Fleet minmaxing/game balance

[genrtul]

So, I've modded the game ever so slightly to achieve a balance point I like.

My expectation for balance is basically this.

Corvettes strongly counter battleships, weakly counter cruisers.
Destroyers strongly counter corvettes, weakly counter battleships.
Cruisers strongly counter destroyers.
Battleships strongly counter cruisers.

I've run a large number of battle tests featuring end game technology (no shipyards, admirals, anomaly or resource bonuses). Weapons consisting primarily of tachyon lances and material disintegrators at a 20k mineral price point. Meaning every fleet costs as near to 20k as possible.

I ran these tests with a number of different modifications to the base ship size evasion.

Vanilla
Corvettes have a 0% modifier. Destroyers have -25%. Cruisers have -50%. Battleships have -75%.

I ran a large series at different numbers but these are the two test groups I find the most interesting.

Test 1
Corvettes at -15%. Destroyers at -30%. Nothing else changed.

Test 2
Corvettes at -20%. Destroyers at -30%. Nothing else changed.

Because these tests ONLY used thrusters and CPU's to achieve their evasion, the top end point of evasion is 60% for corvettes in the vanilla test. The vanilla test however was very shocking. At the base evasion of 60%, destroyers absolutely destroyed them. Not a competitive battle. In fact, even when I simulated evasion to 80% for corvettes, destroyers still beat them. The problem with corvettes, is simply that they can achieve 80, 90, even 100% evasion rates.

Let's talk about the two test groups. Referring to that point of balance I'd like to achieve.

Test 1, Destroyers handle corvettes easily at -15% and -30% respectively. The curious issue is battleships versus destroyers. At this point of evasion, destroyers at 42% evasion, this is a coin toss. Granted, these tests were done with Destroyers designed to fight corvettes, and battleships NOT running an aura. This actually makes me pretty happy with this change. On the flip side of this, cruisers were handling the destroyers quite easily. Lastly, corvettes did to battleships what I think they should, they swarm them and defeat them quite easily.

Test 2, again, the Destroyers obliterate the corvettes easily with them at -20% and -35% respectively. The interesting point of balance again is the destroyer versus battleship balance point. Here, with the destroyers at -35% evasion sitting at 39%, the battleships reliably won, taking over 50% losses in terms of ships, with most of the remaining battleships at or below 50% hp. Cruisers kill destroyers even more effectively.

So here is the sticking point. Test 1 is a better point of balance for destroyers and battleships. With battleships running auras, they would tip the balance and win with losses reliably to the destroyers. With test 2 figures, I feel destroyers will get rolled even harder. However, the big problem is, all the little addons you can get to boost your evasion up even higher with corvettes, to the point where you can actually achieve 100% or higher evasion with vanilla corvettes. Test 1 is at -15% which means that if you pushed your corvettes to 100% evasion, they would be at 85% evasion in game, and at this point destroyers PROBABLY will no longer directly counter them. With test 2 numbers, they would be at 80% and destroyers would counter them (I tested corvettes at 80% vs destroyers and they lost, albeit closely). Once corvettes hit 80-90% they are hitting the point where they are no longer countered by destroyers.

Right now, I am actually considering pushing corvettes to -25% and destroyers to -35%. Destroyers will obliterate corvettes, cruisers will obliterate destroyers, and corvettes will still obliterate cruisers and battleships. The -25% on corvettes will prevent them from hitting the god-mode tier where they start beating destroyers.

I'll let you guys know how it goes.

So my takeaway from this is that we've been severely undervaluing range. I guess you must have observed a lot of damage being done by the tachyon lances before the corvettes got in range? This now makes destroyers the best late-game hull just because they can mount lances, presumably. Or maybe just battleships for the auras.

 

[Kaldrakos]

i have not, but you just said yourself crystal forged plates were good, so why are you still debating my opinion on this?

Crystal forged plate is a random tech which is very rare to get , since you can only get it from elite crystal debris .

 

[Pahvimato]

So my takeaway from this is that we've been severely undervaluing range. I guess you must have observed a lot of damage being done by the tachyon lances before the corvettes got in range? This now makes destroyers the best late-game hull just because they can mount lances, presumably. Or maybe just battleships for the auras.

