General Summation of Stellaris's Warfare Problems

[DrFranknfurter]

I think that's how the game should be designed, because the solution in that example is to protect the trade base itself.

I wish this were possible. Specifically to be able to protect a civilian trade Starbase using as many Military Starbases in the system as I think it needs. I want Shipyards/Anchorages/Trade Hubs/Solar Panel Networks to be more powerful but vulnerable (destructible) targets in war, I also want to be able to put my defences directly on top of my vulnerable spots instead of having to put them on distant chokepoints that are only valuable because the enemy has no option but to pass through them.

I don't actually mind Doomstacking specifically, one unstoppable fleet involved in an epic battle against impenetrable defences sounds better than being expected to juggle 10+ fleets in a series of tiny battles against 10+ identical starbases in relatively unimportant and defenceless systems. I think sometimes I may have argued against doomstacking when I think what I really want to argue against is dull, uninteresting, predictable and annoying warfare. Sometimes that is the same thing, often it is quite different and doomstacking is just one symptom of a bigger, more nebulous problem.

I do want clear and obvious targets in war. Destroy this big glowing thing to win the war instead of miserably trying to spot that one unoccupied system somewhere that's holding up victory, sometimes securing those last few acceptance points in a war is like getting blood from a stone.

So I want to be encouraged to beeline for the enemy Trade station. Destroying it to prevent the enemy from collecting resource income, or to destroy their anchorages to prevent repair and resupply, their shipyards to prevent fresh reinforcements. But I wouldn't want hitting the targets to be quick or easy. Nor do I really want it to require micromanagement of a dozen fleets and their respective logistical headache.

In my theoretical game of Stellaris (5.0 in 2025) the Trade Starbase could be deemed your most valuable target, invested with the highest fleet power of any system, protected with hangers filled with fighters, minefields, system defence boats, planetary missiles coming from every world and fortress habitat, many redundant repair platforms keeping ships in the fight and bringing them back after disengaging, multiple overlapping planetary shield generators protecting the superstructure of fortifications, multiple FTL inhibitors/shield disruptors/communication jammers to mess with enemy fleets and to top it all off a giant Star Fortress or four spewing out a stream of death from a battery of Plasma Cannons, Particle Lances and Neutron Launchers. In response the enemy has siege weapons (artillery and carriers well outside of M turret range and taking less damage), a colossus roaming around cracking enemy forcefields and blasting away annoying fortress habitats, a juggernaut at the back providing repair, support and a battery of counter-auras while a huge fleet is whittling down defences and enemy reinforcements are rapidly closing in.

I really like the idea of making a ridiculously powerful defensive system where you can anticipate the spectacle the fight will provide. In a game like Sins of a Solar Empire I'd have massive protracted battles where planets really feel under siege - ships sneaking around the fields of fire to bomb the planet, or to raid nearby systems, long-range siege weapons jumping in and bursting down the hardened targets while fleets rush to reposition. Massive Starbases, epic explosions and yet clear, surprisingly easy to read combat (ships staying at a respectful distance, facing one another as they blast each other with volleys of colour-coded lasers and missiles rather than merging into a big deathball like a shoal of hungry fish). It's beautiful when it's all working perfectly and sometimes I wish the best aspects could be brought to the rather forgettable space combat here. (Combat is all Sins has, so it has to be good. Stellaris has so much more variety to the experience it can afford to have weak, almost nonexistent combat).

So while I want high value targets, sadly the current Starbase system only works if you fortify the system EN ROUTE to the high value target. All those gun/missile/hanger modules are completely incompatible with those trade/anchorage/shipyards. The whole concept only works with chokepoints to let one starbase protect another. When the hyperlane connectivity is high enough that chokepoints stop existing it becomes impossible to protect any of your trade/anchorage/shipyard systems without using a wall of fortifications to protect a single system and you just don't have the starbase capacity for that. So you just use a big fleet instead. (That's assuming that it was even possible to delay a fleet with any one of those fortifications. Barring the early game Starbases quickly stop even being a minor impediment).

In Stellaris everything changes when you get a battleship fleet that can fire several times at extreme range on the way towards the starbase without even slowing down (they did at least slow down a little back when we had a combat speed setting). The difference between attacking an outpost and the maximum possible suite of system defences is that the outpost dies after the first volley, the Citadel and some platforms may last till the second or third volley (if it's lucky, maybe even kill a ship or two in retaliation if it's a fallen empire citadel as they're downright fancy). Amusingly a handful of civilian ships trying to evade combat can drag the fleet across the system, and that could actually slow the enemy down longer than all the alloys you could possibly spend on defences combined (but still only a few days).

