Just how hard is the level 10 AI? (Answer: Apocalyptically hard)

[Haagenti]

In the meanwhile, the Haagenti Empire has made more progress. After 4 hours of play, I finally understood that electric shuttles (a gift from the AI) could stop the continous raiding (if you haven't been raided by 500 ships at the same time, you haven't been raided at all). Crack-zap, and they die by the 100s. Finally, I have impenetrable defenses, 4 planets and am done with the micromanagement.

I'm steal-clearing planet number 5 (basically using Etherjets to steal all ships from wormholes and kill/parasite them), thus building an enormous fleet.  Once done, I can finally go on the offensive. Cruisers rule, especially with Munitions support (another gift from the AI). The AI has no defense against my BoD (Blob of Death) of 600 cruisers with 10 munitions support and a bunch of parasites. The BoD brings down forcefields as well (so my bombers can stay out of range). Every once in a while, the enemy deploys a counter BoD, but this is easily broken up by dividing it into parts with EtherJets.

My main concern is energy. I need more safe planets to deploy powerplants. At this moment I have to continually scrap Autocannons and Space Planes just to stay into the green. I must have scrapped 1000s already. On the up side, I have located a III planet. That will give me the level III Engineers I need, and some level III cruisers to further enhance the BoD.

For the future, I have also devised a way to kill AI ships on an industrial scale (shooting them individually goes too slow). The current plan is to setup a killzone with a bunch of Electric Shuttles under a Forcefield and use Etherjets to bring the AI ships in bunches of a 100 or so to the shuttles.  You don't need the shuttles, other ships will work as well but shuttles are fast and efficient.

With this I believe I can still beat the Level 10 AI, even if it starts raiding me with 1000 ship raids, as they have to come to me. If they do not have Tachyon Fighters they are toast. Just EtherJet a 100 ships at a time and bring them to The Place of Skulls. Ten trips: five minutes tops.

Zap! Crackle! Next!

[Revenantus]

For the future, I have also devised a way to kill AI ships on an industrial scale (shooting them individually goes too slow). The current plan is to setup a killzone with a bunch of Electric Shuttles under a Forcefield and use Etherjets to bring the AI ships in bunches of a 100 or so to the shuttles.  You don't need the shuttles, other ships will work as well but shuttles are fast and efficient.

Very interesting strategy, we might have to start referring to this tactic as the 'Haagenti Maneuver'. Perhaps you'd like to add a description of it to the unit types and tactics section of the wiki?

Anyway, it sounds like you have a chance at victory - I'm rooting for you!

[Admiral]

My main concern is energy. I need more safe planets to deploy powerplants. At this moment I have to continually scrap Autocannons and Space Planes just to stay into the green.

Hearing this makes me realize that my "end game corner case" (mentioned earlier, when discussion that the reduced output per generator could be modified by number of planets controlled) is not so much an "end game corner case" in the Haagenti Maneuver. (That's somewhat like the Corbomite Maneuver but much less of a bluff.)

Cheers!

[Admiral]

we might have to start referring to this tactic as the 'Haagenti Maneuver'.

You know... I really should stop posting until I have read all the posts from now on. My doppelgänger clearly came up with the same name ahead of me, yet again.

[x4000]

Steps to reproduce:

Hope that Darke has attached the right save.
Goto Orimicci (bottom left corner).
Select ships sitting on top of the Mellau wormhole (top left corner; for any values of "corner" a circle might have).
Ctrl-right-click to send them through to Mellau.
Ctrl-left-click to flip to Mellau.
Select ships once they all seem to be in Mellau.
Ctrl-right-click to send them to Orimicci.
Observe them only half "unfold" on Orimicci. The Tech III repairers nearby may or may not repair them.
Select ships and ctrl-right-click to send them to Mellau.
Ctrl-left-click to flip to Mellau.
Observe ships stuck in a single pile, not attacking anything in range (if there's anything in range). And being hit by Ion Cannons and the like.

For repairers not repairing:
Load the same game.
Send the same ships through the wormhole.
Select and send the Tech III engineers through the wormhole.
Move the ships around a bit and observe they bunch up; and they're wounded.
Move the stack of engineers around a bit, observer they were bunched up (and still "invisible") from moving through the wormhole.
Move stack of engineers near stack of ships; note lack of zappy-healing-beams.

If you pause the game and wait for a bit, then unpause, suddenly the piles unpack a little, and some of the engineers will start healing.

Okay, cool -- this was all related to the slow collision detection on exit.  The 1.011D version solves this completely, and all of these issues are no longer present (and without any lag whatsoever on my Q6600, I might add -- other performance improvements from the past have really helped out here).  Thanks for providing such detailed steps to do this!

[x4000]

For the future, I have also devised a way to kill AI ships on an industrial scale (shooting them individually goes too slow). The current plan is to setup a killzone with a bunch of Electric Shuttles under a Forcefield and use Etherjets to bring the AI ships in bunches of a 100 or so to the shuttles.  You don't need the shuttles, other ships will work as well but shuttles are fast and efficient.

