1.8.1 - The State of Combat Balance

[Fa1nan]

The Doomfleet had no unbidden ships in it; it consists of ships of all four organic FE types. I also described how the battle went: Cruisers die quickly, causing damage output to drop and once the destroyers are gone, the corvettes just fail to even dent the remaining ships. Since the corvettes do not even evasion tank (because no one fires at them), they are a waste of minerals. And why bring destroyers when you can bring armored cruisers or battleships?

 

[GloatingSwine]

The issue with SteveRaptors Fleet may be that the doom fleet has unbidden ships in it which totally wrack anything which is not a bs at high Range. This may bias all tests against small ships. It may be worth testing the small ship load out against non unbidden crisis

The issue with SteveRaptor's fleet is that it's a mixed fleet of primarily bad designs.

 

[GloatingSwine]

So, apparently, armor is better than shields. I did not test the missile based loadouts with all armor, but considering the performance of the shielded plasma cruisers compared to the Swiss Army Knife and the Deva, I doubt they would beat Plasma.

Armour is better than shields when it isn't directly countered. A large proportion of the DFoD isn't directly countering armour (The Machine FE is the one that is strong armour counters).

Also, armour is cheaper because it requires no power and so you get more ships for your buck. (Probably worth double stacking CFP instead of regenerative hull, to test that with a console empire use "research_technologies 1" to get the event and creature weapons). Having even a few more ships makes a big difference. The 422 fleet loses at 220 ships but wins at 250 with 3/5 of its strength remaining.


It might also be worth trying a GC-4Plas layout.

Can I get someone to try something? Suppose I have a Battleship with a Mega cannon and 4 Kinetic Artillery, and it's late game and I have 10 ranks of the repeatable 'High Density Munitions' and 10 ranks of the 'Loader Efficiency' repeatable techs (so, +50% kinetic damage, +50% kinetic weapon attack speed). Now suppose I'm going up against a more balanced build like the old 422 battleship, but because that player is splitting their research into 4 repeatable techs, they have less of each.

Actually, they don't. Because High Energy Focusing/Flash Coolant are physics techs and HDM/LE are engineering, they can have just as many of both.

 

[4o1XOHBV6In4]

So we have an interesting distinction here it seems:
When both fleets consist of ship classes and weapon configurations available to players it seems that shield based designs dominate hard and as a result shield ignoring designs are generally the way to go.
But when fighting against A/FEs armor based designs give better results, presumably because while they use the same components as us their ships have different modules and / or overinvest in PD (and PD counters almost all shield defeating weapons).

I suppose we jsut have to assume that as a given.

@GloatingSwine
If armor is better against shields when armor is uncountered, then what weapon setup for battleships does better against armor than against shields? According to your statement that should be every or at least many design that don't have energy weapons. But in all tests I know of both kinetic and missile setups end up performing better against armor than against shields. So what gives? Have I just not tested the designs you're thinking of yet? Maybe you could give an example

 

[GloatingSwine]

So we have an interesting distinction here it seems:
When both fleets consist of ship classes and weapon configurations available to players it seems that shield based designs dominate hard and as a result shield ignoring designs are generally the way to go.
But when fighting against A/FEs armor based designs give better results, presumably because while they use the same components as us their ships have different modules and / or overinvest in PD (and PD counters almost all shield defeating weapons).

I suppose we jsut have to assume that as a given.

When both fleets are built by players shields dominate because shield stacked with anti-armour weapons beats armour stacked with anti-shield weapons. AI designs, even for those with energy preference, don't have enough anti-armour. They'll waste slots on lasers and disruptors.

In PvE armour is best most of the time, the Xenophile, Materialist, and Machine FEs and Contingency are all weaker to shields though.

 

[4o1XOHBV6In4]

Edits... I hope you get an alert for my last message anyway. Probably should've made it's own post =/

 

[GloatingSwine]

The standard meta design is generally better against armour than shields. It's really the AI designs that fail against armoured ships because they tend to mix and match weapon sizes a lot, and that means most of their designs have sub-50% AP.

Like the energy preference AI will be building battleships with things like large disruptors and medium lasers and such.

 

[4o1XOHBV6In4]

The standard meta design is generally better against armour than shields. It's really the AI designs that fail against armoured ships because they tend to mix and match weapon sizes a lot, and that means most of their designs have sub-50% AP.

Like the energy preference AI will be building battleships with things like large disruptors and medium lasers and such.

Everything you said is right, yet I don't know that a design with large disruptors and medium lasers would do better against armor than against shield. I would suspect they do bad against armor and worse against shields.

Basically my hypothesis is "Shields are the superior defensive options against all non-naked, non-shield-ignoring ships players can construct, including all regular AI designs."
So my hypothesis predicts, that even terrible weapon combinations as employed by the AI do worse against shields even if they're already doing terrible vs armor. Unless they're shield ignoring (obviously).
So if you know of a design a player could construct (includes all AI designs) that does worse against armor than it does against shields you can prove me wrong (as long as it's not shield-ignoring). If no such design exists then shields are the superior defensive option against all regular AI empires that don't employ sufficient shield-ignoring weapons.
My follow up claim would be that AI empires generally don't employ sufficient shield-ignoring weapons.