There just isn't the epic feeling of this fight being the last stand for mankind when even the best very best possible starbase feels frighteningly flimsy, more a liability than an asset. It's going to die in a single volley only to be used to repair the enemy fleet a few days from now and then turn its guns back on your fleet given a few more days to repair the damage it took.

For me my issue isn't about the raw military power of Starbases but their extremely low durability and the frustrating inability to build them where they need to be. I want them to survive long enough for reinforcements to arrive... but that means several hundred days with current sluggish fleet speeds... and that would be a massive new problem in wars when you have 10+ maximally defended systems per empire and 3+ enemy empires that need crushing to achieve your war goals.

If starbases were as durable as I'd like while remaining as cheap, distributed and mass-produced as they currently are, then wars would take an extra 30 years. That's just not going to happen. So instead I want my starbases to actually be more concentrated, I want (paradoxically) more doomstacking rather than less. I want to trim things down to 3 key systems each with 4 starbases protecting their central, valuable and massive trade/anchorage/shipyard. I don't really want to have to manage a dozen completely identical starbases, each with the combined military strength and stopping power of a damp sponge.

So I don't really mind the concept of doomstacks, they do have some positives:
1. Less Micromanagement (fewer fleets/starbases to manage and babysit)
2. More epic battles (1/10th the fleets/stations means each battle has 10x the risk, drama and visual beauty... and means you can watch almost every battle)
3. More predictable engagements (1/10th the fleets/starbases means 1/10th the mental calculations and subsequent mistakes - easier to learn, less frustrating)
So doomstacks and low fleet counts aren't all bad to me... as long as the single decisive battle they produce is suitably epic that it makes it a fun climax to the gaming night. Sadly it just isn't.

Also as an aside the disengagement mechanic means that the fleet you just swatted in that epic battle may take 0% casualties (0 alloys lost) or 100% (20k-100k alloys lost). I don't like the level of RNG that adds. Manually withdrawing a fleet before a key ship is lost feels more tactical and strategic, but for that to happen you need to care about any one individual ship enough to end the battle early to save it (capital ships in Sins of a Solar Empire), and for there to be a manual element rather than simply rolling a dice to see how many ships you lost this time. Plus the combat report system doesn't like or understand the mechanic. It would much rather there were only ever 1 fleet on either side... any more than that and it gets painfully confused and spouts gibberish at you.

So in my ideal combat overhaul I'd have:
Civilian Starbases (Trade, Anchorages, Shipyards - can stack multiple of the same type in a single system, built with influence, can be destroyed)
Military Starbases (placed freely, can eventually stack multiple in a single system, but only with investment and ascension perks)
Platforms/Minefields (All of this managed from the starbase defence manager screen, so they can be rebuilt with a single reinforce all click)
Planetary defences (Shields that envelop Starbases, planetary Missile/Gun/Hangars)

Your Home system could then have a Trade Starbase that covers almost all your empire, or an Anchorage that provides almost your entire naval capacity, Shipyards that construct most of your fleet and store all your reserve mothballed ships before they are refitted for service. Your Home system is then protected by 4 Star Fortress, 20+ platforms, 3 shield generators on nearby habitats etc.

To crack the heart of the empire you need to get past the Military Starbases, to do that you first have to either:
1. Siege at long range to avoid most of the M slot weapons (artillery/carriers)
2. Attack with overwhelming firepower (everything you have)
3. Disable the shield generators first (bombard/crack planets)

This could be achieved without a massive amount of micro by adding a fleet stance:
1. Normal (Ships move to range set by combat computers)
2. Siege (Ships stay out of Starbase range, attacking if possible)
3. Bombard (Ships move to bombard planets and take down support, evading when planets are disabled)

You may only have 3 key targets instead of dozens, but you'd probably split forces more often if you could leave a small force laying siege to one target while your main fleet attacks head-on and another tries to disable the planet shield that is reducing all incoming damage the starbase is taking.

I'm not sure how clear my thoughts are, or how well reality would align with fantasy, but it is fun to imagine a more ideal combat system in Stellaris v5.0 away in 2025.

 

[methegrate]

How do fleets dominate 'war score', by which I assume you mean surrender acceptance? Relative naval strength only gives at most +50 surrender acceptance, which is not enough on its own to broker even a status-quo peace, let alone an outright victory. And taking planets is relevant if you intend to conquer them - it's basically impossible to conquer a planet you haven't occupied.

Is there another name for the score they use to measure surrender acceptance? Not being snarky. Literally, is there something else I should call this point score that measures the state of war?

My experience is that the fleets determine the end of the war. I can take almost every one of their planets, every single starbase, but if their fleet is still intact then the war is not over. By contrast, destroying their fleet tends to generate most of the war score/surrender acceptance and war exhaustion I need to end the war and, of course, largely knocks them out by destroying their military.

