Revamping SC and GW Attacks and Defenses

[Cordane]

I've been working a long time on suggestions for how Strike Craft and Guided Weapons should interact with the defenses of Stellaris warships. Mostly they've been about changing defenses to move away from unbound layered fields of Point-Defense and Flak fire. My recent post on changes to ship classes, sections, and slots and how those get upgraded over time, made some allowances for utility slots to not just be "half shields, half armor, or possibly the odd crystal hull".

Let me talk first about an issue that has bothered me about this game for a long time: Stellaris does not know how to simulate SC and GW movement. Warship movement is a bit better (still not anywhere near ideal), but the problem with SC and GW movement is that it's basically just warship movement only a little bit faster. But why are SC and GW only a little bit faster? Without even trying to pin down any actual sizes, any of the SC and GW objects should be several orders of magnitude smaller than even a Corvette. Assuming comparable technology and proportional, if not outsized propulsion and maneuvering systems, SC and GW should be far better at acceleration than any warship.

Yet the game's visuals makes them only slightly faster - in part because the engine wasn't built with them in mind, and in part because it allows the visual spectacle of relatively slow-moving blips moving through the laser lightshow of PD and Flak fire. Those defensive systems are IMO horribly implemented and balanced, within the context of the game, with fields of fire that are too large, too layered, and too effective. By having extreme levels of PD/Flak mitigation, the base and/or bonus damage values for GW and SC have to be high enough to attempt balance that they are then completely overpowered when not strictly defended against.

So let's work on some solutions:

The relatively slow SC and GW movements have to go. With much higher accelerations, SC and GW should cross any reasonable battlescape in hours or days, not weeks. With the day clock in use by Stellaris, it doesn't make sense to try to visually simulate them exiting their warships, moving toward their targets, attacking, and in the case of SC returning to their carriers - there just isn't enough time. Attacks need to shift to be more like the direct fire attacks of Kinetics and Energy Weapons, with a quick animation and a resolution of the attack either immediately after it "fires" or with a slight delay.

PD and Flak are too easy to throw out of balance - these are removed and replaced. I spent an inordinate amount of time trying to come up with layers that mimic the effect of PD, Flak, and even defensive fighter screens at the fleet level and at the ship level, but all of them either were too good at their job or required such extreme damage increases to counter them that GW/SC were massively overpowered when no primary defense was offered. I was also trying to stay true to my desire to see all defensive layers be somewhat useful against most attacks. So I came back to Countermeasures.

Countermeasures act like decoys and sensor jamming, causing a given percentage of attacks to instead be wasted on the CM layer. This makes CM a hybrid ablative layer, with attacks that hit drones applying their normal damage against it, but allowing significant percentages of attacks to pass through. Higher Tiers of CM modules block higher percentages of regular attacks, with Sensors reducing the percentage by about the same per Tier (GW and SC instead get better entirely on their own and don't gain this benefit from Sensors as well). There will also be a slightly higher chance and a higher multiple for attacks from smaller slots and the reverse for larger slots, as the decoys are killed easily but don't pass their overkill as much.

I've figured out the numbers for basic attacks (vs. specialist or advanced weapons), by weapon type:

CM         % Chance   +/- Tier   Rate   AvgDmgPen  AvgCMDmg
Guided        40%        10%     0.25      60%        10%
Strike        40%        10%     0.25      60%        10%
Kinetic       20%         5%     1.0       80%        20%
Energy        20%         5%     1.0       80%        20%

Having multiple slots with CM modules of the same Tier doesn't change the percent chance to block the attack (just add the layer points), while CM modules of different Tiers use a weighted average to determine the percent chance (e.g., an M-slot Tier-3 and an L-slot Tier-2 average out to "Tier-2.33" (3x2 + 2x4 = 14 / 6) and would block 53.3% of GW/SC and 26.7% of Kin/Nrg attacks, before they apply their reductions). Note that my "Hull Classes, Sections, and Slots" suggestion has possible mixed slot sizes for utilities, but it should also work for vanilla utility slots.