Tachyon lances are kinda special because they hit instantly. This eliminates part of the problem with the horrible target acquiring. Corvettes still beat them up late game - your evasion can rise to 100%.

I'd imagine the strongest fleet is so many corvettes that you no longer can send in any more and the rest of your fleet just sits there, and then a few dozen as reinforcements. After that start stacking BB auras. Haven't tested, but should work pretty well (in the plz nerf, pl0x sense well).

 

[Invader_Canuck]

So my takeaway from this is that we've been severely undervaluing range. I guess you must have observed a lot of damage being done by the tachyon lances before the corvettes got in range? This now makes destroyers the best late-game hull just because they can mount lances, presumably. Or maybe just battleships for the auras.

Well, it's not so simple, because cruisers obliterate destroyers as hard as destroyers obliterated corvettes before the corvettes entered broken evasion levels (80%+).

So if you spam destroyers, and someone counters with cruisers, you're up the proverbial shit creek without a proverbial shit paddle. However, if you bring some battleships to the party, things might swing back in your favor as battleships counter cruisers about as hard as cruisers counter destroyers! However! If the guy you are fighting brings a ton of corvettes with his cruisers, now we're looking at two hard countering stacks. His corvettes counter your battleships which counter his cruisers which counter your destroyers.

Basically, as long as we get a fix to evasion which clamps the maximum evasion at or around 70% I would say, it's pretty fun balance. Corvettes are very cost effective vs cruisers and battleships (less so against cruisers) but are SLAUGHTERED (I can't state this emphatically enough) by destroyers. In my tests for example, the destroyers with 42 ships would beat 80 corvettes consistently with 35-36 ships left at a base point of around 60% evasion. As the evasion climbed those losses mounted.

 

[Ardent]

http://w, but ww.dod.mil/pubs/foi/Reading_Room/Joint_Staff/12-F-0344-Millennium-Challenge-2002-Experiment-Report.pdf

There was absolutely cheating involved. Here's a list of things Van Riper did that were physically impossible or nearly so:

1. In order to avoid comm intercepts and and jamming by Blue Force, Van Riper had all his communications delivered via bike messenger. The computer simulation had no setting for such a delivery method, so all his messages were considered to be instant and yet completely impossible for Blue Force to intercept. Unless someone has invented motorcycles that drive on water and move at the speed of light, this was obviously cheating.

2. Van Riper mounted large anti-ship missiles on small yachts that had no ability to fire them. Theoretically those boats could fit the missiles, but only with very extensive modifications that Red Force didn't have access to. It's not until over a decade later did the someone produced a weapon that would have worked for Van Riper's purposes. Plus a rival nation buying thousands of Russian klubs in the container configuration would be immediately be noticed and the commander of Blue Force would have taken that into account.

3. Van Riper exploited the computer software's constraints by labeling all his ships as civilian pleasure craft. This meant the computer ignored them even when hundreds of them were rapidly converging on Blue Force. In real life this obviously would not be workable since any human with a brain will rapidly recognize such an event as an attack. Computers however, are not so smart. Moreover, the entire scenario takes place in a situation where Blue Force had just demanded the surrender of Red Force(cough Iran cough) and steamed in to attack. There is zero chance that in a period such high tension that an American admiral would let even one yacht approach his fleet, let alone hundreds.

Could you substantiate any of those claims?

As someone who has participated in war gaming (and has written white papers on milcap) I'm aware that what Van Riper did would undermine the intent of the war game, but this isn't like he pulled a Kobayashi Maru.

In fact, as has been pointed out by multiple sources the actual cheating was undertaken by General Pace in defense of Secretary Rumsfeld and the Bush administration's Middle East intentions vis-a-vis Iraq and Iran:

http://www.theguardian.com/world/2002/sep/06/usa.iraq

https://web.archive.org/web/20060504005348/http://www.military.com/opinion/0,15202,95496,00.html

 

[kmaj]

Quickiest/easiest fix to corvette spam: make Point Defense large only.

 

[LucidFugue]

Quickiest/easiest fix to corvette spam: make Point Defense large only.

That only means that corvette spam forces everyone to go missile tech. I'd sooner cut corvettes to 2 hardpoints so the choice to make one of them pd had a bigger impact on damage output.