The upshot is that, in practice, the fleet becomes the only target that matters. Take two hypothetical extremes. If I ignore their fleet, but take every planet and starbase, the war is still not over. (As has happened to me innumerable times in my own games.) I don't have enough war score or exhaustion to end it. Meanwhile if I ignore their planets and starbases, but destroy their fleet, the war is over. I have most of the score/exhaustion I need, and I can easily take any remaining targets that I need.

The second prong of this makes sense. If you take out the entire enemy fleet, they can't keep fighting the war. But because there's no other way to end the war, the enemy's planets and systems are more or less irrelevant targets.

This feeds into the doomstacking problem. Your only defensive priority becomes your fleet, because losing the fleet is the only way you can lose the war. Meanwhile your only offensive priority is their fleet, because defeating their fleet is the only way you can win it. As a result force concentration is always the best move. It protects the only thing that matters (your fleet) while giving you the best way of attacking the only thing that matters (theirs).

 

[Iosue Yu]

To summarise

1. There isn't any interesting target other than the main fleet in a war.
   1. Despite there are planets to invade and trading posts to occupy, a war is still fought over a main showdown of doomstacks. Because you'd just have enough time to seize all undefended assets when the enemy's main forces are on retreat.
   2. And those assets you can seize aren't even that valuable.
   3. The defenses of these assets are just some Starbases which just takes 10 more days to shoot down. Starbases as defense aren't that deterring at all unless you happen to have some other doomstacks nearby.
   4. And players can repair at captured Starbases but the AI doesn't.
2. The wars are also fought in such perfectly even or controlled lab environment.
   1. There isn't even geography. It's just flat plains of various lengths to travel.
   2. Hazards are never hazardous.
   3. Everything is visible to both sides of sensors have been upgraded unless blocked by Nebula.
   4. No risks of an ambush because ambush is fought the same way you engage a regular fleet.
   5. No supply limiting factors.
   6. You basically operate anywhere in the galaxy, in some uncharted waters, the same way you operate at home or in familiar waters.
   7. No changes of Morales.
   8. War Exhaustion isn't War Exhaustion, as all Pops still provide output the same way.
   9. These all limits the rooms for strategies. Everything comes down to how big the number reading of Fleet Power is. It's improved by good technology and sheer number of ships. Ship experience doesn't even contribute that much.
3. Diplomatically unlimiting
   1. No Aggressive Expansion impacts
   2. No War Scores so you can take as much as you want even as non-genocidals
   3. No or negligible penalties for going over administrative capacity after signing a peace deal
   4. Rigid peace deals that you can't even "let them live" by vassalising them
   5. No dynamic war joining because you can't make new alliances when at war.
   6. Losing a war is often crippling to a point because your capital can just be claimed like any other system. But your Capital planet is much stronger than any other planet, while in most other Paradox titles, your capital is another city just with better benefits.
   7. Can't lead an exodus in times of imminent defeat. So half of your entire species will just wait to be purged.
   8. Conquered Pops only pose economical damages but the risks of deflecting is minimal. And it only lasts so long like 10 years or something. It's totally manageable without much efforts. (I mean the AI can manage it how hard is it really?)

 

[makaramus]

I have to insist on doomstacks incrasing your war exhaustion... it sounds like a good idea to me since its not forcing you to not do anything diffrent but shorten wars and give you less time to conquer stuff.
@Iosue Yu actually there is a aggresive expansion but with new relations system became obsolate. When you conquer a lot you will be seen as "threat" by your neighbors not because you grow strong but because you are expanding via war
its still there but compared to your +500 relations boost thanks to all new stuff its like nothing anymore

 

[Iosue Yu]

I have to insist on doomstacks incrasing your war exhaustion... it sounds like a good idea to me since its not forcing you to not do anything diffrent but shorten wars and give you less time to conquer stuff.
@Iosue Yu actually there is a aggresive expansion but with new relations system became obsolate. When you conquer a lot you will be seen as "threat" by your neighbors not because you grow strong but because you are expanding via war
its still there but compared to your +500 relations boost thanks to all new stuff its like nothing anymore

You know there's also this 2 Energy upkeep for Construction Ships.

Aggressive Expansion right now is similar to this. Too minor to make any significant difference.

 

[Altruist]

[Short:]
I tend to split humans (when it comes to gaming) into 3 sub-species (and I always wonder that we really belong to one species):

1. role-players / diplomats (content)
2. artists (visiually driven)
3. strat & tact gamers (often don't care much about the above 2)

Stellaris tries to be the game for all 3 sub-species and IMHO excells for the first 2 and lacks horribly in the 3rd section. To excell in the 3rd you need somebody in charge who really knows strat & tact games and protects the core-engine like a holy grail from the first 2 groups who always want to add another layer without taking into account wether the core engine can cope with it strat & tact wise. You also need to faithfully defend it from player wishes (of which, let's admit it, 90% are usually insane, game-breaking or overcomplicated) while maintaining an openeness for the other 10% suggestions.