What makes this work is a shift in how GW and SC attacks affect the other defensive layers. Rather than having one or more defensive layers (PD, Flak, fighter screen) that only affect GW and SC, but letting them ignore Shields, basic GW and SC now are affected by all three outer layers (CM, Shields, Armor) but so are basic Kinetic and Energy weapons. I've shifted the bonus damage profile for basic GW/SC attacks to be a 1.25 damage multiplier against Shields, Armor, and Hull. When effective Damage Per Day (base damage per hit, effective Accuracy, rate of fire, with some allowances for range) for all four attacks is the same getting to balanced defensive layers (1/3 of utility slots as CM, 1/3 Shields, 1/3 Armor, combined points equal to Hull), the four attacks get to 0 Hull points at the same time on average. (The Hull damage bonus may need to be adjusted to compensate for early-game ratios of CM/Shield/Armor points vs. Hull points being lower.) They also line up well when going against favorable defensive setups (e.g., GW/SC with no CM, Energy w/o Shields, Kinetics w/o Armor).

Note that I said basic attacks - there should be specialist attacks that penetrate CM layers easier (e.g., Torpedoes have better internal sensors, giving them a lower % Chance but a smaller Rate, as their bigger warhead is wasted on a decoy) or strip CM layers faster (e.g., Swarmer Missile salvos are intentionally "decoy naive", and have a higher % Chance and/or ignore their normal Tier bonus, but a larger Rate, as multiple missiles seek out decoys to destroy). There should also be specialist GW/SC attacks that prioritize damage versus the other defensive layers, and advanced attacks (similar to Kinetic Artillery) that have favorable effects versus multiple defensive layers.

Easily the biggest issue even I have with going to a Countermeasures system is that it does away with Strike Craft attacking other Strike Craft - no World War One-style dogfighting cloud or even WWII-style bomber formations being harassed by fighters. Combined with eliminating the fleet defense aspect of PD and Flak systems, this change is a big hit to the more traditional Stellaris screening ships - I've got one potential mollifying suggestion with having different styles of Hangars, including ones that are designed for shorter operating ranges to concentrate attacks on stalking Corvettes (versus focusing on distant artillery Battleships). Another is to have separate Computer or Auxiliary modules that give (non-stacking) aura effects to boost CM % Chance for allied ships or inhibit % Chance for enemy ships.

One aspect that would possibly get overlooked is how the more organic space foes would deal with CM-type effects. Would their attacks be affected in the same way by CM modules, and would attacks against them be reduced similarly? Also, there are probably multiple variants that can be designed for GW/SC to get through or knock down the various defensive layers, and Energy already has a full menagerie of options, but Kinetics might need some real help to come up with variety.

 

[Pancakelord]

Interesting, I agree strikecraft are not as good as they ought to be and much of this stems from underlying mechanics. Whilst PD/missiles stem from (in my opinion) a too-simple tracking system that over-rewards spamming PD.

With much higher accelerations, SC and GW should cross any reasonable battlescape in hours or days, not weeks. With the day clock in use by Stellaris, it doesn't make sense to try to visually simulate them exiting their warships, moving toward their targets, attacking, and in the case of SC returning to their carriers - there just isn't enough time.

Just ignore the in game calendar. The battle system runs on it's own terms, it's pointless to try and apply real-world physics to this. It's best think of combat movement speed in relative terms (e.g. missiles are X times faster than SC). But this quickly runs into a problem which is... instant-hit weapons.

Attacks need to shift to be more like the direct fire attacks of Kinetics and Energy Weapons, with a quick animation and a resolution of the attack either immediately after it "fires" or with a slight delay.

I'd argue the opposite, I think hit-scan weapons are a big mistake in the combat system. If everything becomes instantaneous you effectively dominate with whatever has the highest range. I.e. Kinetic artillery builds. then just spam that like nuts for the alphastrike to overcome any firerate limitations.

