleets should have caps vs saying that the entire game needs to be scrapped and started from scratch except for keeping some of the names the same. The latter is new information, and is not so much a magic bullet as it is a magic supernova.

I'm going to back up a bit to what seems to be our core disagreement. I want people to not send all their ships to attack one thing. So do you. Therefore I want "send all my ships to attack one smaller thing" to be less effective (e: and more expensive, alongside other mitigations already in the game) . So do you. I consider penalising out-of-combat fleet intersections to be undesirable in its own right, for reasons I've already articulated, and unnecessary. You disagree.

If you consider penalising out-of-combat fleet intersections to be desirable or necessary, what, exactly, is the balance or gameplay issue with fleets docking at the same berth in peacetime, or travelling together in peacetime, or docking together or temporarily interacting en-route to different targets during wartime? Why does this need to be prevented out of combat?

If you consider it simply unimportant, and my concerns unfounded, what is the con of limiting any new mechanical penalties to in-combat that outweighs the cons of having to rebuild everything else about the game up to and including the entire map system?

I’m not calling for a full reset of the game — just a focused rework of the military system, which I think is the root cause of the problem. That’s not a supernova, it’s a targeted overhaul — like others we've seen before.

Yes, we agree on the goal: doomstacking should be gone. I just believe that relying on isolated tweaks within the current system (like only penalizing in-combat) won’t achieve that. The incentives baked into the game still lead to stacking.

Out-of-combat penalties aren't the solution, just one example of how force concentration could be discouraged. Whether fleets dock in peace or pass each other in wartime isn't the real issue — it's that none of it meaningfully impacts strategy right now.

If deeper changes aren’t on the table, then sure — in-combat only tweaks are the least disruptive. But I think they won’t shift the gameplay loop enough to matter. That’s the tradeoff I’m pointing at.

 

[Critical Ethics]

I’m not calling for a full reset of the game — just a focused rework of the military system, which I think is the root cause of the problem. That’s not a supernova, it’s a targeted overhaul — like others we've seen before.

Yes, we agree on the goal: doomstacking should be gone. I just believe that relying on isolated tweaks within the current system (like only penalizing in-combat) won’t achieve that. The incentives baked into the game still lead to stacking.

Out-of-combat penalties aren't the solution, just one example of how force concentration could be discouraged. Whether fleets dock in peace or pass each other in wartime isn't the real issue — it's that none of it meaningfully impacts strategy right now.

If deeper changes aren’t on the table, then sure — in-combat only tweaks are the least disruptive. But I think they won’t shift the gameplay loop enough to matter. That’s the tradeoff I’m pointing at.

I think what you're missing is that the changes you're describing would require significantly altering the way the galaxy map itself functions. It's a sparsely connected node network and a bunch of current game concepts are built off that setup. If you own fleet A and fleet B, and your decisions on where to berth fleet A or what route you send it to location X impact what routes are available for fleet B to get to location Y - that would be so irritating to play in the node system Stellaris uses. Never mind trying to get the AI to not end up accidentally blockading its own fleets into its own territory.

You could maybe get it to work by going full 3D space map and switching from hyperlanes to warp so friendly fleet movements have minimal impact on the routing of other friendly fleet movements, and I think you'd find a lot of people who would be in favour of this for many reasons, but that's going to impact combat in ways far beyond preventing doomstacks, as well as requiring a redesign of most other aspects of the game.

 

[daniel060]

Am I understanding correctly that there's only one asset for the DSC? It seems like there should be one for each ship type. (Yes, I know that's a lot more work. Sorry!)

 

[Anfros]

Someone's been playing Terra Invicta

 

[Imp0815]

I think what you're missing is that the changes you're describing would require significantly altering the way the galaxy map itself functions. It's a sparsely connected node network and a bunch of current game concepts are built off that setup. If you own fleet A and fleet B, and your decisions on where to berth fleet A or what route you send it to location X impact what routes are available for fleet B to get to location Y - that would be so irritating to play in the node system Stellaris uses. Never mind trying to get the AI to not end up accidentally blockading its own fleets into its own territory.

You could maybe get it to work by going full 3D space map and switching from hyperlanes to warp so friendly fleet movements have minimal impact on the routing of other friendly fleet movements, and I think you'd find a lot of people who would be in favour of this for many reasons, but that's going to impact combat in ways far beyond preventing doomstacks, as well as requiring a redesign of most other aspects of the game.

You can already get a glimpse of such a rework: just ramp up the hyperlane density in the generation settings, and the "sparsely connected node network" becomes a fully connected node network.