For an immensely complex game like Stellaris within a RL capitalistic driven economy there are, of course, many other factors which make things difficult but, for a moment, sit back, have in mind what I tried to write about above and look at the changes and DLCs Stellaris had in the last years...

I bet there is no single person in the Stellaris-team left who dares to say she or he has still a profound overview of the weight and decision making process within Stellaris. Most likely it will be rather a kind of research project to look up what goes on... (in a bizarre way they have succeeded in programming something which has grown beyond them... wether you call this a failure, chaos or an achievement is IMHO an open question).


[For those who like to read longer explanations...]
Designing and especially maintaining a good strat & tact game is a very difficult task.
Even companies which really should know how to do it, fail all too often (see Stellaris, Civ...)

First, everybody says AI... well, that's SciFi but not the present state of software. The software is, in comparison to a mean human, dumb. It's much better than it used to be, though. Which is rather the problem than the solution. With a truly dumb software and very limited hardware, nobody would be as insane as trying what the Paradox programers tried with their games: no more completly fixed scenarios you have to beat, or fixed scenarios with a bit of randomizing thrown into it. Instead they tried to go with pointers and a weight-system everything is interconnected with and the software makes decisions based on those... and fails miserable. As many good gamers have suggested, the Stellaris engine would actually profit from reintroducing some fixed build- and behaviour patterns which is, in terms of programming, a step back.

The interesting question is, why this new approach of programming fails so miserably.

There are, of course, still quite severe hardware restrictions if you really want to make everything interdependent. Real AI-programming is still in its childhood or rather embryo-stage. Another thing, as long as you don't really know how humans make the decisions they do, it is very difficult to simulate it. And here, too, as much more as we know today, research and knowledge is still in a very early stage.

So they come up with that pointer and weight-system (which is quite a brilliant idea IMHO).
But to make a good strat & tact programming code, it needs, besides many other things, also good strat & tact players. And not just any good strat & tact players who do what they do rather intuitive but players with a true understanding why they are doing things (and those are rather rare). I'd consider myself a good strat & tact player but I (and I think most good players will agree) must admit, as much as I do the necessary steps like deeply diving into the in and outs of how the game works, for some of the most important decisions I rely on a rather intuitive understanding. Those are mainly the different stages every 4x strat & tact game has and the most important is the switch from peaceful expansion to military expansion. Every strat & tact gamer knows how difficult it is to teach even humans to "see" when it is the best time to switch because there are so many variables and it is slightly different in every game.

So far I have said strat & tact players, what I actually mean are programmers and those groups are often enough not identical. Programming is math and algorithms and that genius spark how to transfer economy, research, military and strat & tact into code. In Stellaris is added the additionaly dificulty of how and what weigth is added to possible decisions. With each new ability/DLC those weights probably need to change. To make things even more difficult, when tens of thousand players playing the game, they come up pretty fast with some devious tactics you have never even dreamed of as the original game designer. This, too, would need adjustment of the given weights and how the "AI" makes its decisions.

And probably the most difficult: For one or 3 games everything is fine but then a good player recognizes patterns in the decision making process and ruthlessly starts to exploit them (and although this is often enough a rather intuitive process, humans excell at recognizing patterns and other than software not only exact repeating patterns but also even similar behaving patterns). We start to "outthink" the "AI" which often enough is just us abusing software patterns. I haven't even a starting clue how this can be solved software-wise.

 

[Iosue Yu]

[Short:]
I tend to split humans (when it comes to gaming) into 3 sub-species (and I always wonder that we really belong to one species):

1. role-players / diplomats (content)
2. artists (visiually driven)
3. strat & tact gamers (often don't care much about the above 2)

Stellaris tries to be the game for all 3 sub-species and IMHO excells for the first 2 and lacks horribly in the 3rd section. To excell in the 3rd you need somebody in charge who really knows strat & tact games and protects the core-engine like a holy grail from the first 2 groups who always want to add another layer without taking into account wether the core engine can cope with it strat & tact wise. You also need to faithfully defend it from player wishes (of which, let's admit it, 90% are usually insane, game-breaking or overcomplicated) while maintaining an openeness for the other 10% suggestions.

For an immensely complex game like Stellaris within a RL capitalistic driven economy there are, of course, many other factors which make things difficult but, for a moment, sit back, have in mind what I tried to write about above and look at the changes and DLCs Stellaris had in the last years...

I bet there is no single person in the Stellaris-team left who dares to say she or he has still a profound overview of the weight and decision making process within Stellaris. Most likely it will be rather a kind of research project to look up what goes on... (in a bizarre way they have succeeded in programming something which has grown beyond them... wether you call this a failure, chaos or an achievement is IMHO an open question).