@4o1XOHBV6In4 : Did you want to verify the test you ran verifying that point defence was actually worth building against missiles when compared to building more ships with weapons to kill stuff faster?

Just ran some really quick tests, I hope to revisit this topic later though. For now it seems that the A-200 outperforms the A-202 and both win against the T102 design. This is true for a minerals : minerals comparison and I used the following designs:

Spoiler: A-200 "Unshield"

Spoiler: A-200 "Unshield-PD"

Spoiler: T-102

If you're wondering about the empty weapon slots on the A-20X designs... I just think there are no weapons for those slots that are worth their minerals (and the additional power you need). Not against heavy shielded meta builds anyway.

Anyway, here are the results in detail (Min. = minerals cost; # = number of surviving ships):

    Design        | Min. |  #  |    vs    |  #  | Min. | Design

    Unshield-PD   | 1497 | 100 |    vs    | 167 |  896 | T102
                         |  62 |    vs    |   0 |
                         |  62 |    vs    |   0 |
                         |  61 |    vs    |   0 |
                         |  63 |    vs    |   0 |
                         |  60 |    vs    |   0 |
 
    Unshield      | 1487 | 100 |    vs    | 166 |  896 | T102
                         |  63 |    vs    |   0 |
                         |  66 |    vs    |   0 |
                         |  63 |    vs    |   0 |
                         |  64 |    vs    |   0 |
                         |  64 |    vs    |   0 |

 

[moyang]

Everything you said is right, yet I don't know that a design with large disruptors and medium lasers would do better against armor than against shield. I would suspect they do bad against armor and worse against shields.

I Second this. Once I tested Torpedo Cruiser loadouts against Materialist FE fleet, with different # of shields. And I used same number of ships in each test, so one can say this test was skewed for shielded ships, because they are more expensive.

The size of battle: 80 Cruisers vs 20 Battlecruisers and 40 Escorts

Results: All loadouts managed to win.
Maxed shield: lost ~30
2 shields: lost ~20
1 shield: lost ~10
no shield: lost ~20

 

[GloatingSwine]

Everything you said is right, yet I don't know that a design with large disruptors and medium lasers would do better against armor than against shield. I would suspect they do bad against armor and worse against shields.
[/CODE]

That's not what I'm saying though?

I'm saying that mixed weapon loadouts which do not focus on armour piercing do worse against armour than they do against shields. And that includes all AI empire designs.

You can see it in Fa1nan's test, where an all-armour design does better against the Doom Fleet than an all-shield design, winning with fewer losses.

Shield ignoring weapons aren't the point, the Arc/Bomber synergy only works if you monobuild it and the AI doesn't. The point is that at very high armour states you add far more effective hitpoints than you can with shields, even if those shields regenerate (because they get focused down), and the AI doesn't mount the weapons required to counter that.

Most AI fleets are going to have an aggregate AP rating somewhere in the region of 30-40%, and if you're at 90% armour you're running about 250% effective hitpoints against that.

 

[anomanderus]

I feel like a brainlet because I never go into this much detail with ships, just ending up with a swarming fleet that overruns the enemy.

 

[Buddlschlumpf]

I feel like a brainlet because I never go into this much detail with ships, just ending up with a swarming fleet that overruns the enemy.

Which is also valid as in Stellaria, sheer Numbers always win. But you do. not always have the numbers on your side, then Design Kicks in.

 

[4o1XOHBV6In4]

Everything you said is right, yet I don't know that a design with large disruptors and medium lasers would do better against armor than against shield. I would suspect they do bad against armor and worse against shields.

Let me try this more clearly. What I wrote in the quote is equivalent to:
"Mixed designs do not do better against armor than against shields. They do bad against armor but they do even worse against shields".

Or, if you prefer: Shields > armor vs random mixed loadouts.

That's not what I'm saying though?

I'm saying that mixed weapon loadouts which do not focus on armour piercing do worse against armour than they do against shields. And that includes all AI empire designs.

I understand you to say: Armor > shields vs random mixed loadouts.

So it seems we're at a disagreement. We've observed that against the A/FE Doomstack armor does in fact perform better and I don't disagree. What we disagree on is if the same is also true for regular AI empires. You said against regular AI empires armor > shields, I believe the opposite to be true.
Basically I tried to find a weapon combination players (and therefore regular AI empires) have access to that beats shields but is countered by armor. In different words, a loadout L so that armor > shield against ships with loadout L. I haven't found any other than torpedoes, arc emitters, strike craft [and possibly some gimped naked designs]. We said those are irrelevant as the AI doesn't use them enough to make a difference. I don't think there's a weapon combination in what's left over that beats shields but not armor. If you know of one please do share.
If you wish I'm confident I can produce AI designs that beat armor better than they beat shields.

