Analysis of Point Defense Efficiency

[Exultant (OP)]

We've had several rather spirited discussions about Gauss vs. Rail efficiency on the forums, but Steve's changes to the way beam PD works has thrown all of that out the window.

I've had a bunch of time to play around with my new PD calculator, which you can find here: https://aurora2.pentarch.org/index.php?topic=13619.msg170874#msg170874, and I've encountered a few rules-of-thumb by playing around with 10cm rail vs. full sized gauss.

There are a few assumptions that need to be made to compare PD efficiency:

1.) Trying to block every shot with final fire PD is massively sub-optimal when anywhere near tech parity with your opponent.
2.) Crew and MSP costs to account for the 1% misfire rate must be considered when comparing HS efficiency. Because it's not reasonable to just assume you're firing the weapons to stop a single volley of missiles, as a baseline I'm going to consider MSP costs (in HS) to fire against 10 separate volleys (Defined as a # of missiles striking the fleet in a single 5s increment, independent of salvo size) of equal size. For note, you can fit 400 MSP in 1 HS storage.
3.) Capacitor 3 is used for railguns so you are never caught with your proverbial pants down.

I'm going to go ahead and assume forum user bean is correct when they said that due to the changes, there is no reason to use less than full size gauss as there's no way to catch the upside of the greater variability.

This has us comparing full gauss to full 10cm Rail. Since we want greatest efficiency, we'll use quad turrets to get a 20% discount on turreting costs.

Catching leaks falls to CIWS and/or shields. Both systems must be added per-ship, so are inefficient from a HS consideration. CIWS is also expensive to fire. A single setup is twin 50% shots, so equivalent to a more swingy single gun turret, and the BFC range is capped to 30kkm, meaning it's chance to hit will be reduced compared to your beam PD. With CIWS, you pay for better tracking and you pay for ECCM on every setup. At Ion era tech, it costs approximately the same amount of MSP to fire a CIWS as it does a full size single gun gauss turret.

Shield tanking on recharge alone is very difficult unless you intercept most shots and you have good recharge. Still, at very low leak %s it's cheaper to catch the periodic leak with shield than CIWS, especially since shields help against beam weapons too.

That gives us 2 theories for optimization.

1.) Compare Rail to Gauss with similar leak % rates (I've been aiming for 2-5% in my tests).
2.) Compare Rail to Gauss with similar total CIWS needed to catch the leaks.

The first one is volley independent. With similar leak rates, you will need similar CIWS totals regardless of the number of missiles downrange (as long as you have enough shots to make the set number against every missile).

The second one accounts for the fact that it takes a different total number of shots to hit a missile between Rail and Gauss unless you're already moving at 4x tracking speed. This allows you to make less total shots, relying more heavily on the CIWS to pick up the leaks, especially if you can keep the total CIWS needed to 1. It also is very volley-size dependent. Bigger volleys mean bigger total numbers of missiles needing to be killed by CIWS. This is useful for low-volley sizes, but against spoiler missile spam you're going to be in for a lot of hurt.

Given the above, I'm going to compare systems at similar % leak rates.

Looking at the spreadsheet, we can see very quickly that it is MUCH cheaper to fire a railgun than it is a gauss turret, even when considering that misfire is rolled per turret, not per shot. We can also see that as railgun range techs increase, it gets more expensive to fire. Generally speaking, Rail becomes more expensive than gauss somewhere around 60kkm railgun range. While better gauss tech does increase turret cost, you need less total turrets to shoot down the missiles. Gauss 8 rolls for misfire half as much compared to gauss 4 without being twice as expensive (though close). Gauss range does not affect turret costs.

Finding 1: As long as we keep railgun range below 60kkm, you will get more longevity out of a railgun vs. a gauss turret if you are MSP limited.