[For those who like to read longer explanations...]
Designing and especially maintaining a good strat & tact game is a very difficult task.
Even companies which really should know how to do it, fail all too often (see Stellaris, Civ...)

First, everybody says AI... well, that's SciFi but not the present state of software. The software is, in comparison to a mean human, dumb. It's much better than it used to be, though. Which is rather the problem than the solution. With a truly dumb software and very limited hardware, nobody would be as insane as trying what the Paradox programers tried with their games: no more completly fixed scenarios you have to beat, or fixed scenarios with a bit of randomizing thrown into it. Instead they tried to go with pointers and a weight-system everything is interconnected with and the software makes decisions based on those... and fails miserable. As many good gamers have suggested, the Stellaris engine would actually profit from reintroducing some fixed build- and behaviour patterns which is, in terms of programming, a step back.

The interesting question is, why this new approach of programming fails so miserably.

There are, of course, still quite severe hardware restrictions if you really want to make everything interdependent. Real AI-programming is still in its childhood or rather embryo-stage. Another thing, as long as you don't really know how humans make the decisions they do, it is very difficult to simulate it. And here, too, as much more as we know today, research and knowledge is still in a very early stage.

So they come up with that pointer and weight-system (which is quite a brilliant idea IMHO).
But to make a good strat & tact programming code, it needs, besides many other things, also good strat & tact players. And not just any good strat & tact players who do what they do rather intuitive but players with a true understanding why they are doing things (and those are rather rare). I'd consider myself a good strat & tact player but I (and I think most good players will agree) must admit, as much as I do the necessary steps like deeply diving into the in and outs of how the game works, for some of the most important decisions I rely on a rather intuitive understanding. Those are mainly the different stages every 4x strat & tact game has and the most important is the switch from peaceful expansion to military expansion. Every strat & tact gamer knows how difficult it is to teach even humans to "see" when it is the best time to switch because there are so many variables and it is slightly different in every game.

So far I have said strat & tact players, what I actually mean are programmers and those groups are often enough not identical. Programming is math and algorithms and that genius spark how to transfer economy, research, military and strat & tact into code. In Stellaris is added the additionaly dificulty of how and what weigth is added to possible decisions. With each new ability/DLC those weights probably need to change. To make things even more difficult, when tens of thousand players playing the game, they come up pretty fast with some devious tactics you have never even dreamed of as the original game designer. This, too, would need adjustment of the given weights and how the "AI" makes its decisions.

And probably the most difficult: For one or 3 games everything is fine but then a good player recognizes patterns in the decision making process and ruthlessly starts to exploit them (and although this is often enough a rather intuitive process, humans excell at recognizing patterns and other than software not only exact repeating patterns but also even similar behaving patterns). We start to "outthink" the "AI" which often enough is just us abusing software patterns. I haven't even a starting clue how this can be solved software-wise.

I've at times discussed what AI is and how the CPUs of Stellaris aren't AI. But it's more academic than practical. And to be honest, the CPUs have half the characteristics of AI, not sentient AI but still AI.

What we know about AI so far is that

1. It is made of a Knowledge Database;
2. Its decisions are made by giving a complex search in the Database;
3. Database is generated by Machine Learning;
4. Database updates live as soon as new knowledge has been gained.

What Stellaris CPUs are doing is like a simplified version of 1 and 2. Only that the basic Database is made of human input instead of running the first learning algorithms. And those weights are arbitrary.

If we want CPUs to approach closer to an AI-status, we can still do it with realistic hardware by doing offline updates that the Paradox studio runs at home. Similar to anti-Virus detection, Paradox can update us AI thought patterns.

Weights should be generated by statistics. It means Paradox can collect a lot of "Play Cases" and receive input from actual human players by some means. And then these Play Cases get turned into conditions and weights after normalising into a valid number.

The technology is tedious. But logically it is better than assigning arbitrary numbers to decision making. Guess work means you don't actually know which point you should aim at and when you'd arrive. But using statistics, despite technologically challenging, the solution is finite.

 

[MichaelJanuary]

Combat in my imaginary Stellaris II

FUEL DEPOTS
Gas Giants are a source of hydrogen/helium isotopes used for interstellar travel. Ships have to refuel every so often from starbases with fuel depots that are kept stocked up by gas-giant cloud scoops. A cloud scoop also has is its own fuel depot. Fuel is also a major component in the manufacture of ships, missiles and fighters. The cloud scoop and its attendant fuel depot can be disabled due to combat damage.