Again, whether fleets in transition count toward system limits is just a detail.

 

[Ajacoba]

You can already get a glimpse of such a rework: just ramp up the hyperlane density in the generation settings, and the "sparsely connected node network" becomes a fully connected node network.

Again, whether fleets in transition count toward system limits is just a detail.

The game is absolutely not designed for that though

Starbase capacity is way too limited to defend such wide borders and citadels are too expensive in construction cost and upkeep to put them in several dozen systems

It would however again encourage players to flank like crazy and abandon doomstacks

 

[grommile]

citadels are too expensive in construction cost and upkeep to put them in several dozen systems

So you put them in high-value systems, instead.

 

[Ajacoba]

So you put them in high-value systems, instead.

I mean, all systems are high value
If the enemy captures enough "empty" ones your energy and mineral production may plummet, and with them your alloys and consumer goods will suffer

 

[Critical Ethics]

You can already get a glimpse of such a rework: just ramp up the hyperlane density in the generation settings, and the "sparsely connected node network" becomes a fully connected node network.

No it doesn't, a fully connected node network is where all nodes connect to all other nodes, not just all local nodes.

Even if we go with a locally fully connected node network you get things like:

  A---1---2
 / \ / \ / \
B---4---5---Y
 \ / \ / \ /
  7---8---X

A and B are fleets. X and Y are targets. If I send A to X and B to Y then they will intersect. How do you robustly tell an AI that A-Y B-X is the "superior" attack plan?

Let's say A-X B-Y is a better fleet match up than A-Y B-X. How do you robustly develop a pathfinding system that gets them there efficiently without intersecting? Or am I supposed to micro them both individually?

Let's assume it's only a soft cap:

  A---1---2
 / \ / \ / \
3---4---5---Z
 \ / \ / \ /
  B---8---9

I want A and B to converge and doomstack Y. I go A12Z and B45Z. They both arrive and murder everything. If something targets one of the split stacks the other can trivially jump in to assist. The out of combat doomstack penalty does not apply mechanically, it's just annoying to work around. So why bother?

Again, whether fleets in transition count toward system limits is just a detail.

It's not just a detail, it fundamentally changes the entire concept.

Can you not intercept or redirect traveling fleets (not reinforcements)? If you can't, you've removed a fundamental game conceit. If you can:

  A---1---2
 / \ / \ / \
B---4---5---Y
 \ / \ / \ /
  7---8---X

A45X + B45Y, or A45Y + B45X. An enemy fleet jumps from 2 to 5. If it's a hard cap what happens to your "extra" ships? If it's a soft cap then you're just describing "doomstack penalties only apply during combat" but with more steps and also limiting where you can station your ships during peacetime.

 

[Ajacoba]

Honestly, just keep the doomstack system around and teach players the advantages of flanking, ambushing and other silly tactics that encourage them to use catapults, jump drives or, at the very least, more than one ship design (so that their artillery X slot ships don't get rofl-stomped by torpedo defense platforms directly at the hyperlane)

 

[Critical Ethics]

Honestly, just keep the doomstack system around and teach players the advantages of flanking, ambushing and other silly tactics that encourage them to use catapults, jump drives or, at the very least, more than one ship design (so that their artillery X slot ships don't get rofl-stomped by torpedo defense platforms directly at the hyperlane)

That's one of the reasons I'm hopeful for doomstack combat trade penalties (e: as one component, not a magic bullet cure all). Seeing your trade plummet every time you over commit to a fight is very visible reminder to maybe try out something trickier and the higher ship attrition from not doomstacking will feel less harsh when you're sitting on a pile of filthy lucre.

 

[Imp0815]

No it doesn't, a fully connected node network is where all nodes connect to all other nodes, not just all local nodes.

Even if we go with a locally fully connected node network you get things like:

  A---1---2
 / \ / \ / \
B---4---5---Y
 \ / \ / \ /
  7---8---X

A and B are fleets. X and Y are targets. If I send A to X and B to Y then they will intersect. How do you robustly tell an AI that A-Y B-X is the "superior" attack plan?

Let's say A-X B-Y is a better fleet match up than A-Y B-X. How do you robustly develop a pathfinding system that gets them there efficiently without intersecting? Or am I supposed to micro them both individually?

Let's assume it's only a soft cap:

  A---1---2
 / \ / \ / \
3---4---5---Z
 \ / \ / \ /
  B---8---9

I want A and B to converge and doomstack Y. I go A12Z and B45Z. They both arrive and murder everything. If something targets one of the split stacks the other can trivially jump in to assist. The out of combat doomstack penalty does not apply mechanically, it's just annoying to work around. So why bother?