Assuming your turret gear tech always stays at parity with your BFC tracking tech, the final thing that matters is your vessel speed vs. the gauss tech. With it being cheaper, as ship speed approaches 4x tracking speed, the better and better railguns get. even at max gauss tech, 8 shots at 300HS is equal to 4 shots at 150HS, and Gauss is always at least twice as expensive cost-wise.

Finding 2: In all cases, if you can go 4x tracking speed, always use 10cm rail, assuming 4x shots.

Likewise, at a ship speed at 2x tracking, you'll be at parity with Gauss 4, after 10 volleys, so long as you are using railguns with 10kkm range. If you up the range at all (assuming capacitor 3) railguns get more expensive initially (but will ultimately win the MSP game for long protracted missile defense)

Finding 3: If you are at ion era techs and your doctrine does not involve high speed fleets (>2x tracking speed), gauss should be used.

Finding 4: At Gauss 6, rail becomes superior at 3x tracking, so long as you keep with 10kkm range railguns.

Finding 5: In nearly every near-parity case, Rail becomes more efficient after 20-30 volleys, so long as railgun ranges are kept low.

Do note that I have not included missile decoy considerations on the attacking missiles. My hunch is that railguns may be superior only in that more shots = more chances to actually target the real missile, even though your % chance to hit it is lower, but I have no data to back that up.


Final Considerations:

All of these comparisons have been in cases where we didn't have to worry about total ship size. As we are all aware, the faster we make the ships go, the less payload space we actually have. Achieving 3x tracking speed on a beam ship leaves very little room for guns unless you are making significant sacrifices elsewhere. Attempting to get a large number of railguns on fast ships will have you building many more fast railgun ships (with all the associated costs) than you would a slower gun barge with gauss turrets. The ultimate efficiency test would be to find out how many guns you can fit into a ship, and consider total BP cost of the ship vs. % of the total PD power that ship represents.

[nuclearslurpee]

even at max gauss tech, 8 shots at 300HS is equal to 4 shots at 150HS, and Gauss is always at least twice as expensive cost-wise.

Note that the bolded bit here is not quite true, because the 150-ton railgun requires a power plant which adds nontrivial tonnage. On the other hand, railguns on slower ships require smaller BFCs than Gauss turrets, so how the cost/tonnage balance plays out will depend on tech level, fleet speed, etc.

A critical factor working against railguns is that having a high fleet speed also has a higher cost, you will need larger engine mass fractions and/or higher boost modifiers (and corresponding fuel consumption). Some players like to cheese the NPRs at a tactical level with ships that are basically a spinal laser bolted to a massive engine with 20+ layers of armor, aside from the fact that this gets boring quickly (IMO) this approach has severe strategic costs despite its apparent tactical efficacy. Keeping a slower fleet speed and using Gauss turrets may be a more cost-effective approach when these factors are considered.

A key advantage of railguns is that they are clearly better for anti-ship work, with twice the effective weight of fire and lower costs per weapons failure. 10cm railguns should never be a primary anti-ship weapon but once the range is closed they can end a battle quickly. On the other end of this scale are laser or meson turrets, these are the least efficient as pure PD weapons but they offer the most anti-ship dual-purpose utility (and, since 2.2, anti-laser-warhead capability).

Personally, I find each option to be equally viable as each has different strengths and weaknesses, which I think is the sweet spot for Aurora's game balance.

[Exultant (OP)]

You're quite correct about the power plant issue, which slipped my mind. The other side to that is that a 100% boosted 1.1 HS powerplant at the same tier as Gauss 8 tech can power 30 10HS railguns which can have a range of 90kkm compared to Gauss's 60, so while technically gauss is more space efficient, it's only by 0.03 HS per gun and 60% range.

You and I are speaking the same language about boosting a ship fast enough for railguns to become efficient in a fleet setting. Even at 2x tracking speed in the NPE era, I was struggling to fit more than 7 rail onto a 10kton ship because I was using 50% of my total hull for engines. I'm currently considering a doctrine of slowing down my PD ships compared to my beams, and using them as a staging area during the main engagement, separating the beams to chase down stragglers after the incoming missiles have been depleted. I have some spoilers to test this theory on, but NPRs may be a bit more challenging if I get ambushed while away from the PD.