Super high evasion is clearly the source of corvette strength. I'm against caps where it can be avoided, so I'd want to investigate buffing small weapon accuracy, while buffing medium and large weapon damage and range to compensate. Probably also putting an accuracy malus on the defensive computer, and evasion malus on the offensive computer.

That means you can outfit destroyers with many smaller hardpoints to counter corvettes, or with fewer larger hardpoints to counter cruisers/other destroyers. So long as the accuracy buffs can prevent evasion stacking reaching those thresholds, then the balance starts to fall back towards weapon counters.

 

[Tale]

Quickiest/easiest fix to corvette spam: make Point Defense large only.

Quick and easy fixes often just turn problems into meta strategy.

 

[Blobbox]

That only means that corvette spam forces everyone to go missile tech. I'd sooner cut corvettes to 2 hardpoints so the choice to make one of them pd had a bigger impact on damage output.

Super high evasion is clearly the source of corvette strength. I'm against caps where it can be avoided, so I'd want to investigate buffing small weapon accuracy, while buffing medium and large weapon damage and range to compensate. Probably also putting an accuracy malus on the defensive computer, and evasion malus on the offensive computer.

That means you can outfit destroyers with many smaller hardpoints to counter corvettes, or with fewer larger hardpoints to counter cruisers/other destroyers. So long as the accuracy buffs can prevent evasion stacking reaching those thresholds, then the balance starts to fall back towards weapon counters.

I think the way to go is to increase the raw damage of bigger weapons, while at the same time making smaller weapons ignore a percentage of their target's evasion chance.

For example have small/medium weapons reduce evasion by 75/50%. This would cap evasion effectiveness against small weapons at 25%, which makes corvette eHP a lot less scary.

Of course, ships would have to choose a target according to their weapon loadout, but this would make bigger weapons better against bigger ships and smaller weapons more effective against smaller ships.

 

[LucidFugue]

I think the way to go is to increase the raw damage of bigger weapons, while at the same time making smaller weapons ignore a percentage of their target's evasion chance.

For example have small/medium weapons reduce evasion by 75/50%. This would cap evasion effectiveness against small weapons at 25%, which makes corvette eHP a lot less scary.

If corvette evasion vs small was capped at 25% it would invalidate most evasion bonuses. They natively get at least that much. Hence why I'd boost their accuracy instead - it has the same effect (increasing chance to hit) but keeps the range of values nice and broad. That way accuracy is a dimension to consider in fleet design. If you choose highly accurate beam weapons to counter evasion stacking corvettes, I can build aggressive kinetic corvettes who are gonna get hit no matter what, but put out a. lot of hurt. If evasion is just a flat 25% irrespective of leader, battle computer or weapon type, what's the point?

Of course, ships would have to choose a target according to their weapon loadout, but this would make bigger weapons better against bigger ships and smaller weapons more effective against smaller ships.

This is really important and the other weak spot in the AI. I'd love to see the ability to select AI preferences when designing ships - stuff like target preference and range preference. That way you can design your large ships as multiple small hardpoint corvettes killers who close distance, or giant flying guns that sit back at range and blast the enemy capital ships.

 

[Nephrahim]

Could you substantiate any of those claims?

As someone who has participated in war gaming (and has written white papers on milcap) I'm aware that what Van Riper did would undermine the intent of the war game, but this isn't like he pulled a Kobayashi Maru.

In fact, as has been pointed out by multiple sources the actual cheating was undertaken by General Pace in defense of Secretary Rumsfeld and the Bush administration's Middle East intentions vis-a-vis Iraq and Iran:

http://www.theguardian.com/world/2002/sep/06/usa.iraq

https://web.archive.org/web/20060504005348/http://www.military.com/opinion/0,15202,95496,00.html

That's not really what happened though.

Like people say they "Cheated" when they re-floated the boats but what else would they have done? Ended it after one day and gone home? They had spent 200 million dollars on the exercise, of course they weren't just going to declare red team the winner and end it. They had other parts to test.

 

[Gamesguy]

OK I just did dozens of tests. All tests were conducted with level 4 technology and equal mineral cost. The corvette build was 2 small x-ray laser, 1x barrier PD, 3x small anti-matter reactor, 2x small crystal forged plate, hyperdrive 2, defensive combat computer advanced, impulse thrusters, and lvl 4 sensors.