It's not just a detail, it fundamentally changes the entire concept.

Can you not intercept or redirect traveling fleets (not reinforcements)? If you can't, you've removed a fundamental game conceit. If you can:

  A---1---2
 / \ / \ / \
B---4---5---Y
 \ / \ / \ /
  7---8---X

A45X + B45Y, or A45Y + B45X. An enemy fleet jumps from 2 to 5. If it's a hard cap what happens to your "extra" ships? If it's a soft cap then you're just describing "doomstack penalties only apply during combat" but with more steps and also limiting where you can station your ships during peacetime.

I already stated that I don’t see the point of the discussion about what happens in transition or not. That’s a detail that needs to be tested and balanced by game designers.

If you’re still curious how this could work, just take a look at Hearts of Iron IV — it’s basically the same system, and units still suffer penalties if they exceed local supply even in transtions or just idling around.

It’s just a matter of numbers and tuning — and the AI handles the systems in HoI IV pretty well. At least well enough to provide a fun gameplay experience, where major mistakes aren’t the norm.

 

[Imp0815]

The game is absolutely not designed for that though

That’s why I always talk about a rework of the systems.

 

[Ajacoba]

The AI also perfectly counters doomstacks by flanking you and turning wars into a Micro-Management nightmare, lol

 

[Ishyro]

In my opinion there are 2 issues in the combat system that need to be solved before even looking at doomstacks.

1. Fleets can kinda be everywhere at the same time (at least be anywhere fast enough to defend and attack at the same time).
We could solve this simply making the combats longer, and sublight travel slower (for military vessels only).


2. Temporary losing control of systems doesn't cost anything, you just take it back and that's it.
So when you are invaded while you are invading, the correct answer 90% of the time is to ignore it, finish your own invasions, and only then take care of defending.

If a fleet in combat is "stuck" and can't do anything else for longer, and you can't just ignore the enemies roaming in your systems, you are forced to split fleets do defend multiple fronts.

That will not help when there is only a single front, but then doomstacking should be the correct response to that kind of war (imo, maybe not everyone will agree).

Any solution that try to make doomstacking costly or whatever doesn't solve the issue, because then the answer is not to avoid doomstacking, but to improve the economy until doomstacking is affordable.

 

[Ajacoba]

In my opinion there are 2 issues in the combat system that need to be solved before even looking at doomstacks.

1. Fleets can kinda be everywhere at the same time (at least be anywhere fast enough to defend and attack at the same time).
We could solve this simply making the combats longer, and sublight travel slower (for military vessels only).


2. Temporary losing control of systems doesn't cost anything, you just take it back and that's it.
So when you are invaded while you are invading, the correct answer 90% of the time is to ignore it, finish your own invasions, and only then take care of defending.

If a fleet in combat is "stuck" and can't do anything else for longer, and you can't just ignore the enemies roaming in your systems, you are forced to split fleets do defend multiple fronts.

That will not help when there is only a single front, but then doomstacking should be the correct response to that kind of war (imo, maybe not everyone will agree).

Any solution that try to make doomstacking costly or whatever doesn't solve the issue, because then the answer is not to avoid doomstacking, but to improve the economy until doomstacking is affordable.

Any system or planet that is conquered is removed from your economy though (until the war ends or you reclaim it)

Also fleets absolutely can not be everywhere at once, it can take literal years to travel through your empire

That's why we have hyper relays and star gates to make it slightly faster - but even those tools have their limits

 

[Imp0815]

Any system or planet that is conquered is removed from your economy though (until the war ends or you reclaim it)

Also fleets absolutely can not be everywhere at once, it can take literal years to travel through your empire

That's why we have hyper relays and star gates to make it slightly faster - but even those tools have their limits

And on a side note, we already had this problem — and they "fixed" it by introducing hyper relays, and a bit further down the line, by making jump gateways buildable.
So the issue of doomstacking was indirectly acknowledged: players were encountering the problem of their stacked fleets being immobile, and the solution was to put them on a train — allowing them to be nearly everywhere at once.
So going back to this now is just reintroducing the same complaint without fixing the cause.

 

[Critical Ethics]

I already stated that I don’t see the point of the discussion about what happens in transition or not. That’s a detail that needs to be tested and balanced by game designers.

If you’re still curious how this could work, just take a look at Hearts of Iron IV — it’s basically the same system, and units still suffer penalties if they exceed local supply even in transtions or just idling around.

It’s just a matter of numbers and tuning — and the AI handles the systems in HoI IV pretty well. At least well enough to provide a fun gameplay experience, where major mistakes aren’t the norm.