Personally, I was always on team Gauss under the old system, advocating for 17% turrets all the way once you hit Gauss4 tech. Now that the changes are here, I was surprised to find how well railguns hold up, primarily due to the MSP costs. The choice between them really does seem like a sliding scale now depending on tech and fleet speed, whereas before I never really considered build cost to be a largely important factor.

Since small amounts of CIWS seems to be an almost necessity unless you're willing to house massive quantities of MSP on your ships, we've even lost the tech-cost argument, since we're strongly encouraged to research gauss anyway.

[nakorkren]

Since small amounts of CIWS seems to be an almost necessity unless you're willing to house massive quantities of MSP on your ships...

Slightly off topic, but rather than carry CIWS, why not use that hull space for shields? Having any shields at all helps protect from microwave weapons, prevents accumulation of (small amounts of) damage from battle to battle, and removes a nice chunk of damage "off the top" from any missile salvos that take more than a few minutes to reload. Plus (as far as I know) shields don't use MSP in combat, whereas I assume CIWS does.

[Exultant (OP)]

Shields recharge per 5s = RechargeTech*ShieldHS/60.

To recharge-tank a volley, you need Recharge = ASM Damage * # Leaks/Increments between volley.

Against my missile-using spoilers, they launched ASMs every 15s and each leaked missile that hit me dealt 7 damage.

Every missile leaked would take me HS = 7/(3recharge/60i)/3 increments = 47 HS of shield to tank on recharge alone. My ships are only running 20HS of shield (60 total points), so I could withstand ~9 missiles hitting the ship before I lost shields, not including the shield recharge which would be (assuming 1 missile leaked per volley) 27 increments, so I'd get 3 extra missiles blocked before I lost shields, so 12 missiles. This does not include size efficiencies.

12 missiles sounds great, but when NPRs and spoilers go missiles, they go all-in and you're going to take way more than 12 leaked missiles before you run them out of storage.

Shield tanking starts to look a lot better with bigger HS ships, because a shield that big represents a smaller total fraction of your available HS. At my 5% HS doctrine for shields right now, that means I'd need to be making ~50kton warships to shrug off single missile leaks without worry.

[nuclearslurpee]

Keep in mind, shield recharge is much less of a limiting factor against enemies using reduced-size launchers, which are usually the most challenging case for point defense due to the larger possible salvo sizes. The other major challenge is AMM spam, which is hard to stop by any means but incurs relatively low damage from leakers.

[nakorkren]

@Exultant: As I alluded to about slow-reload salvos and as nuclearslurpee pointed out more explicitly, the benefit of shield recharge is in helping you absorb leakers from enemies using reduced-size launchers to increase their salvo size, which for PD is the primary threat due to large salvo sizes overwhelming beam PD. You're not going to be tanking salvos entirely on recharge unless you're just way bigger than your opponent.

Also, I've heard a general rule of thumb tossed around that your shielding should be roughly 1/3 of your total armor, i.e. if you have 6 rows of armor x 100 columns you'd have 600 armor and be aiming for 200 shield. If you follow that design philosophy, you end up with more significant shield capacity (albeit less armor). I've been playing around with pretty large ships recently (60 to 80ktons) with heavy shielding (8% HS is armor, 11% HS is shields, 25% engines, i.e. 1640 armor and 1422 shield) and a single dreadnought of that size has absorbed the full magazine capacity of an NPC's entire fleet through a combination of PD, shield absorption, and shield regen without taking much if any armor damage. This is the origin of my caveat above about "unless you're much bigger than your opponent."

Another consideration is that shields a very helpful for mitigating laser-warhead missiles. Laser warheads generate lower damage but deeper penetration than normal warheads, and either avoid beam PD entirely (if you're relying on railguns or gauss) or significantly reduce the PD effectiveness by detonating at higher BFC ranges and hence lower hit percentage chances. Since shield efficacy doesn't change with range the damage is applied at, when beam PD efficiency per HS goes down as with laser-warhead missiles, shields look more attractive.

