Author Topic: 10 planets, diff 10 ... what could go wrong  (Read 28658 times)

Offline RockyBst

  • Full Member Mark III
  • ***
  • Posts: 219
Re: 10 planets, diff 10 ... what could go wrong
« Reply #60 on: April 15, 2014, 06:23:51 pm »
Don't worry, I found some ultimate Limburger to keep the threat off my back, as I'm sure you'll see in the attached save.

Looks like this one spawned in the regular amount of strategic reserve ships for me, so not quite sure what went wrong with the other one.

Offline Draco18s

  • Resident Velociraptor
  • Core Member Mark V
  • *****
  • Posts: 4,251
Re: 10 planets, diff 10 ... what could go wrong
« Reply #61 on: April 15, 2014, 06:55:15 pm »
12.31.10: Zupreape command goes down. The penetrator appears to be immune to the cloaker starship sitting next to it and is still getting hit (bug or intentional? I'm assuming intentional).

Intentional I believe.  After firing a penetrator is immune to cloaking, cloaking boosts, and cloaking superboosts.

See http://www.arcengames.com/mantisbt/view.php?id=2751

Offline Aklyon

  • Core Member
  • *****
  • Posts: 2,089
Re: 10 planets, diff 10 ... what could go wrong
« Reply #62 on: April 15, 2014, 07:20:11 pm »
Don't worry, I found some ultimate Limburger to keep the threat off my back, as I'm sure you'll see in the attached save.
Does it involve warheads?

Offline RockyBst

  • Full Member Mark III
  • ***
  • Posts: 219
Re: 10 planets, diff 10 ... what could go wrong
« Reply #63 on: April 15, 2014, 07:38:22 pm »
Well, the CPA did involve some warheads. But that is not the master cheese...


12.31.33: I can see the very instant when all the 4,625 ships on Zupreape all realign themselves on the Murdoch wormhole. Dear lord I hope those carriers aren't coming along for the long as well.

12.31.52: 12,076 younglings tiger to Murdoch. Allow me to repeat that. 12,076 youngling tigers. Forget Murdoch going down, that's more than enough to terminate Rakalurn too. I'm going to lose those fabs.

12.32: 300 attack on Murdoch. And plenty more coming. Luckily it seems like the special forces are now retreating to the AI homeworld.

12.32.13: Make that 3,000 attack. With another 1,000 in carriers.

12.32.17. Okay, at this point I need to prioritise. Firstly, the Mk. 5 plasma sieges are getting ferried back to my homeworld. If I lose the fabs, I can't rebuild them. Mk. 5 leeches can stay where they are for the time being, but I'll try to pull them out before they pop. The same with the maws, try to retreat them before they go.

Now, as for lightning torpedo frigates ... I think 25 of them are going to play a little game of keep away with the AI. When the command station looks like its going down, load them into the jumpship and send them to the other side of the grav well. I'm praying this keeps the youngling tigers occupied for long enough for them to atrophy, assuming they don't just go straight through the wormhole to Rakalurn.

12.32.20. Missile silo going up on Rakalurn. 3 Mk. 2 lightnings queued up. At this point I have little to fear from AIP increases beyond slightly larger waves.

12.32.40: Plasma 5s rescued. The only leeches left are the Mk. 5s, and they're at 70% health. Time to evac them. Scrap level is enough to start spider construction on Rakalurn. 4,162 attack on Murdoch.

12.33.01: Oh lord. CPA declared. Hitting in 10 minutes, so (supposedly) 16,709 ships. Spire jumpship is now retreating all remaining starships to Rakalurn.

12.33.22: The tigers are here. 16,094 attack on Murdoch. That's just completely absurd. Lightnings and sentinals are doing their circle strafing thing again, everyone else has been evacced.

12.33.27: Murdoch command goes down. Jumpship is evacuating sentinal frigates.

12.33.32: Oh man, nasty thought. Do younglings in carriers attrition?

12.34: No. No they don't.

12.34.29: 15,216 threat. Jumpship is busy hopping lightning frigates from one side of the grav well to another. The 17 remaining sentinal frigates which haven't been evacuated are told to fire on some of the remaining carriers to get them to disgorge more younglings.

12.34.30: Defences are going up on Rakalurn. Beams, lasers, mines etc. Plus all the 'spare' forcefields no longer needed on Murdoch.

12.35.30: There's a raid starship on Zarpar killing Mk. 5 snipers turrets. Zarpar, as in my homeworld. Sniper turrets, as in the ones I turned off accidentally along with the spiders when I wanted them not to crash my economy by building 6 caps at once.

Thanks goodness he didn't decide to go for the completely unshielded lightning missiles instead.

