AI War Beta 7.025-7.028 "Extermination Protocol MkI" Released!

[RockyBst]

Fixed a bug where lightning and armored warheads could not be given direct attack orders because... well, they didn't have a conventional attack anymore, after the conversion to the "explosion after death" mechanic that allows their new aoe logic.
This also was preventing Neinzul Rocketry Corps warheads from actually making attack runs on a target on the same planet. They'd dutifully get to the target planet and then just sit there and provide remarkably cheesy opportunities to enterprising and shameless players.

You mean my spirecraft translocators can't herd Neinzul rocketry warheads into position over hostile wormhole exits anymore? Bah, I only managed to wipe out about two reprisal waves with that trick

[keith.lamothe]

You mean my spirecraft translocators can't herd Neinzul rocketry warheads into position over hostile wormhole exits anymore? Bah, I only managed to wipe out about two reprisal waves with that trick

Don't worry, the Neinzul Rockefort cheese will be remembered, along with Jumpship Stallton

[Toranth]

Ok, so I had some time today to play some more, and need to adjust my earlier comment about the Champion Alt-Response.  I was playing a Progress 5, Response 5 game, and had grabbed a few ARS/FacIVs to get Cruisers unlocked for my 8 Champions.  The continuous response was good (1 spawn every 2.5-ish minutes) - it was enough that I couldn't afford to ignore it for hours on end, but not so high that it became invincible or was able to overwhelm my fortress Homeworld after just 45 minutes.  As long as I spent 5-10 minutes cleaning up the threat every hour or two, it was under control.  If the Response Intensity had been 10, I think

The Nemesis Exowave was still crazy, though.  I screwed up and had popped a Nuclear Eye unintentionally, resulting in an early AIP jump, so the Exowave launched at hour 4, and it consisted of 977,676.96 Strength - 197 Nemesis frigates in each batch.
They killed me, of course.  A bunch of HBCs, core Turrets, and 2 Mod Forts were unable to kill even one of them.  Not surprising, considering that blob of ships had more than 8 billion HP in shields that needed to be destroyed before the ships themselves could be hurt.  Without a Planetary Armor Inhibitor, this was basically impossible.
I'm going to try to do some experiments and see how many Nemesis frigates are survivable with what level of defenses, and I'll come back with my opinions on that later.


On the hacking changes, whoa!  It's suddenly getting dangerous to hack again.  I played a quick game where I started by hacking two Fabricators (turret controllers), unattended.  One covered, one unattended.  Between them, they produced about 800 threat (a fair amount of it was escaping from my fleet... was it supposed to do that?).  It also cost me most of my Mk I-II fleet.  The switch from zombies to Threat also makes them much more dangerous.  A group of Infiltrators paired up with some Mk III AI Raids to pick off one of my ARS outposts, much to my surprise.

Then, a few hours later, I did an 80 point SuperTerminal run.  When it started, I'd only spent the 75 HaP from the previous 2 fab hacks, had 150 HaP available, and spawns were about 100 units in size.  That's about 3 times larger (and a mark or two higher) than they'd been in the past.  A good upgrade.  If I hadn't beachheaded the Terminal system, I would have had to abort much sooner.
At the end, when I quit after dropping 80 AIP, the individual spawns were 300-350 units in size, and usually Mk V (with some Mk IV sometimes).
The final spawn, originally 150,000 Strength, was about 15,000 Mk V ships.  It tore my defenses apart, but between them and my fleet (and a careful Warhead) only about 2000 survived to become real threat.  While the final spawn was still way more powerful than the immediately preceding base pulse (about 50 times), the balance was closer.  It was still about the same size as the 7.025 final pulse... but I was better prepared this time.


I also encountered what I think was a bug where the 'Seconds of vigorous response' counter reset to zero after a reload, but I was not able to consistently reproduce it.

[keith.lamothe]

The Nemesis Exowave was still crazy, though.  I screwed up and had popped a Nuclear Eye unintentionally, resulting in an early AIP jump, so the Exowave launched at hour 4, and it consisted of 977,676.96 Strength - 197 Nemesis frigates in each batch.

Ouch.  Yea, that's not so much an attack as a race to see which one can land the last hit.