To restate/summarize my hypothesis: Shields are useful for a wide range of reasons and are an important part of a well-rounded PD system. Even a small amount of shielding (defined as a single shield generator at the max size you can build, once shield tech gets decent around say delta/epsilon) will help significantly with PD against enemies using reduced size launchers or box launchers (which are the biggest threat vs PD).

[Exultant (OP)]

You both raise good points, and I do believe that I discuss the interval factor in my math above. I haven't encountered reduced size launchers, but since reload time increases faster than size reduction, that makes shields more attractive.

I would like to point out that I just discovered a few hours ago that in the current version of Aurora, CIWS is completely bugged and does not actually stop missiles, so currently CIWS is a non-factor. Shields are our only option for leak catching at the moment if you don't want to armor-tank. As we all seem to be in agreement on, this pushes ship design to be bigger so we can have stronger shields.

(8% HS is armor, 11% HS is shields, 25% engines, i.e. 1640 armor and 1422 shield)

You and I currently have very different design philosophies. I'm running 9% armor, 5% shield and 50% engine at the moment on beam-craft and had even toyed with going bigger so I could chase that elusive 3x tracking speed, which doesn't give me the spare tonnage for shields the way you have set up. Are you a missile doctrine, or are you very high tech? I just got to Ion in my game.

I haven't even begun to look at anti-laser warhead defenses yet. I had been toying with the idea of 10cm lasers for that role, but they're hideously inefficient. More than likely I'd try to swap to an AMM defense doctrine if I was encountering laser warheads.

The biggest factor that has me not as worried about laser warheads is that ion tech is something like a 30-40% conversion factor. A S6 missile that I can make will give me a WH9, so that gives me laser warheads at 3-4 damage (Same depth penetration as a WH9 missile) and my range increment is currently I think 40K? I'm currently researching Gauss 40k velocity, (and gauss range is free) so until much higher levels of laser warhead tech I should be safe. Firing a laser warhead from outside the lowest increment seems to be massively inefficient in a cost-to-damage ratio. Laser warheads will become more deadly at higher efficiencies and once the increment exceeds gauss range (or gauss range at that tech level, since the jumps are farther between) or vs. railguns which are cost-incentivized to stay low range. I do agree that shields are going to be highly effective against laser warheads due to diminished total damage.

[nakorkren]

You and I currently have very different design philosophies. I'm running 9% armor, 5% shield and 50% engine at the moment on beam-craft and had even toyed with going bigger so I could chase that elusive 3x tracking speed, which doesn't give me the spare tonnage for shields the way you have set up. Are you a missile doctrine, or are you very high tech? I just got to Ion in my game.

I'm on magneto-plasma engines, but I'm using 250% boosted engines for that particular ship. That means my 25% of HS being engines makes my ship go 10km/s, putting my ship at 2.5x racial tracking speed so I can use 10cm railguns as primary PD weapon and lets it choose when to engage with enemy fleets. That makes the ship a fuel hog and have only a 8B km range, so I tow it into systems of interest with an armored tug and support it with a dedicated tanker/jump tender. If you want to see what an earlier version of that ship class looked like at Ion tech level, I posted about it in a different thread https://aurora2.pentarch.org/index.php?topic=11100.msg170053#msg170053 Note that the earlier version attempted to use both 10cm railgun and AMMs, I later shifted to eliminate the AMMs and went with all PD on the ship being 10cm railguns. I normally escort these dreadnoughts with carrier(s) loaded with 10cm railgun fighters and/or missile fighters, which gives me some tactical flexibility if needed.

Back on topic for PD, your bug report on CIWS not actually destroying missiles is really significant given the changes to missiles, and your test of it seems sound! Thanks for identifying that, hopefully Steve sees it and confirms/refutes/fixes.