12.36: This strategy of jumping ships to remote corners of the grav well is working insanely well. Limburger supreme. I think I can keep the entirety of the threat locked down here for as long as I need.

Current threat: 9,226. How much of that decrease is due to chipper shredding ... I wouldn't like to hazard a guess. Strength on Murdoch: 130,000. A.k.a around the size of the special forces fleet that I nuked in that 9/9 heroic game. And if it wasn't for the fact so many are Mk. 5, I might just do the same here.

12.36.01: A nuke starts building on Rakalurn. Y'know, just in case.

12.37.23: Oh god. There's a neinzul regeneration chamber on Zupreape. I wondered where all those pesky little almost dead younglings were going.

12.37.26: Jumpship swings by Rakalurn to pick up one of the rescued Mk. 5 plasma starships.

10.37.37: Regen chamber explodes.

12.37.46: Blast, the brief sojourn on the jumpship allowed them to get in range to take out one of the snipers.

12.38.13: 2,200 bomber wave to nowhere. I should probably have put a Mil 2 up on Rackalurn. Doing that now.

12.39.28: 122,000 strength on Murdoch. The cheese stops them from doing much, but it ain't exactly killing stuff quickly.

12.41: 3,300 bombers to Rakalurn. Salvage time!

Oh, and these two waves are syncing with the CPA.

12.43: CPA hits.

  • 114 Mk. 2 ships
  • 257 Mk. IV ships
  • 4,120 Mk II ships from strategic reserve.

So, not quite the 16,000 you promised then.

Still, I remove all but one of the lightning torpedo frigates from murdoch - I think chasing around just that one guy is enough to keep the threat over there occupied.

12.44: Attack on Rakalurn, couple of hundred. And just after defences finished building, too, it's like the AI wants to feed me salvage.

12.44.31: Bombers waves hit. 3,500 attack. 13,300 threat.

12.44.43: 5,700 attack. Time for the first lightning missile I think.

12.44.55: Eh, took out 1,500 or so. Good enough.

12.45.10: Second missile pops 500 zenith siege engines. 2,700 attack. I think I can manage the rest without another missile, but there is one in reserve just in case.

12.46: Actually, I could probably have done it without the second missile too. Oh well. Both human resistance and marauders show up again - the marauders once again from behind the command station. They go after the still-not-cloaked penetrator.

12.47: Last ship on Rakalurn dies, ignoring of course the 400 or so the maws are still digesting. 20 million or so scrap gives the economy the boost it's been needing for for a while. That single lightning ship over on Murdoch has done an admirable job of keeping that 115,000 strength threat locked down.

12.47.35: Human resistance pop in to help with Murdoch. As there's only about 20 of them, they die rather rapidly.

12.49: While waiting for the now much-faster refleet, I decide that Murdoch could use some beach heading. 24 engineers, one mobile builder. Mil 2 command and 48 missile turrets are up in a jiffy. I'd send in rebuilders, but they they'd just start trying to rebuild mines and stuff that have moved to Rakalurn.

12.50: Full cap of needlers, missiles and MRLS Mk. 5 go up. They'll go down again soon enough, but at least it gives the poor threat something to chase other than that lightning frigate.

12.51: Wow. Those turrets lasted essentially no time at all and took out all of 2,000 strength. Yeah, without the jumpship cheesing I'd definitely be dead right now.

12.52: The lightning torpedo frigate reports that he's lonely. So 31 of his friends, and 70-odd sentinel frigates, warp in to join him. My one regret is that I can't get Lyudmila over there to join in the fish-in-a-barrel.

12.53: Docks report a full refleet on Rakalurn. Now time for starships.

1,000 bombers? That is tiny baby wave.

12.55: This threat clearing is going much faster now that I'm jumping 100 ships around the planet. Last of the current crop of disassembler guardians dies. Man, I hate those things. Fast, and eat starships for breakfast. Snipers retarget the remaining ships based on fastest speed guardians/starships first. 96,000 strength remaining.

12.56: Two dire guardians spawning soon. Ignorable.

12.57: While I'm waiting, 4 regular transports go up on Rakalurn.

12.59: 800 ships quietly load into transports and file through Murdoch to Zupreape. The threat is still intent on the snipers.

13.03: The 500 remaining ships on Zupreape are terminated, along with the needler guard post. That still leaves the Mk. 3 fort at the top, but I'm going to ignore that for now as it's safely out of the way. The jumpship takes a brief break from the sniper ferrying duties and brings over a colony ship and a few engineers. Warp jammer and a full cap of turrets for Zupreape, please.

AIP: 164
Enemy strength on Murdoch: 62,000

I have to say, this is absolutely the most broken strategy I've ever devised. It not so much the killing power of the sentinels and lightnings, although that's fairly significant for a full cap (~5,000 strength per minute), more the fact that the AI just never gets a shot off. Those threat ships have been on planet for over half an hour now, I'm half expecting them to nip off to join the threat fleet somewhere else soon.