Instead all weapons should have travel time (with the possible exception being PD - depending on how the combat engine is coded, projectiles tracking other projectiles could get pretty CPU intensive or lead to some nasty problems like recursion if mishandled), this would mean each KA shell would travel and not instantly hit (currently they instantly hit and then the engine "catches up" by rendering in a shell on a ray/line from the gun to the target), each laser beam would be a pulsed beam of light travelling through space, each burst of an autocannon would be a thick stream of rounds flying through space (again if they're treated as an aggregate projectile, its less processor intensive).

• This would also lead to more of a need to use diverse weapons, as overkill would be an issue (because these things cant retarget unlike missiles)
  • Indirectly buffing missiles, which can retarget thanks to them being "smart" munitions.

If all weapons are given travel time, Strike craft, with their front-loaded travel time, but long(er) tail damage+ability to retarget enemies, suddenly don't look like such a crap option, and higher tier strike craft could definitely do with some more interesting features to distinguish them some more from basically being "missiles that dont blow up".
An old feature from V1.0 of stellaris was that hyperdrives could FTL out from a star system without being at the edge of the gravity well. I remember once seeing a mod that let your fleet do micro-jumps during combat, too.
This indicates the engine is more than capable of some interesting things. such as:

• Specialised Tier 3/4 anti-strike-craft variants "interceptors/Adv. Interceptors" could be introduced that are fitted with Micro-hyperdrives, letting them launch and automatically FTL-jump right to enemy strikecraft in the system, with a cool down (e.g. can be used every 7 days and only triggered when a valid enemy is X distance units away).
  • This lets them have their lower movement speed but still allows them to close on enemy fighters over long distance extremely quickly.
    • I've screwed around with strike craft speed in the past through modifiers for personal mods, and currently if you set it to be too high, it'll spazz SC out completely and their dog-fighting loops get extremely large (going beyond the range of the solar system's gravity well sometimes lol), also reducing their real DPS as they're rarely in-range to hit eachother if they move too fast.
      • so making SC speed too great leads to more problems than it solves, hence offsetting this with a "micro hyperdrive".
• Bring back bombers as tier 2 units - with a special focus on countering BB(spam), Titans/Juggernaughts & Colossi.
  • You could have a Tier3 "Heavy bomber" [Tier 4 = Adv. Hvy Bmbr] variant that can be as a hard-counter to starbases(e.g. giving them an implicit +500% bonus damage vs bases & increase resistance to PD/Flack damage, but reduced resistance from strikecraft and generally slower travel time, so you'd need to compose your fleets with extra interceptors to protect them),
    • This would allow for starbases to be made into true fortresses (right now they're a pushover after the first 20-30 years of the game) that require specialist bomber-heavy fleets or titan guns to be cracked open.
  • You could have rarer tier 4 "Hyper-bombers" that will FTL back to their carrier to rearm, rather than having to fly back, halving their sortie down-time. (this also keeps the visual spectacle, as you'd see them all launching in wings several times over the course of a battle).

By having unique units dedicated to different ship targets, the game will be better able to balance different SC vs other weapon types - e.g. without a bomber-type SC I'll always skip using strike craft at all, and instead focus on spamming torpedoes as there is no way to balance SC against other weapon types currently as they are a "one size fits all" option.

On missiles VS PD... I don't think the current system works (well) at all, by tying chance to hit to the tracking of each PD gun vs evasion of each missile, it quickly devolves into spam more PD to win. But at the same time I'm not sure if we need to throw it all out entirely (e.g. to go with your countermeasures).

I think the game needs to decouple tracking/chance to hit from Missiles/Strikecraft and PD, instead it needs a battlewide calculation to better model the role of sensor arrays and the general chaos of having so many targets moving around.