You can't say that you "wouldn’t recommend splitting a supply system into states, but rather applying it universally — to make it easier for players, automation, and the AI" as an argument in one post and then say that whether or not the supply system is split into states is irrelevant in another.

You can't argue that what happens in transition is pointless when the point is that either answer is flawed.

You can't port in systems from completely different games with completely different maps and completely different gameplay loops and then call it just a combat rework - HoI 4 doesn't even have an exploration mechanic.

I mean, you absolutely can do all those things, I'm not your mum, but it's not how discussing stuff works. A discussion is two or more people actively engaging with each other's points. Yeah it's absolutely fine to dismiss stuff like individual civics or specific origins as "that's up to the devs to work out", but when people are discussing the actual core mechanics of your concept, the actual part you think will solve all the problems and things cannot be solved without, then for it to count as a discussion you have to actually read, parse, internalise, and engage with those arguments even if they bring you to conclusions you don't like (assuming the other person is actually engaging, there's no point in putting the effort into arguing in good faith with someone who's not (not a dig at you right now, just putting that important qualifier in there on general principles)).

So yeah, I've pointed out numerous, very detailed issues with applying persistent system ship limits to Stellaris without ripping out the game map and associated concepts and completely starting over with everything those chain down to. If you want to argue against them, fire away. If you want to argue it would still be worth doing to avoid the nightmare scenario of only applying combat limits in combat, sure. Lord knows I'm more favourable than most to large scale sweeping changes, but you're awful optimistic about player acceptance of such a thing, and vastly underestimating how much of the game it would impact. More importantly if we were starting from scratch anyway then there's potentially solutions other than hard, persistent system limits (SotS SotS SotS) and describing starting completely over like that as adding system limits to Stellaris is like describing a stew as adding boiled potatoes to a sandwich.

 

[Imp0815]

You can't say that you "wouldn’t recommend splitting a supply system into states, but rather applying it universally — to make it easier for players, automation, and the AI" as an argument in one post and then say that whether or not the supply system is split into states is irrelevant in another.

You can't argue that what happens in transition is pointless when the point is that either answer is flawed.

You can't port in systems from completely different games with completely different maps and completely different gameplay loops and then call it just a combat rework - HoI 4 doesn't even have an exploration mechanic.

I mean, you absolutely can do all those things, I'm not your mum, but it's not how discussing stuff works. A discussion is two or more people actively engaging with each other's points. Yeah it's absolutely fine to dismiss stuff like individual civics or specific origins as "that's up to the devs to work out", but when people are discussing the actual core mechanics of your concept, the actual part you think will solve all the problems and things cannot be solved without, then for it to count as a discussion you have to actually read, parse, internalise, and engage with those arguments even if they bring you to conclusions you don't like (assuming the other person is actually engaging, there's no point in putting the effort into arguing in good faith with someone who's not (not a dig at you right now, just putting that important qualifier in there on general principles)).

So yeah, I've pointed out numerous, very detailed issues with applying persistent system ship limits to Stellaris without ripping out the game map and associated concepts and completely starting over with everything those chain down to. If you want to argue against them, fire away. If you want to argue it would still be worth doing to avoid the nightmare scenario of only applying combat limits in combat, sure. Lord knows I'm more favourable than most to large scale sweeping changes, but you're awful optimistic about player acceptance of such a thing, and vastly underestimating how much of the game it would impact. More importantly if we were starting from scratch anyway then there's potentially solutions other than hard, persistent system limits (SotS SotS SotS) and describing starting completely over like that as adding system limits to Stellaris is like describing a stew as adding boiled potatoes to a sandwich.

You see, I haven’t quite found the right words for it yet, but it’s starting to look like we both have completely different perspectives on what it means to implement or rework systems — what’s needed and what’s feasible for Stellaris overall.

I understand the point that reworking combat and fleet progression would be a major overhaul, since the whole game truly revolves around it. There isn’t really a clear metric for progress beyond seeing increasing fleet power over the span of the game. So yes, if you touch doomstacks, you’re forced to rewrite big parts of the game — but that says more about the issue than about the feasibility.

So maybe there isn’t any argument left — because I think the game is really starting to overstay its welcome with flavor DLCs for over two years now, while the fundamentals remain the same boring loop. Either the game changes, or it burns out — and we truly need to set our hopes (and money) on a successor.

I get the point that the changes I see as necessary are big ones — but I believe they’re needed to reach a place where all these new flavors actually make sense, and aren’t just smoke and mirrors.