Offline Aklyon

  • Core Member
  • *****
  • Posts: 2,089
Re: 10 planets, diff 10 ... what could go wrong
« Reply #64 on: April 15, 2014, 07:47:17 pm »
That is one jumpship whose has certainly paid for itself, even at spirecraft costs.

Offline keith.lamothe

  • Arcen Games Staff
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 19,505
Re: 10 planets, diff 10 ... what could go wrong
« Reply #65 on: April 15, 2014, 08:30:33 pm »
Well, the CPA did involve some warheads. But that is not the master cheese...
No, no it wasn't.  The Jumpship Cheese is... masterful, indeed. 

The game bows (and breaks) before your sheer combination of audacity and lack-of-decency ;)  I daresay that you may be able to hold off an arbitrarily large force of normal threat (or threatfleet, probably) by such means.  There may be specific AI units that could counter it, but in their absence... 

An exo would also counter it implicitly by not caring about what shoots at it while it's enroute.  In theory I could add a new threat behavior that tasks a clump of threat with exo-like behavior, but I'm not sure that's really a general solution here (and detecting when to do so, and when not to do so, would be tricky indeed).

Anyway, well done.  Cheese of the Month award.  Possibly of the Year, we'll see.  But you know I have to nerf that now ;)

The question is: how to do so, without unduly harming the Jumpship's other utility?


Quote
Thanks goodness he didn't decide to go for the completely unshielded lightning missiles instead.
Oh, definitely.  Wouldn't it be a shame if he were taught to do that?


Quote
12.36.01: A nuke starts building on Rakalurn. Y'know, just in case.
It's the only way to be sure.


Quote
12.37.23: Oh god. There's a neinzul regeneration chamber on Zupreape. I wondered where all those pesky little almost dead younglings were going.
You mean one of those chambers was actually being useful?  To the AI?  I'm not sure I've ever seen it work before.


Quote
1,000 bombers? That is tiny baby wave.
I don't know what's up with that.  Maybe the intervals are just that different sometimes.  I'll see if that save generates any explanatory wave loggery.


Anyway, looks like the AI may as well call gg on this one, but perhaps it has bit more dastardliness left.
Have ideas or bug reports for one of our games? Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Draco18s

  • Resident Velociraptor
  • Core Member Mark V
  • *****
  • Posts: 4,251
Re: 10 planets, diff 10 ... what could go wrong
« Reply #66 on: April 15, 2014, 09:51:09 pm »
Anyway, well done.  Cheese of the Month award.  Possibly of the Year, we'll see.  But you know I have to nerf that now ;)

The question is: how to do so, without unduly harming the Jumpship's other utility?

I don't know if you can.  Self-attrition on friendly worlds (regardless of other conditions) is just uncalled for.  Revoking teleporting under the same just feels like a kick to the family jewels.

I don't think any other penalty would matter.  Paralysis?  Nope.  Just slows down how quickly you can unload the units.  Full weapon cooldown on unloaded ships?  Nope, the AI units would still retarget (not to mention most drone spawners have very short reload times).  Paralysis on unloaded units?  Would undermine one's ability to hop in close to something and smash it to bits (they'd get blown apart first).

Offline keith.lamothe

  • Arcen Games Staff
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 19,505
Re: 10 planets, diff 10 ... what could go wrong
« Reply #67 on: April 15, 2014, 10:22:52 pm »
Maybe: units getting out of a jumpship may not enter another transport for, say, 20 seconds.
Have ideas or bug reports for one of our games? Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Draco18s

  • Resident Velociraptor
  • Core Member Mark V
  • *****
  • Posts: 4,251
Re: 10 planets, diff 10 ... what could go wrong
« Reply #68 on: April 15, 2014, 11:01:29 pm »
I'm not sure that would actually mean any thing except in hostile territory.  It doesn't counter the cheese here (though I admit I have no idea how rapidly he needed to switch sides on the gravity well).

Offline TechSY730

  • Core Member Mark V
  • *****
  • Posts: 4,570
Re: 10 planets, diff 10 ... what could go wrong
« Reply #69 on: April 16, 2014, 12:50:29 am »
I think teaching the AI when it is no longer worth chasing down the "small stuff" (or at least not with the full attacking force) compared to moving onto the next planet is the only real "counter", even though that is probably the trickiest to code...

Offline RockyBst

  • Full Member Mark III
  • ***
  • Posts: 219
Re: 10 planets, diff 10 ... what could go wrong
« Reply #70 on: April 16, 2014, 03:04:10 am »
Give me a jumpship, a single chokepoint on an X map, and a single ship with radar dampening, and I reckon I could could hold off the CPAs from the showdown devices. Yep. 