If you've got hundreds/thousands of strikecraft buzzing around fighting eachother and thousands of missiles in flight you need an incredible Combat information centre & fleet sensor network to process all that, so that your PD can track and kill something - without dealing friendly fire. The current tracking calculation seems to skip all of that, which devlolves missiles/SC vs PD into the spammy state we have today.​

Now, It's been a while since I've had to think up an equation for anything like this so bear with me, but i've given a purely illustrative example of what I talk about above

Spoiler: Equation example

This is probably mathematically flawed when it comes to fleet scaling - If I gave this some more thought I could come up with something less hackish or more representative of all the ships in a battle - any nearby allied space station could also be factored in as a massive sensor array [e.g. a star fort could have size 10] giving your team a big advantage in tracking and killing enemy SC/missiles,
Equations:

A basic example with lots of corvettes, a fiew cruiser carriers and tier 3 sensors, (In theory chance to hit rises when you use more large ships with high tier sensors, representing their bigass sensors & networked targeting arrays - but large ships often lack as many PD hardpoints per alloy unit spent).

And chance to hit =/= ability to hit, you still have to field the point defence to shoot things down.
So chance to hit simply tells you what % chance you have to hit a missile if it's within range of a PD gun, but your ability to shoot a slavo down relies on having enough PD to do the job (i.e. Net PD shots per tick = #pd*firerate, then mathematically averaged per tick).

 

[Cordane]

I'd argue the opposite, I think hit-scan weapons are a big mistake in the combat system. If everything becomes instantaneous you effectively dominate with whatever has the highest range. I.e. Kinetic artillery builds. then just spam that like nuts for the alphastrike to overcome any firerate limitations.

There are a lot of things to unpack from the vanilla alpha strike builds:

• Currently, Stellaris has almost no penalty for attacking at the maximum value of a weapon's range.
  • My preference would be to have direct-fire weapons take effective Accuracy penalties as the range gets longer, either by applying a similar penalty per range increment and making the increments longer for long-range weapons, or applying smaller penalties per increment for long-range weapons and having the same number of increments as the rest.
• L- and X-slot weapons - those with the greatest vanilla direct-fire ranges - have too easy of a path to hitting Cruisers, Destroyers, and even Corvettes - the ships that are most likely to carry weapons that can't match the L/X-slot range advantage.
  • Spinal cannons (X-slots in vanilla, Cannons of any size in my comment in the "Hull Classes, Section, and Slots" thread) should have horrid Tracking versus anything not 1-2 sizes larger than the warship they're mounted on.
  • The Tracking/Evasion system needs to be completely revamped, especially for the largest weapons - I have previously spent and would again need to spend a whole thread on this subject by itself.
• Too many of the top Tier weapon types are L/X-slot exclusives.
• The largest and latest weapons have the longest Cooldowns, "balanced" by higher per-shot damage, making their size advantages even larger.
• The only negative for range-focused ships and fleets is that (certainly vanilla) Stellaris makes very little effort to actually maintain the range the warships want.

Alpha strikes are an issue in and of itself - hitscan weapons (versus projectile-tracked) don't make this any worse.

 

[Pancakelord]

Too many of the top Tier weapon types are L/X-slot exclusives.

I agree. I'd also to say that some classes lack top tier L/XL slot equivalents - in part due to weapons never really getting a second, comprehensive, review after the introduction of G/P slots - like high Tier missiles for XL slots.

The largest and latest weapons have the longest Cooldowns, "balanced" by higher per-shot damage, making their size advantages even larger.

I think there must be some sort of technical limitation in the game, as I've never understood why they went with this method, rather than purely increasing dps you could have far more interesting [and fun] effects:

Imagine XL missiles with an air-burst effect, able to wipe out/cripple whole squadrons of SC/Corvettes within their detonation radius [e.g. dealing AOE damage over time vs armor/hull by spraying adjacent targets with "acid", or shorting out engines/debuffing evasion via "EMP"], or weapons that can hit multiple targets per shot - like lightning arcing from ship to ship, rather than blowing them up one at a time.​

The only negative for range-focused ships and fleets is that (certainly vanilla) Stellaris makes very little effort to actually maintain the range the warships want.