FTL
All ships travel at warp. Warp is slow. For example, lets say warp is 1 light year per month. Ships in warp cannot change orders. Beyond a certain range (say 6 light years), there is a penalty to warp speed due to the energy requirements. So the further you jump, the slower your fleet moves. A 10 light year jump would be slightly slower than two 5-light-year jumps, but a 15 light year jump would take significantly longer than three 5 light year jumps, and take significantly more fuel.

SUB-SPACE SENSORS
Ships in warp can be tracked by sub-space sensors if you have the tech. This would reveal their speed, vector and likely destination, IF you have sensors in range. If your sensors/tech is good enough, it may reveal ship numbers, types, etc. Otherwise you may have an incomplete picture of what the detected fleet looks like.

INTERCEPTS / FRONT LINES
If you detect a fleet in warp heading towards one of your systems, you can launch an intercept fleet to the same system (if you have a fleet in the area). Depending on speed, range, vector, you may even arrive in time to catch them in that system. Its possible for fleets to jump past your front line, but due to the energy/time penalties on very long jumps, such fleets will be progressively easier to intercept the further they are trying to jump.

RETREAT
Fleets that retreat from combat will retreat at warp to the last friendly system they were in, or the nearest friendly or neutral system if that system has been captured. Retreating fleets can be tracked, and even intercepted, if you can get a fleet to their destination in time.

METALS
Rocky planets, asteroids and habitable planets are all sources of metals. Metals can be refined into alloys by refineries. Alloys are the primary component used by shipyards to manufacture ships, fighters, missiles and ammunition (ref EVE, SOTS2, Polaris Sector).

MANUFACTURING
All manufacturing of ships, missiles, fighters, ammunition is done by ORBITAL SHIPYARDS. Components in the manufacture include ALLOYS, HE3, EXOTIC ELEMENTS (strategic minerals -motes, gas, crystals).

ORBITAL SHIPYARDS
You can build one orbital shipyard for every 10 pops on a planet. Orbital shipyards can be disabled by combat damage during raids on the star system.

LIMITED AMMUNITION
Ships can hold limited quantities of fighters, missiles and kinetics, and could run out in extended engagements, leaving them entirely dependent on energy weapons. To resupply, ships would need to return to a friendly starbase,

WEAPON EFFICIENCY / FLEET COMBAT
Missiles should be fast and devastating, while point defense is limited. It should be relatively easy to overwhelm any ship/fleets point defenses, as long as you have missiles/fighters. Standard tactics would be for fleets to engage in long range missile duels until out of missiles, then close for fighter attacks until fighters run out, then close to energy or kinetic range (assuming either side is still in the battle). Even point defenses would run out of ammo in extended missile/fighter duels, leaving ships at the mercy of incoming missiles/fighters.

ASSAULTING STARBASES
Any fleet that assaults a starbase, would run out of long range missiles and fighters and point defense ammunition way before the starbase does, and so it would ALWAYS be a losing proposition.

RAIDS / SIEGES
To take a system with a sufficiently strong starbase, would require:

• several 'raids'
  • Target gas-giant fuel extractors, disabling them.
  • Target orbital infrastructure of inhabited planets
• This would degrade the starbase supply of missiles/fighters, while also destroying the in-system manufacturing base for fighters/missiles.
• Once sufficiently degraded of missiles, fighters, the starbase itself would be vulnerable to fleet attacks, as it cannot move (and point defense is not that effective).

FLEET ORDERS
Fleets would have configurable orders, allowing you to engage in a 'Raid' or 'Assault' or 'Invasion' or 'Fleet Combat', which would change their targeting priorities, their likelihood to disengage or retreat. A Raiding fleet would have a high chance to retreat or disengage, and would target cloud scoops and orbital shipyards, disabling them for an extended period, before targeting fleets or starbases. An assaulting fleet would target fleets or starbases first, but would retreat or disengage when damaged or out of supplies. An invading fleet would be the least likely to retreat, and would destroy/disable all orbital infrastructure then lay siege to inhabited planets.


GLOBAL AND LOCAL RESOURCE MODEL
The only global resource is credits. All other resources are locally stored (at the system level) up to the capacity of the local starbase and any inhabited planets. A starbase with one or two inhabited planets would have ample local storage, that could be increased by tech, as well as infrastructure (modules, buildings).

You can set up trade routes between starbases to ferry resources back and forth (X per month). The limit on number and size of trade routes can be determined by tech, policies, government type, and probably also infrastructure. e.g., you could add a 'trade module' to a starbase to increase its trade volumes. When a system is captured, a fraction of the stored resources are also captured. To blockade a star system, you would launch a raid on the star system, and all its trade would be suspended for a fixed number of months. Continuous raids would therefore serve to blockade it. You could also interdict trade routes (capture a % of traded goods per month). A fleet set to 'Interdict' would idle on the edge of a star system (outside of range of defenses), for a fixed number of months, while they bleed off a percentage of all trade goods each month. Thus they are subject to being intercepted if the enemy gets a fleet there in that time frame. Systems being 'interdicted' would be unable to cancel their trade routes due to communication jamming.