 

[Torakka]

I first tried to get a benchmark for what amount of minerals were needed to defeat the Doomfleet. I tried a fleet of 220 "422"-Battleships since these were supposed to beat the Doomfleet according to Torakka.
For me, they lost horribly. The ranged engagements had 85 and 87 FE destroyer remaining, the point blank one 183 and 173. I'm not sure why, but in hindsight I notice that I've put the weapons into different slots. Though I doubt that such a thing can make this kind of difference.

Thanks for these tests! I did my testing with 1.8.0 and there might have been some changes regarding targeting or some other in-combat behaviour, although none were mentioned in the patch notes. (Except downgrading the damage of Flak Cannon.) Another possible difference is that your leader might have spawned with trait(s) that affect combat, but mostly I'd suspect that for some reason the bombers of DFoD just behaved more rationally in your tests.

About the armour versus shield thing:

1. One armour plate costs less minerals than same tier shield. Additionally, the armour plate does not need energy necessitating less reactors reducing the cost of the designs even further.
2. Tweaks in the weapon targeting in 1.8.0 should make all weapons better at focus firing, reducing the effectiveness of shield regen (as it is less likely that there will be breaks from being shot to regen the shields back to full).
3. The fallen empire ships in DFoD have weapon loadouts that tend to perform better against shields than armour (only decent armour penetration, but high shield damage or even shield ignore).
4. However, with sufficiently high number of repeatable techs, it is possible to reach the maximum 90% damage reduction from armour without any armour plates.

Also, @moyang ran some test that suggested Enigmatic Decoder being the ultimate aux slot item instead of capacitors.

 

[elitesix]

Has anyone tested whether the armor penetration % is actually based on defensive armor values instead of defensive armor percentage? By this I mean that you can stack armor - usually without using up your utility slots - and get to 90% reduction on battleships, and then % doesn't increase even if you add more armor. However the armor value displayed beside the armor percentage continues to increase if you add more armor.

Is it possible the armor penetration % is in fact based on the armor value instead of the armor %? If so, stacking armor values well past the value to get 90% could be very effective, by making armor penetration %s have less effect.

 

[Torakka]

Has anyone tested whether the armor penetration % is actually based on defensive armor values instead of defensive armor percentage?

Not sure if anyone has empirically tested it, but a dev (grekulf) confirmed here and here that the damage reduction percentage is used, not the absolute value of armour points. Therefore, stacking armour past the value required for 90% damage reduction is completely useless in every scenario.

 

[GuardianGI]

I don't know it has been mentioned in this thread, but strike craft repeatable techs have 10% bonus to fire rate and damage vs 5% of other weapons.

 

[moyang]

Not sure if anyone has empirically tested it, but a dev (grekulf) confirmed here and here that the damage reduction percentage is used, not the absolute value of armour points. Therefore, stacking armour past the value required for 90% damage reduction is completely useless in every scenario.

I tested it before and can confirm it's true. Any more armor after armor cap is wasted.

I don't know it has been mentioned in this thread, but strike craft repeatable techs have 10% bonus to fire rate and damage vs 5% of other weapons.

Problem is, their HP remains the same, and they get intercepted pretty fast.

 

[GloatingSwine]

Basically I tried to find a weapon combination players (and therefore regular AI empires) have access to that beats shields but is countered by armor.

The contention isn't that it "beats shields", AI trash doesn't beat anything, just that it is worse against armour and so if you use high armour designs you take less casualties and can beat larger opposing fleets.

Also, the AI has a group of relatively fixed layouts it uses. It will always use energy torpedo corvettes as soon as it unlocks them (no AP), it has two destroyer designs of which is is much more likely to use the one with all PD modules, six cruiser designs, and about five battleship designs one of which is super rare. Against those designs, armour is better.

As a reference, the energy preference AI's most common cruiser design once it has reached the end of the tech tree is 1 Large Plasma, 2 Medium Disruptor, 1 Medium Laser, and one Whirlwind Missile. That means that four out of its five weapons are strongly countered by armour. And none of its other designs are much better, the only one that is is 2x LPlas/2x M Disruptor (and even then the two medium slots are functionally dead against armour).

And that's the energy preference AI, the only one that really mounts strong anti-armour at all (albeit it does love disruptors to an unreasonable degree, midgame it's not uncommon to see it mount nothing but because it weights researching them quite highly, again GG if you have armour because they have no AP). If it was Kinetic Preference it would be loading up on trash like autocannons.

 

[GloatingSwine]

Thanks for these tests! I did my testing with 1.8.0 and there might have been some changes regarding targeting or some other in-combat behaviour, although none were mentioned in the patch notes. (Except downgrading the damage of Flak Cannon.) Another possible difference is that your leader might have spawned with trait(s) that affect combat, but mostly I'd suspect that for some reason the bombers of DFoD just behaved more rationally in your tests.

I also did this test at 220 and they didn't win.

It very probably is how bombers behave, the angle the two fleets meet at is going to make a difference how long their big silly detour lasts and how long it takes for them to start having an effect.

Also, whilst it is possible to hit 90% without any armour plates it's quite difficult, you need rather a lot of ranks of +armour to do so even on battleships.