My first test was a mirror match, same corvettes, only instead of defensive combat computer I tried aggressive and sentient. Both were utterly destroyed by the corvettes with defensive computers. The same holds true for every ship class and every weapon(except missiles, but missiles were worthless against PD). For some absurd reason defensive computers are vastly better even on battleships. Then I tried a variety of ship classes and load outs against the evasive corvette horde. All builds emphasized hp/armor tanking(I tried shields, they were horrible). The best weapon was the large autocannon, the worst were missiles(they all got shot down by PD) followed by lasers/railguns.

I tried a variety of weapon sizes, and contrary to what the weapon descriptions say and all logic, larger weapons are flat out better. You can test this yourself. Take a destroyer and fit a large weapon and nothing else, now have it fight against a DD with 4 small guns and nothing else. The destroyer with the large gun will win. This gets even more lopsided with bigger ships.

Mixed fleets...actually work. Substituting half the destroyers/cruisers/battleships with evasion tanked corvettes improved the outcome significantly in all scenarios.

Here's some data.

20 cruisers vs 85 evasion tanked corvettes(same mineral cost), outcome was cruisers won with 40% HP remaining and 16/20 ships remaining.

50 destroyers vs 100 corvettes, DDs won with 44.1% hp and 48/50 remaining. Note that the destroyers would have done better if I had given them an extra small gun over a PD. Two PD per destroyer really was overkill.

10 BB vs 100 corvettes, corvettes won with 91% hp and 24/100 remaining:

10 BBs vs 100 corvettes(the BBs actually cost the same as 125 corvettes), this time with the arc emitter. Battleships won with 11% hp and 7/10 remaining. A weird thing happened in this test, most of the corvettes were greatly slowed by the battleship subspace aura and trickled in slowly to die. This seems to happen when the battleship mounts a large only long range weapon like lances, kinetic artillery, and the arc emitter. It doesn't happen when battleships mount shorter ranged weapons like normal lasers. Retesting without the subspace aura didn't change anything, the corvettes still trickled in a few at a time and died. In fact the BBs performed better without the slow aura...winning the battle with 18% HP and 9/10 surviving.

With that in mind, I redid test #3, this time the battleship has no aura and more plates instead of generators. The BBs this time managed to eek out a victory with 4% hp remaining:

Some other tests:

20 cruisers vs 43 DDs, cruisers won with 14% hp and 9/20 remaining.
20 cruisers vs 10 BBs, BBs won with 27% hp and 9/10 remaining.
Same as above, except the BBs have sentient combat computers instead. BBs did much worse and cruisers won with 65% hp and 9/20 remaining.

With small and medium guns instead:

The same 20 cruisers as the first test, except with small and medium weapons vs 85 corvettes. Cruisers won with 28% hp and 17/20 remaining, this is significantly worse than with large guns.
The same 50 destroyers as the second test, except with small weapons and an extra armor plate(due to extra slot), and one PD swapped for a laser. Destroyers won with 33% hp and 48/50 remaining. Same as above.

I did a few other tests, here are my conclusions:

1. Fit defensive computers on everything, even battleships. This makes a huge difference.
2. Corvettes=BB, but BB>CA>DD>Corvette. Note that this is identical weapons, if you specialize, then BBs get a lot better. BBs with autocannons will eek out a win against the corvette swarm, where as corvettes need energy weapons or armor hurts them too much.
3. Larger weapons are better than smaller weapons, period.
4. Fighters are useless.
5. Missiles are useless against PD. Dedicating one slot to PD on a corvette, 2 on cruisers and DD, and 4 on battleships is sufficient to reduce missile damage by at least 90%. I tried all types of missiles, they all suck against PD.
6. Area of effect weapons like the arc emitter and the flak cannon are all worse than just fitting as many autocannons as you can.
7. Shields are strictly worse than HP tanking with hull plating.
8. Regen plating is somewhat worse than HP tanking on battleships, but not horrible.

Corvettes are not broken, it is Defensive Combat Computer(Advanced) that is broken. Also, large weapons> small weapons, the description lies when it says larger weapons have a hard time hitting small ships.