The suggestion I would have is to cap the time threat fleet spends engaging attackers. So, if all irreplacables on their current planet are gone, and they've been under attack for > 10 minutes, do the normal calculation of where to go next ignoring the currently engaging forces.

Please, please do not nerf the jumpship itself - that single Titanite asteroid I found has given me incredible, amazing amounts of utility. I wouldn't have gotten close to hour 6 without it, and I'd say probably 3/4 guardposts removals since then have been due to it. But realistically, how often to people even find a Titanite asteroid? It, along with the first ARS having sentinels and the railpod/reprocessor fabs, was one of the few strokes of luck the RNG gave me.

Also, this 'keep threat locked down' cheese can be done in a similar manner by just throwing a stream of cheap ships (younglings, for example) through a wormhole.

EDIT: As for the wave itself, here's your answer. The wave interval multiplier is normally around 5/6.

Quote
4/15/2014 10:59:41 PM (7.020)
-----------------------------------
Determining Next Wave Interval; Current Game Time: 12:38:11
baseInterval = floor( ( 14 - this.AIDifficulty ) * 3 * 60 ) = 720
unmodifiedRandomLow = 2
unmodifiedRandomHigh = 12
Because Diff >= 7 and entry_points = 0, modifiedRandomLow is 2 and modifiedRandomHigh is 9
Because game time > 2 hours, using modified range.
randomRoll = rand(2,9) = 2
randomAmount = floor( randomRoll * player.AIDifficulty * 10 ) = 200
realInterval = baseInterval + randomAmount = 920
NextWaveSizeMultiplier = realInterval / baseInterval = 1.28
= So will announce next wave for this player at T+920, around 12:53:31

4/16/2014 12:02:38 AM (7.020)
-----------------------------------
Starting CreateWaveToPlanet at Game Time: 12:53:32 ; Player.AIType: Vanilla ; Player.AIDifficulty: 10 ; EffectiveAIP: 164 ; AITechLevel: 2 ; IsSchizo: False
MultiplierFromWaveInterval = 1.28
« Last Edit: April 16, 2014, 03:34:16 am by RockyBst »

Offline keith.lamothe

  • Arcen Games Staff
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 19,505
Re: 10 planets, diff 10 ... what could go wrong
« Reply #71 on: April 16, 2014, 08:56:01 am »
Yea, I realized this morning that nerfing the jumpship has the tiger by the wrong end.  I've been considering for a little while now adding a rule where, rather than all threat on a planet either going with the "no interesting human presence here" AI branch (which looks for new places to bust up) or with the one of the "fight it out" branches, I could have it:

- if AI threat strength on planet > total human strength on planet * 10
- and there are no non-mobile irreplaceable units to kill
-- then have enough threat do "fight it out" to maintain 10x superiority
-- have the rest of the threat act as if there is no interesting human presence here and find something else to kill

So in this case you could potentially have tied up a chunk of the threat indefinitely, but the rest would have cruised on.  And quite possibly won.

Though I have great confidence another cheese would be devised ;)
Have ideas or bug reports for one of our games? Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Draco18s

  • Resident Velociraptor
  • Core Member Mark V
  • *****
  • Posts: 4,251
Re: 10 planets, diff 10 ... what could go wrong
« Reply #72 on: April 16, 2014, 10:27:09 am »
Though I have great confidence another cheese would be devised ;)

Two or three planets performing this tactic (with 25 times as many human units being the distraction). :P

Although that would lock up a chunk of fairly limited size (500 times as big as the human force).

Offline keith.lamothe

  • Arcen Games Staff
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 19,505
Re: 10 planets, diff 10 ... what could go wrong
« Reply #73 on: April 16, 2014, 10:32:18 am »
Two or three planets performing this tactic (with 25 times as many human units being the distraction). :P

Although that would lock up a chunk of fairly limited size (500 times as big as the human force).
Yea, if you can pull off this kind of dance across 3+ planets then I think that crosses the line from "cheese" to "you're a machine, have fun" ;)  And even there it wouldn't have the theoretically-unlimited capacity it currently does on just 1 planet.

I have no problem with (and indeed I heartily encourage) clever human tactics that establish inordinately high kill-to-loss ratios (or "strength-tied-up"-to-"strength-defending" ratios), so long as it isn't a "if I do this trick, my ratio is infinite in my favor" situation... kinda like this one :)
Have ideas or bug reports for one of our games? Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Aklyon

  • Core Member
  • *****
  • Posts: 2,089
Re: 10 planets, diff 10 ... what could go wrong
« Reply #74 on: April 16, 2014, 01:00:20 pm »
With enough moving MSDs on hand, you could contain threat indefinitely*