You mentioned a feasibility analysis of defending against those beasts and that would be very useful (it may help set their costs better).  Also, do you have the logs from the nemesis exo?  I think it puts it in LogicLog_MinorFaction_AltNemesis.txt .  I just want to make sure it's not counting some factor way more than intended, etc.  1 million strength doesn't come easy.


Glad to hear the hacking response is more interesting in general.

The final spawn, originally 150,000 Strength, was about 15,000 Mk V ships.  It tore my defenses apart, but between them and my fleet (and a careful Warhead) only about 2000 survived to become real threat.  While the final spawn was still way more powerful than the immediately preceding base pulse (about 50 times), the balance was closer.  It was still about the same size as the 7.025 final pulse... but I was better prepared this time.

So it sounds like it accomplished the goal of being interesting, though perhaps a bit too harsh at that relatively early stage in the game?

You mentioned some fairly sizeable per-pulse spawns.  Did they at least rise to the level of being non-boring?

I also encountered what I think was a bug where the 'Seconds of vigorous response' counter reset to zero after a reload, but I was not able to consistently reproduce it.

Hmm, not sure why it would do that.  Was it loading from an earlier version or from a 7.026 save?  Anyway, let me know if you see it again.

[Toranth]

The Nemesis Exowave was still crazy, though.  I screwed up and had popped a Nuclear Eye unintentionally, resulting in an early AIP jump, so the Exowave launched at hour 4, and it consisted of 977,676.96 Strength - 197 Nemesis frigates in each batch.

Ouch.  Yea, that's not so much an attack as a race to see which one can land the last hit.

You mentioned a feasibility analysis of defending against those beasts and that would be very useful (it may help set their costs better).  Also, do you have the logs from the nemesis exo?  I think it puts it in LogicLog_MinorFaction_AltNemesis.txt .  I just want to make sure it's not counting some factor way more than intended, etc.  1 million strength doesn't come easy.

  I'll attach the AltNemesis logfile from that game, and the hacking logs.

Glad to hear the hacking response is more interesting in general.

Yeah, the increased response is very noticeable.  In that game, I didn't spend my last 100 or so HaP, in part because I didn't need to, but also because the response may have taken more effort to handle than the additional fab or K was worth.  There's no longer a risk-free "Hack all the things!" ability, as long as your HaP is positive, that's for sure.

The final spawn, originally 150,000 Strength, was about 15,000 Mk V ships.  It tore my defenses apart, but between them and my fleet (and a careful Warhead) only about 2000 survived to become real threat.  While the final spawn was still way more powerful than the immediately preceding base pulse (about 50 times), the balance was closer.  It was still about the same size as the 7.025 final pulse... but I was better prepared this time.

So it sounds like it accomplished the goal of being interesting, though perhaps a bit too harsh at that relatively early stage in the game?

You mentioned some fairly sizeable per-pulse spawns.  Did they at least rise to the level of being non-boring?

I've attached the hacking logs from that game; you can see how big the spawns were.  I had beachheaded the SuperTerminal world, expecting a fierce response.  While the spawns themselves were unable to match the turrets and fortresses, they would have shredded my fleet if it had been alone.  The last test run I did (7.025), the end spawns were base strength 2200, with 1 wild-roll.  This time, they were base strength 3500 with 3 wild rolls.  Including the wild-rolls, I think the average spawn was 3-5 times more powerful, higher mark, and frequently triggering faster as well.

Beachheading preparation is now much more important that it's been since setting up for the 10,000 point AIP reductions back when the AIP floor was capped.

Although... the response I experienced may have been weaker than it should have been.  Looking in the logs, I see that the "AddStrength_PreRollStrengthTimesX" wild-rolls aren't doing anything for the SuperTerminal (no change in strength).

[Fleet Unity]

On the Champion Nemesis thing I tried 8 champions on normal with the nebula missions as well and both AI Spawned over 300 shadow frigates after four or so hours about 100+ each and no matter what I tried even with super fortress and modular ones there was no way to defeat them all and I tried it many different ways.
Edit: This was on difficulty 7 as well with two or so players and only one Home World I can't remember if it was three or two players and were disabled at the time.
Edit:2 I mean the Exo Wave is what I could not seem to defeat not the random wandering ones.
Edit:3 Also I think at the time our champions were only 8 Frigates I deleted the save file and restarted that game.