Ask if you want me to test something specific.

 

[Pahvimato]

*SNIP*

The fact that defensive computers are better than most choices is not really broken, but a small balance problem, while corvettes are. And L weapons and S weapons have the same DPS, the accuracy isn't that much smaller. Apart from lances/arc/other instant stuff, they tend to mostly aim at the same targets though. This is where their problem is - not in their accuracy in and of itself, but rather the horrible target acquiring which causes them to concentrate on single targets (which might work against BB, but is not very useful against DDC).

As Invader_Canuck pointed out, the problem with corvettes is that the 80% you can get easily is nowhere near the maximum. If I'm not mistaken, the corvettes you used in those tests only had 60% - I believe what I and the others here are asking for is for corvettes to be useful in the way you described. With some luck with anomalies, some users have even reported evasion rates of over 100%. This is, of course, not a hard thing to fix. In the stellaris dev diary there's already a note that corvettes are going to get nerfed, I assume that means no more dodging absolutely everything, but corvettes will retain their role as a screening ship for the bigger ones, like you seemed to hint at in your post. Removing their evasion completely would be silly.

 

[Ardent]

That's not really what happened though.

Like people say they "Cheated" when they re-floated the boats but what else would they have done? Ended it after one day and gone home? They had spent 200 million dollars on the exercise, of course they weren't just going to declare red team the winner and end it. They had other parts to test.

So the accusation that Van Riper cheated is being made but cannot be substantiated. The accusation that Pace cheated is clearly substantiated by the fact that the exercise continued.

I get why Pace cheated. I have spent 5 days cooling my heels at a desk with nothing to do because my command got blown up on day one of an exercise. That's a huge waste of time for everyone involved, and MC02 was comically over-sized to begin with. Which contributed to the conclusions it drew being bogus, from my perspective. That and it was initiated with a specific political outcome in mind, which also tends to damage findings.

I don't understand the efforts to slander (and libel) Van Riper, though.

 

[Blobbox]

If corvette evasion vs small was capped at 25% it would invalidate most evasion bonuses. They natively get at least that much. Hence why I'd boost their accuracy instead - it has the same effect (increasing chance to hit) but keeps the range of values nice and broad. That way accuracy is a dimension to consider in fleet design. If you choose highly accurate beam weapons to counter evasion stacking corvettes, I can build aggressive kinetic corvettes who are gonna get hit no matter what, but put out a. lot of hurt. If evasion is just a flat 25% irrespective of leader, battle computer or weapon type, what's the point?

Perhaps there was a misunderstanding, it would not be capped at 25%, but scale at 25% efficiency against small weapons.

Say, with my changes applied, there is a ship with A) 0% evasion and B) 100% evasion getting shot at by a large, a medium and a small weapon with 100% accuracy each.

A) the ship gets hit 100% of the time by all three weapons.
B) the ship gets hit 0% of the time by the large weapon, 50% of the time by the medium weapon and 75% of the time by the small weapon.

So a ship at 100% evasion would be impossible to hit for large weapons, but the accuracy of small weapons would only be reduced by 25%.

 

[FlyingPhoenix]

I was going to start a thread saying what Gamesguy said, without the test results, but yeah, it is clearly the evasion mechanic which is broken, rather than the strength of corvettes, and it would be nice if it was the evasion and accuracy mechanic which was fixed, rather than giving Corvettes a nerf.

And shouldn't this be the way it works? Cruisers and Destroyers should be anti corvette to some extent, but corvettes should be the counter to battleships. Otherwise battleships have no counter other than a larger battleship, which takes away from strategy.

 

[Pahvimato]

I was going to start a thread saying what Gamesguy said, without the test results, but yeah, it is clearly the evasion mechanic which is broken, rather than the strength of corvettes, and it would be nice if it was the evasion and accuracy mechanic which was fixed, rather than giving Corvettes a nerf.

And shouldn't this be the way it works? Cruisers and Destroyers should be anti corvette to some extent, but corvettes should be the counter to battleships. Otherwise battleships have no counter other than a larger battleship, which takes away from strategy.