Trade resources can include metals, alloys, food, as well as fuel, fighters, missiles. A starbase in a frontline system with no colonised planets to supply it, can therefore be kept stocked up on fighters, missiles, fuel, ammo by trade routes that supply X of each a month (unless its blockaded by a raiding force).


SUMMATION
The FTL mechanism would allow you to spread your fleets along and behind border systems, and concentrate them in response to enemy warp movement, whether in attacking or defending. It would allow for feints, holding actions, and dispersed combat. You might for example send a small force to raid one star system, and the moment the enemy launches an intercept, you send a larger force to raid or interdict or assault a second system. It would require several raids on a strongly defended/populous system to degrade its defenses to the point where it can be assaulted or invaded. All warfare would be based more on attrition of enemy infrastructure until your fleet has an advantage.

The only de facto infinite weapon you have is energy weapons. Once a fighter is destroyed, you have to build another one. If you have launched all your missiles, or fired all your ammo, you have to go fetch more, or build more. There is also a place for 'Tender' ships, which can refuel or resupply your fleets, or even slowly manufacture replacement fuel, fighters, missiles, ammo. You might have to park it over a gas giant or in an asteroid field for a while, while it does this. Very large ships (battleships, Titans) might even have access to a ship module which provides this, so they can act as a tender in a pinch. A constructor ship would probably be able to double as a fleet tender.

 

[KingAlamar]

Ahhh ... the good old Doomstack discussion returns. Nice to know that some things never change

What I think is really interesting is how multiple systems interplay that really drive the importance of having Doomstacks. IMHO part of the tradeoffs we have with limited numbers of static defenses, chokepoints, space terrain, FTL type, map generation, time to "conquer" a system, player bandwidth, etc. actually reinforce the importance of Doomstacks.

EDIT: Removed explaining the reasons why this is the case as others have more or less already addressed it.

EDIT: Also removed wish for an "evolving gameplay loop" as part of the solution to various issues the game has ...

 

[KingAlamar]

Combat in my imaginary Stellaris II

<SNIP>

SUMMATION
The FTL mechanism would allow you to spread your fleets along and behind border systems, and concentrate them in response to enemy warp movement, whether in attacking or defending.

I was thinking along similar lines with one gotcha. I HATE the old problem of 'whack-a-mole' so much that if we went back to a more open gameplay that warp allows that the AI to assist with some of the downsides of such a system.

FYI: I remember an OLD game -- Birth of the Federation -- and it had a feature or two similar to what you explain. I really miss the concept that I could give some type of "Intercept" order to my fleets. If we could trust the AI to execute orders given then that would help a LOT.

EDIT: I only mentioned the Intercept order because you talked about it. I LOVE that feature and don't know of too many other games that have something similar.

 

[GC13]

Doomstacks existed long before chokepoints were added to the game. Doomstacks exist because there are no targets of strategic importance in a war other than the enemy fleet, and no strategic asset of yours other than your own fleet.

Military strategy is all about being able to concentrate more force than your enemy, but the realities of war give you the incentive or necessity even to split your forces; in Stellaris you just tell everyone to fly together because there is no reason to do anything else and every reason to do so.

I said this years ago, but look at a game like League of Legends if you want to see military strategy happening in real time: teams will naturally concentrate to pursue or defend strategic objectives, then disperse to seek smaller advantages when there are no such objectives to push. Stellaris is like a game of ARAM: there's literally nowhere for your fleet to be other than together and facing the entire enemy fleet. Objectives win games.

 

[KingAlamar]

Doomstacks existed long before chokepoints were added to the game. <SNIP>

I don't disagree. Specifically I would contend that the design of the game actively ENCOURAGES formation of Doomstacks. Oddly changes around 2.0 & 2.1 to mitigate doomstacks actually encouraged their formation even more.

If getting rid of doomstacks would make the game more fun [debateable] you'd have to create an enormous opportunity cost to using a doomstack. Something akin to the idea that if even a SMALL enemy fleet runs through your empire unchallenged it would do completely disproportionate damage to your ability to continue a war. Sun Tzu would laugh at the downsides of this system because, ironically, you'd be forced to protect everywhere not knowing where an enemy would strike -- thus also being weak everywhere

 

[MichaelJanuary]

FYI: I remember an OLD game -- Birth of the Federation -- and it had a feature or two similar to what you explain. I really miss the concept that I could give some type of "Intercept" order to my fleets. If we could trust the AI to execute orders given then that would help a LOT.

EDIT: I only mentioned the Intercept order because you talked about it. I LOVE that feature and don't know of too many other games that have something similar.