[keith.lamothe]

@Toranth: Haven't had time to look at all of it, gotta get some sleep, but one thing I wanted to clarify is that the nemesis exo you ran into wasn't actually a million strength, it was a million exo points.  The confusion is my fault as the logging calls the variable adjustedStrength, but:

adjustedStrength *= this.GetTotalResponseMultiplier( intensity ) = 977676.96
strengthPerChampion = (human shadow frigate).GetCostForAIWave() = 300
strengthPerReferenceUnit = (mkI hunter killer).GetCostForAIWave() = 3960
exoCostPerRefernceUnit = (mkI hunter killer).GetReferenceCostForTier() = 32768
theoreticalExoCostPerChampion = exoCostPerRefernceUnit * ( strengthPerChampion / strengthPerReferenceUnit ) = 2480
championsPerExo = ( ( adjustedStrength / theoreticalExoCostPerChampion ) / 2 ).IntValue = 197

Indicates a total strength of 197 * 300 = 59,100 strength.  And that 300 is the "special cost" of the shadow frigate rather than the actual strength value, that was concocted for those Heroic waves to account for the modules, etc.

Anyway, not saying it was any less of a steam roller, but I figured it was worth clarifying the metrics being used.

[Kahuna]

Btw. Shouldn't tech level be V when playing vs 2 The Core AIs? It's II at the moment.

[Coppermantis]

I think that the thing with The Core is that it just doesn't listen to tech level, even though the scale is still ticking as usual. It might be a bug though, but all the ships seem to be Mk. V as they ought to be.

[Toranth]

@Toranth: Haven't had time to look at all of it, gotta get some sleep, but one thing I wanted to clarify is that the nemesis exo you ran into wasn't actually a million strength, it was a million exo points.  The confusion is my fault as the logging calls the variable adjustedStrength, but:  <snip>

Ah, so it wasn't a million strength, it only felt like a million strength.  Got it.   

Looks like Nemesis frigates are listed at 192 Strength for actual gameplay, right?  So this Exo was combined roughly 77,000 strength actual, and 118,000 theoretical.
Everything a human can normally build on one planet measures about 85,000 strength.  That's a lot closer than I was expecting.

[keith.lamothe]

Ah, so it wasn't a million strength, it only felt like a million strength.  Got it.   

When the rocks fall and everyone dies, the exact tonnage remains somewhat vague.

Looks like Nemesis frigates are listed at 192 Strength for actual gameplay, right?  So this Exo was combined roughly 77,000 strength actual, and 118,000 theoretical.
Everything a human can normally build on one planet measures about 85,000 strength.  That's a lot closer than I was expecting.

In theory, but the 192 doesn't count the modules.

With:
1) The module firepower (which is widely varied due to randomized picking)
2) The very high speed of champs in general
3) And, possibly most importantly, the tendency of the shields to provide overlapping coverage (as you noted)

The result is a pain much larger than the sum of its parts.

Which I think is great, but it does mean that the overall size of the hammer has to be adjusted, because while in theory a single nemesis frigate might be as dangerous as a cap-and-a-half of mkI fleet ships (taking the 300-strength figure), in practice 10 of them are much more dangerous than 15 caps of MkI fleet ships, and 100 of them are waaaay more dangerous than 150 caps of MkI fleet ships (even though that would be akin to a 7,500-fighter wave on low caps).

[keith.lamothe]

Update: 7.027 just to do a quick balance iteration on the champion nemesis exos.


@Toranth: So, breaking down the math, the multipliers that led to your demise from the nemesis exos were:
Multiple champions, * 3.34  (it's 1, plus 0.334 for every champ past the first)
Cruiser-level champions, * 1.67 (would have been *2 if battleships)
Difficulty 9, * 4.5 (difficulty 7 is 1)
Nemesis Intensity 5, * 1.5 (intensity 4 is 1)

So ultimately this does look largely provoked:

- 8 champs is pretty "all in"; particularly once these hit BB size and you've got a lot of levels and unlocks (not hard to do on intensity 5) I imagine you can steamroll most AI worlds with no other losses.  The homeworld assaults would presumably require assistance, but you'll have a good bit of it from the splinter faction fleets, etc.  In other words, if anything's actually going to stop your brutal conquest of the galaxy, it would be these exos.