Well, the best way to nerf corvettes would be to fix evasion, yes. Why it's specifically corvettes I'm complaining about is their ability to go beyond 100% evasion - something that shouldn't happen. After that's fixed, it would work just like that - I hope. As I noticed in my own tests, and Invader_Canuck noticed in his, the evasion where corvette spam is viable is around 80%. Why corvette's are too strong is that that can still be increased, if you know your minmax.

RIght now I think the best thing to do is wait for the Clarke patch, that should be live near the end of May, which isn't that far off. We are a few guys dicking around with the console, paradox has a QA department. Let's hope they found a good way to fix it, and if not, we can continue pointing out the flaws in evasion itself after that.

 

[Gamesguy]

The fact that defensive computers are better than most choices is not really broken, but a small balance problem,

It's quite a large balance problem actually.

while corvettes are. And L weapons and S weapons have the same DPS, the accuracy isn't that much smaller. Apart from lances/arc/other instant stuff, they tend to mostly aim at the same targets though. This is where their problem is - not in their accuracy in and of itself, but rather the horrible target acquiring which causes them to concentrate on single targets (which might work against BB, but is not very useful against DDC).

As Invader_Canuck pointed out, the problem with corvettes is that the 80% you can get easily is nowhere near the maximum. If I'm not mistaken, the corvettes you used in those tests only had 60% - I believe what I and the others here are asking for is for corvettes to be useful in the way you described. With some luck with anomalies, some users have even reported evasion rates of over 100%. This is, of course, not a hard thing to fix. In the stellaris dev diary there's already a note that corvettes are going to get nerfed, I assume that means no more dodging absolutely everything, but corvettes will retain their role as a screening ship for the bigger ones, like you seemed to hint at in your post. Removing their evasion completely would be silly.

Again, corvettes are not broken if they only break once they get passed 80% evasion. Evasion stacking is broken, thus the obvious solution is to make evasion subject to diminishing returns, much like armor is.

 

[genrtul]

OK I just did dozens of tests. All tests were conducted with level 4 technology and equal mineral cost. The corvette build was 2 small x-ray laser, 1x barrier PD, 3x small anti-matter reactor, 2x small crystal forged plate, hyperdrive 2, defensive combat computer advanced, impulse thrusters, and lvl 4 sensors.

My first test was a mirror match, same corvettes, only instead of defensive combat computer I tried aggressive and sentient. Both were utterly destroyed by the corvettes with defensive computers. The same holds true for every ship class and every weapon(except missiles, but missiles were worthless against PD). For some absurd reason defensive computers are vastly better even on battleships. Then I tried a variety of ship classes and load outs against the evasive corvette horde. All builds emphasized hp/armor tanking(I tried shields, they were horrible). The best weapon was the large autocannon, the worst were missiles(they all got shot down by PD) followed by lasers/railguns.

I tried a variety of weapon sizes, and contrary to what the weapon descriptions say and all logic, larger weapons are flat out better. You can test this yourself. Take a destroyer and fit a large weapon and nothing else, now have it fight against a DD with 4 small guns and nothing else. The destroyer with the large gun will win. This gets even more lopsided with bigger ships.

Mixed fleets...actually work. Substituting half the destroyers/cruisers/battleships with evasion tanked corvettes improved the outcome significantly in all scenarios.

Here's some data.

20 cruisers vs 85 evasion tanked corvettes(same mineral cost), outcome was cruisers won with 40% HP remaining and 16/20 ships remaining.

50 destroyers vs 100 corvettes, DDs won with 44.1% hp and 48/50 remaining. Note that the destroyers would have done better if I had given them an extra small gun over a PD. Two PD per destroyer really was overkill.

10 BB vs 100 corvettes, corvettes won with 91% hp and 24/100 remaining:

10 BBs vs 100 corvettes(the BBs actually cost the same as 125 corvettes), this time with the arc emitter. Battleships won with 11% hp and 7/10 remaining. A weird thing happened in this test, most of the corvettes were greatly slowed by the battleship subspace aura and trickled in slowly to die. This seems to happen when the battleship mounts a large only long range weapon like lances, kinetic artillery, and the arc emitter. It doesn't happen when battleships mount shorter ranged weapons like normal lasers. Retesting without the subspace aura didn't change anything, the corvettes still trickled in a few at a time and died. In fact the BBs performed better without the slow aura...winning the battle with 18% HP and 9/10 surviving.