To be clear, by intercept, i meant sending your fleets to the same star system that the enemy was heading towards, in the hopes of catching them there. Not intercept as in meet them in interstellar space.

 

[Ragnarok Ascendant]

I don't disagree. Specifically I would contend that the design of the game actively ENCOURAGES formation of Doomstacks. Oddly changes around 2.0 & 2.1 to mitigate doomstacks actually encouraged their formation even more.

If getting rid of doomstacks would make the game more fun [debateable] you'd have to create an enormous opportunity cost to using a doomstack. Something akin to the idea that if even a SMALL enemy fleet runs through your empire unchallenged it would do completely disproportionate damage to your ability to continue a war. Sun Tzu would laugh at the downsides of this system because, ironically, you'd be forced to protect everywhere not knowing where an enemy would strike -- thus also being weak everywhere

It'd be nice if the majority of someone's resource output was concentrated in space rather than in planets, and if you had a stance that targeted resource platforms. Then you could tear through someone's economy by blowing up every mining and research platform you could find.

 

[MichaelJanuary]

Please no. I do not want to have to deal with supply chains of any sort in my space opera genocide simulator. Key strategic resources required to construct advanced spaceships components? Yes, please. But supply chains? Nope, nope, nope.

Its all in the implementation.

A global resource model is fine up to a point, but I would like the ability to cut systems off. Like if I blockade or interdict a system with shipyards, they should not be able to draw resources from the global pool by magic and continue building ships. I had one game where I had occupied a system with the enemy megashipyard, but my troopships were on the other side of the galaxy. So I parked my fleets at the megashipyard and just kept destroying one ship after another for years as they came off the production line. The bloody AI never stopped building.

There are many systems for how to do that. You could do it be requiring a route to the sector capital or empire capital, or you could simply disable or pause production while enemy ships are in-system. (The system is flagged as blockaded). Or you could slow construction down by a fraction, and let the enemy ships siphon off some of the resources (The system is flagged as interdicted). But having a completely surrounded or occupied system continually building ships is just stupid and immersion breaking.

The same for a mining planet. With no farms. You can blockade it for years, but the planet will never starve. The fact that its running a deficit on food (or crystals or motes or energy) is irrelevant in current Stellaris.

 

[ctl3]

Generally I oppose local resource model in this thread because it's more a economy rework than military. More specifically I disagree following statement:

"Blockage should be a thing in interstellar warfare so a local resource model should be implemented."

However I do agree following statement:

"We've designed a economy system with local resource model and proved it's better than current economy system so let's implement blockage mechanism."

When we make a proposal of mechanism change, we should consider if the scale and the domain is proper. Blockage is a minor aspect of warfare but local resource model is a major overhaul of economy system. Overhauling the entire economy system for the sake of blockage mechanism is kind of reverse the causation.

 

[HallowedError]

What if we had to populate starbases and even mining systems, making them specialized habitats. This would make them more valuable and spread out the things you want to blockade/capture so that doomstacks are less effective. I would think this would make systems you've upgraded more valuable giving a system wide bonus for how many pops you have working it instead of the baseline automated mining/research hubs.

If Blue attacks Red in one stack Red can get a raiding fleet past them and murder/blockade mining systems with valuable pops and potentially crippling their economy depending on what system they've chosen to prioritize.

 

[Ragnarok Ascendant]

....Okay, making basic construction cost pops is probably out of the scope of war and into the topic of 'The Economy Rework That Will Cause More Rage Than Megacorp's'.

 

[HallowedError]

Half the suggestions would have an impact on how economy works but I yeah, I get your point.

EDIT: I also wasn't trying to say it should cost pops but I think upgrading should open up slots to put pops into

 

[Iosue Yu]

If you think closely, occupying uninhabited star systems is a way of blockade. The resources go to the occupying country, deduced from the original owner.

The difference goes to only inhabited planets have to be landed and invaded before they start contributing resources to the occupying country.

Now the main issue goes to if certain actions are treated as blockading a planet, where those resources go and how long does it take to begin the starving process for that besieged planet.

An easy solution is to always store a local surplus for only Food and Energy for each planet. When the local surplus is filled, the rest goes to the country's store. It isn't even that difficult to implement because you only need to apply a local modifier called "stored energy %n" and "stored food %n". The only needed codebase change is to allow manipulation of this variable number. It should be already supported unless the Scripts are really that bad basic addition and subtraction are not supported.

When a planet is being besieged, it stops contributing to the country pool, and all resources it generates get stored locally. If there's a deficit, it gets funny effects after running out of local food and energy. When the siege is finished, all the stored resources get transferred to the country.

It's not an economic overhaul. If you think it is, then you are unnecessarily complicating things.