- Diff 9 is supposed to hurt pretty bad


That said, I think I'm just going to remove the highestChampMark factor entirely, as it does feel a bit treadmill-y and clearly the math is just fine without it.

For now I'll also try knocking the multi-champ factor down from 0.334 per extra champion to 0.2.  It's tricky to balance as each additional champ does more to increase your offensive capacity than it does for your defenses.  Though champs can certainly pull off some striking cheese with the deployable shields and heavy pausing.

[Toranth]

Well, the experimental results:
Only the Trader can save you.


This time, as part of getting more stuff to be strong, I ended up going all the way to Battleships accidentally.  So, the response was about 236 Nemesis frigates from each AI.
My homeworld was about as optimally set up for this sort of defense as I could figure out, with one entry wormhole the Exowave could come through, and with a distance of about 65,000 between that wormhole and my Home Command Station.

The map also conspired to help, as there was a Grav Drill between one AI homeworld and my Homeworld, so one of the two batches arrived more than 2 minutes after the other.

I played the game normally for the first 4.5 hours, with no cheating, just planning ahead.
I built 4 Mod Forts, all 5 Mk I Fortresses, a huge selecting of mark I-III Grav Turrets, mines, Mk I-III turrets, Mk V Sniper and Spider, all supported by a Mil III Command Station, of course.

With Trader goodies:
With Armor Inhibitor, Radar II, 2 OMDs - Survive.
Without Armor Inhibitor - Die.

Then I did some resource cheating and built the SuperFortress, too.
With SuperFortress and Radar II - Survive.
With SuperFortress, 2 OMDs - Die.
With Armor Inhibitor and SuperFortress - Survive.
With Armor Inhibitor, SuperFortress, and Radar II - 'Easy' win.

Then I did some more resource cheating, and built all 25 Mod Forts plus all the Mk I-II normal Fortresses.
With no goodies - Die to the first group of 236.
With only 2 OMDs - Die.
With 2 OMDs and Radar II - Survive.
With 2 OMDs and Armor Inhibitor - Survive.


Then I did some more cheating, and made it so the second group arrive within 30 seconds of the first.  This went less well for Humanity.
With all the Fortresses and stuff, no Trader Goodies - Die.
With all Fortresses, etc, and Armor Inhibitor - Survive.
With all Fortresses, etc, and SuperFortress - barely survive.  My Home Command Station took damage every time I ran this test.
With just the original defenses (Mk I forts) and Armor Inhibitor, SuperFortress, Radar II, and 2 OMDs - Lost 50% of the runs.  Home Command Station FF network took lots of damage every time.
With just the original defenses (Mk I forts) and missing even one of (Armor Inhibitor, SuperFortress, or Radar II) - Die.


So then I did some additional cheating, killing off some of the Nemesis as the were inbound, to control group size.
Using No Trader Goodies:
Two groups of 100 arriving simultaneously - Die.
Two groups of 100 arriving 2 minutes apart - Survive, barely.
Two groups of 150 arriving 2 minutes apart - Die.


Basically, with strong homeworld defenses and a HUGE distance (75% of the entire system) you have enough time to kill about 150 Nemesis Frigates before they kill your Home Command Station.  Trader goodies can increase this:  Radar II makes the Nemesis get a little closer to each Grav Turret before being able to kill them, keeping them slowed longer.  The Armor Inhibitor basically does 20,000,000 damage to each Nemesis when it enters the system.  The OMDs kill one very 10 seconds, and in this case accounted for 10-ish kills each.
At short range, where the entrance wormhole was only 30,000 distance away, a group of 100 Nemesis is able to destroy the Home Command Station before you can kill them.  Again, Trader Goodies can help, Armor Inhibitor in particular, but it's a race of their firepower against your Home FF network, not your overall defense strength.