With that in mind, I redid test #3, this time the battleship has no aura and more plates instead of generators. The BBs this time managed to eek out a victory with 4% hp remaining:

Some other tests:

20 cruisers vs 43 DDs, cruisers won with 14% hp and 9/20 remaining.
20 cruisers vs 10 BBs, BBs won with 27% hp and 9/10 remaining.
Same as above, except the BBs have sentient combat computers instead. BBs did much worse and cruisers won with 65% hp and 9/20 remaining.

With small and medium guns instead:

The same 20 cruisers as the first test, except with small and medium weapons vs 85 corvettes. Cruisers won with 28% hp and 17/20 remaining, this is significantly worse than with large guns.
The same 50 destroyers as the second test, except with small weapons and an extra armor plate(due to extra slot), and one PD swapped for a laser. Destroyers won with 33% hp and 48/50 remaining. Same as above.

I did a few other tests, here are my conclusions:

1. Fit defensive computers on everything, even battleships. This makes a huge difference.
2. Corvettes=BB, but BB>CA>DD>Corvette. Note that this is identical weapons, if you specialize, then BBs get a lot better. BBs with autocannons will eek out a win against the corvette swarm, where as corvettes need energy weapons or armor hurts them too much.
3. Larger weapons are better than smaller weapons, period.
4. Fighters are useless.
5. Missiles are useless against PD. Dedicating one slot to PD on a corvette, 2 on cruisers and DD, and 4 on battleships is sufficient to reduce missile damage by at least 90%. I tried all types of missiles, they all suck against PD.
6. Area of effect weapons like the arc emitter and the flak cannon are all worse than just fitting as many autocannons as you can.
7. Shields are strictly worse than HP tanking with hull plating.
8. Regen plating is somewhat worse than HP tanking on battleships, but not horrible.

Corvettes are not broken, it is Defensive Combat Computer(Advanced) that is broken. Also, large weapons> small weapons, the description lies when it says larger weapons have a hard time hitting small ships.

Ask if you want me to test something specific.

I've only been able to reproduce your large vs. small weapon tests when I didn't use any utility slots except energy on my ships. If I use a sensible amount of defenses the small weapons win, and they win solidly. How were your ships outfitted when doing these tests?

Large weapons do win sometimes in large fleet engagements, if they manage to quickly kill one of the earlier ships during the starting phase when they have range advantage. It's still a majority of small wins in ~10 tests.

I also confirmed that armor works as we thought. Armor increases EHP by Arm/60. A laser works in a strange way though. When it says it decreases armor by 50% that doesn't mean that it treats the enemy as if it had 50% less armor stat, but rather decreases the armor's damage shaving by 50%. So if fighting a ship with 60 armor that's 50% reduction -> 25% reduction which is like having 20 armor instead of the 33% reduction that you'd have with 30 armor. So lasers are even better at piercing armor than expected.

And I made a separate thread about this, but might as well post it here. Accuracy and evasion are additive, meaning a weapon with 76% hitrate against a 60 evasion ship doesn't hit 0.76*0.6=46% of the time as might be expected but just 0.76-0.6=16% of the time. This is a big part of why evasion is so good.

Also having a hard time reproducing getting anything to beat corvettes. I use stormfire autocannons and a PD with 3 HP plating in utility (best engines and defensive computer) and that build beats anything I make.

 

[Rabiator]

The problems are basically:

• Targeting choice: When all ships fire on the same opposing ship there is going to be MASSIVE overkill and longer range doesnt matter.
• Ship movement: The only option for ships seems to be "close in with the enemy". That is bad because staying at range would be the logical choice for a fleet armed with long range weapons.

This sums up to A PROBLEM OF ALGORITHM and the "non-strategy way" in which Paradox treats their combat. I think it has to be swapped for tactical combat with human input (plus a decent AI version) where ships are NOT treated as part of one giant formula but instead can choose one of a set of sub-programs to follow. I always cringe when I see my shiny big ships close up to the "melee" instead of staying out of range. So ...

• we need "checkboxes" for ship behaviour AND/OR
• tactical combat with human input.