The fact that vanilla ships don't try very hard to control their range is probably why they went with a fixed accuracy calculation (though I'd prefer it if we had a more robust tactics/formation system, to tell the combat AI how we want it to fight, something like automatic party combat rules in a CRPG, instead of the current combat computers - then variable accuracy would totally make sense).

But, as things are, it'd be even more frustrating if your ship accuracy varied with their location on the battlefield, without giving you some way to influence that after the battle begins or via better pre-selectable AI combat rules.​

Whilst there are more things to it than just hitscan vs simulated projectiles, the biggest issue remains if you dont remove histcan, it's not damage that's the problem, its reaction time, a weapon that can instant-hit will always trump one that cannot.

in game terms it's the "Δticks" (tick = the smallest scale of the combat simulation, I think its 1/10 of an ingame day, if i'm remembering the default value of ticks_per_turn correctly) between your hitscan strike and their simulated strike.​

​

If one side uses hitscan and the other doesn't (or uses far less), on the first tick of battle, side B gets some fraction of it's fleet deleted within that first day of combat (on the 1st 10th of the day probably) and cannot then launch fighters, or fire back missiles etc. But if projectiles were all simulated this would be less gamey, because if those shells had travel time, fleet B could still fire back missiles or launch SC, even if those ships are doomed to die to the KA. Projectile simulation shouldn't need to pull on (m)any other parts of the game, so it's probably something that could be offloaded to a spare CPU thread (if done right).

• Fighters need time to get airborne and missiles can take a little while to launch a full salvo, this is a much bigger deal for these two weapon types specifically, as their whole shtick is range - just like L/XL weapons (and unlike S/M slots), but they mechanically can't compete with something that has no lead time before dealing damage, whilst having lead time of their own.
  • Turrets have 0 tracking time, it looks like they turn to face their targets, but in reality they snap fire if needed, further compounding the instant-kill nature of L/XL, whilst missiles have to arc to turn to their targets and fighters follow long attack curves.
• Neutering L / XL damage wont help here, if you neuter damage too much, people will just gravitate to another weapon, killing build variety, if you don't neuter it by enough, people will just spam out more ships to compensate. And nerfs never go down well, even if they are warranted.
• Likewise, killing accuracy at certain ranges, as mentioned above, would lead to frustrated players raging as their ships move out of "optimum range" with no way for them to correct for that.

Either way, this is all academic. I would like a proper overhaul of this, as combat is a huge part of the game. But I just don't see a big overhaul to weapon mechanics happening at this point in Stellaris' lifetime, be they hitscan>simulation, or evasion/tracking/accuracy mathematics getting replaced with something else.

If I had to come up with a more realistic/low-effort (read: band-aid) fix to improve the state of hitscan vs not hitscan weapons (to level the playing field even a little for Missiles / SC),

Spoiler: I would:

1. Force all hit-scan weapons to enter battle "unloaded" making them perform a reload as their first action, rather than getting a free shot (in the case of L/XL weapons that almost immediately start in-range)
   • Is it unrealistic to start a battle with an unloaded gun? Probably. But we're playing a game that includes talking mushroom people, so 'realistic' is relative. One can always waive this as time for the guns to charge. or whatever.
   • In the case of larger guns this would at least prevent them firing for the first few days of an engagement, giving ships with non-alpha G/H weapons a chance to do something before they're struck down.
   • And small guns would be loaded/charged up by the time they get in-range so this doesnt affect them.
2. Add small -evasion debuffs (i.e "inertial mass") to all ship components(except thrusters)
   • (e.g. a -evasion penalty if you add armor components to your ship, and a small -evasion penalty for adding shields).
     • This would make un-armored, un-shielded, thruster-heavy ships a weird little niche - glass cannons, in RPG terms, able to dish out damage but if they get hit they go boom very quickly.
3. Separate in-combat speed from in-system speed (there are mods that work quite well at faking/approximating this via ship events)
   • Then add -combat_speed modifiers to all components (and +combat_speedto thrusters).
     • This means massive, heavily armoured carriers are going to be the big un-evasive targets they ought to be, but because they have so much armor, slowing them down so much, they'll stay at range and actually make the most of what they've got (rather than YOLOing into enemy gun range).
     • And because this would be a separate speed property from non-combat speed, it wouldn't make ships moving around the galaxy even slower than they already are.
4. Add missile and strike craft repeatable/upgrade techs
   1. +% Missile flight speed.
      • The faster a missile goes, the fewer ticks it spends within range of a PD battery, giving it a better chance of getting through before the PD gun can reload and fire on it.
   2. +% Missile hull.
      • meaning more PD shots to bring down each missile, when you factor in PD fire rate and ranges, they can only get 1-3 shots off if it's 1 gun vs 1 missile, depending on how fast the missile is going.
   3. +% Strike craft wing-size. (more fighters launched per wave).
   4. +% Strike craft flight shields/Hull. (survivability)

Taking points 2/3 together: if components slowed down in-combat speed and reduced evasion (with armor being the most egregious), AP could be further buffed per module to compensate, giving:

• mid-tier missiles a reason to be used in a long-range anti-armor niche (where early-game missiles are good for everything, till PD shows up, or for knocking over starbases - and lategame missiles basically don't exist) as armor is pretty meh currently on most ships, to the point that it's quite easy to chew through even with gauss, so no need to bother with missiles that could be downed by PD.
• A greater reason to re-introduce specialised strikecraft, like bombers which get implicit dmg bonuses vs armor (or even special attacks that can bypass armor entirely, dealing direct hull damage.. if you can drop the enemy shields e.g. with a null void beam).
• If all ships are moving slower in-combat, with the biggest moving the slowest, this also means generic strikecraft (if other types are not added) would also be more useful as PD screens, because they'd have more time to dogfight or bring down missiles before the fleets finally mash up. Right now, in vanilla, things get so packed on top of each other that SC don't have a chance to shine so you're better off using just PD rather than SC, on an alloy cost per PD gun, basis.

Something i've noticed is that when missiles are launched within PD range [i.e. when ships are all in a big-fat-mess] they often instantly die, because the PD can knock out the missiles as they appear tick-for-tick, rather than being overwhelmed by the whole salvo at once.

These changes would atleast give G/H weapons a fighting chance in more cases, and may make fleets stick to their proper ranges a bit more.

 

[Cordane]

... {hitscan} ...

There are some issues with having instantaneous-effect versus tracked-projectile weapons in Stellaris, but switching all attacks to tracked-projectile is not a viable solution. The combat engine isn't even capable of handling the calculations and graphics for the limited number of SC and GW "projectiles" in the battle environment, never mind having 4-10 times as many if all "projectiles" are included. For no reason other than this, it's just not a good solution. Moving everything over to instantaneous effect, or at least instantaneous calculation and slight delay to application of effect, makes for a much cleaner environment for calculations. Would I love to see an RTS fleet battle game using the Stellaris IP, one that has really detailed controls and options? Absolutely, but this game is just not built for it.

Relatedly, Stellaris does have some elements of its combat system that are exaggerated in distance and time to make it easier for the player to be involved at the "clock = calendar" pacing and system map scale. The biggest roadblock to any change is the game clock - ship movement (at least at the beginning of the game) actually ties fairly realistically to what's shown on the system map, but the forced extreme length of battles ensures that fleets would quickly overrun each other if they started at a more realistic range (a handful of light-seconds or a few million kilometers). So they end up starting battles at visually light-hour ranges - I think the attacks themselves need to resolve on screen much closer in time on the game clock to how they would at more realistic ranges, despite the much larger scale distances on screen.