I would suggest that response intensity 10 be aimed at producing a total of 150-200 Nemesis Frigates for 8 Battleship Champions.  That's enough that it would be very hard to survive without Trader Goodies.  A multi-system defense in depth is required at that point.
Intensity 5 with 8 BB Champions should be a total of no more than 100.
Intensity 5 with 1 BB Champion should total around 50.  In that case, you've already unlocked the stuff that's actually going to help defend yourself - the Mod Forts and Nebula Ships outweigh the Champion itself.

So ultimately this does look largely provoked:

Yeah, it is pretty much self-inflicted under the current settings.  I just don't think it should be instant-death at time X.  Even Astro Trains Intensity 10 isn't that sharp a cliff.

That said, I think I'm just going to remove the highestChampMark factor entirely, as it does feel a bit treadmill-y and clearly the math is just fine without it.

Actually, my first instinct would be to do the opposite - Up the Hull Size factor, and lower the per-Champion factor.  However, your numbers would result in, hmm...
26,000 * 2.4 (count) * 1 (no size factor) * 4.5 (Diff 9) * 1.5 (Intensity 5) = 421,200;   421,200/2480 = 170 total Nemesis frigates in Exowave 1.  212 in wave 2.
26,000 * 2.4 (count) * 4.5 (Diff 9) * 6 (Intensity 10) = 1,684,800 = 680 total Nemesis.
26,000 * 1 (count) * 4.5 (Diff 9) * 6 (Intensity 10) = 702,000 = 284 total Nemesis.
26,000 * 1 (count) * 1 (Diff 7) * 6 (Intensity 10) = 64 total Nemesis.
26,000 * 1 (count) * 1 (Diff 7) * 1.5 (Intensity 5) = 16 total Nemesis.




Edit: Attached the image of my HW defenses:

[RockyBst]

I just don't think it should be instant-death at time X.  Even Astro Trains Intensity 10 isn't that sharp a cliff.

Time to first death at 10/10 astro trains would be 14 minutes ... that's when the first nuke trains rolls into your homeworld

[keith.lamothe]

Time to first death at 10/10 astro trains would be 14 minutes ... that's when the first nuke trains rolls into your homeworld

Special Delivery!


@Toranth, thanks much for the very thorough testing  I agree with targeting, say, a total of 200 nemesis champs on the first nemesis exo on Diff 9 with 8 champs on alt-nemesis-10.

I can consider reinstating the hull-size-based multiplier, but ultimately it doesn't feel right that the response gets heavier simply because you've improved the weapon in-game (edit: FS is an exception to this, as that's actually a general escalation of the war; and Showdown Devices scales like that to account for actually being a victory condition rather than just a weapon, though it certainly needs more work).  I prefer to not "punish" what amounts to playing better.  Also, given how long those exos take to charge it's likely the player will have at least Cruiser by the first exo, and BB by the second at the latest.  I'd prefer it if getting to BB before the first exo was simply a good play, not an additional +33% or whatever to the hammer.  Does that make sense?

One thing I need to do is make the AI's champs vulnerable to aoe (and thus warheads) again, just hadn't gotten around to dealing with that report from you.  That would at least provide an "out" for a player caught moderately unprepared for the ferocity of one of these.

Anyway, as to the math:

26,000 * 2.4 (count) * 4.5 (Diff 9) * 6 (Intensity 10) = 1,684,800 = 680 total Nemesis.

Yea, ow.  So I'm thinking to get this down to the target of 200:
- change the per-extra-champ bonus from +20% down to about +14.3%, meaning that having a full set of 8 is an effective doubling of the exo size.
- drop the multiplier for Intensity-10 down from 6 => about 2.1, scaling down to about 1.05 at Intensity 5, and probably making intensities 1-4 get 0.2, 0.4, 0.6, and 0.8

That would result in:

26,000 * 2 (count) * 4.5 (Diff 9) *  2.1 (Intensity 10) = 491,400; @ 2480 per-each that's ~= 198 nemesis per exo... wait, you said 680 total for the current math, did you mean 340 in each exo and thus 680 total?  Or 680 in each exo, for a total spawn of 1,360 (good grief) ?