Very interesting strategy, we might have to start referring to this tactic as the 'Haagenti Maneuver'. Perhaps you'd like to add a description of it to the unit types and tactics section of the wiki?

Anyway, it sounds like you have a chance at victory - I'm rooting for you!

Yeah, definitely -- that's really cool.

[x4000]

My main concern is energy. I need more safe planets to deploy powerplants. At this moment I have to continually scrap Autocannons and Space Planes just to stay into the green.

Hearing this makes me realize that my "end game corner case" (mentioned earlier, when discussion that the reduced output per generator could be modified by number of planets controlled) is not so much an "end game corner case" in the Haagenti Maneuver. (That's somewhat like the Corbomite Maneuver but much less of a bluff.)

For this one, I'd also point out that certain ship types cost more energy than others, and if you're using parasites, etc, then you are going to be more energy-poor.  We might still need to tweak the efficiency reduction, but I'm not convinced quite yet.

[Haagenti]

As input for your energy reduction tweak: I'm currently using three II and three I reactors per planet, and considering adding a fourth. This made life hard when I had only one planet, but it's easing up now. I now have three safe planets (I cleared a fourth but haven't settled it yet, to avoid giving the AI a second raiding point). I think I need 3-4 more planets to settle all my future power needs.

The parasites (at 400/pop) don't help, but I need them to convert massive amounts of space planes and autocannons. I'm not sure for what yet: perhaps 1000 autocannons/planes are sufficient to bring down an Ion Cannon by themselves.

Very interesting strategy, we might have to start referring to this tactic as the 'Haagenti Maneuver'. Perhaps you'd like to add a description of it to the unit types and tactics section of the wiki?

I would, but methinks that X will modify the game in response to my strategy, thereby invalidating it, so I haven't taken the time yet. I can only hope (beg, beg) that X will name a planet after me for this strategy.  

I'm currently leaning to the opinion that EtherJets are unbalancing the game and will have to be nerfed. They give me the ability to choose the time and place of battle, both in offense and in defense, and that means that I can handle almost anything.

I believe that with Etherjets and a reasonable starting position, that if I can survive the first hour, almost any AI-7 can be easily defeated (possibly with the exception of the AI that cloaks everything even when firing and maybe the Attritioner as Etherjets are not very sturdy). What I've seen of AI-8, the same probably holds as well. Don't know about 9 but the premise (...that if I can survive the first hour...) is certainly false for most AI-10 that do serious raiding.

Anyway, it sounds like you have a chance at victory - I'm rooting for you!

Yeah, definitely -- that's really cool.

Thanks, guys.....that keeps me going through the sometimes tedious mass-murder.

My next ship tactic will probably involve a mobile Forcefield and Electric Shuttles (the Haagenti Maneuver 2.0). I think that almost nothing can stop this Juggernaut: it will simply chug towards the AI Command Center and Warp Gate, take them out and then happily zap everything that converges on it. The more, the better.

Only EtherJets (who will simply drag the shuttles out of the field) or massed Snipers/Sniper Turrets to blast them out will counter this. Infiltrators need to get close so they will be zapped. Bombers can reduce the forcefield, but they need to stay out of range (and there need to be a lot of them). And I need only 15 or so shuttles, so there is plenty room for cruisers under the field.

BTW: To save time on experimenting: does anyone know whether Ion Cannons will penetrate Mobile Forcefields?

[PhilRoi]

intresting

[x4000]

As input for your energy reduction tweak: I'm currently using three II and three I reactors per planet, and considering adding a fourth. This made life hard when I had only one planet, but it's easing up now. I now have three safe planets (I cleared a fourth but haven't settled it yet, to avoid giving the AI a second raiding point). I think I need 3-4 more planets to settle all my future power needs.

The parasites (at 400/pop) don't help, but I need them to convert massive amounts of space planes and autocannons. I'm not sure for what yet: perhaps 1000 autocannons/planes are sufficient to bring down an Ion Cannon by themselves.

That's a lot of ships for just a couple of planets, so I'm thinking the energy is seeming okay for the moment with that.  We'll see how it goes, but right now I'm feeling pretty okay about that based on your description.  Thanks for the notes!

Very interesting strategy, we might have to start referring to this tactic as the 'Haagenti Maneuver'. Perhaps you'd like to add a description of it to the unit types and tactics section of the wiki?

I would, but methinks that X will modify the game in response to my strategy, thereby invalidating it, so I haven't taken the time yet. I can only hope (beg, beg) that X will name a planet after me for this strategy.  

I'm not sure what to do to invalidate that at the moment, actually.  I might make the AI send in a little "tracer" guy first to have your shuttles blow their attacks before the main fleet gets there, or something like that.  I'm going to add that to the short-term list, anyway, and we'll see where that goes.

As for the planet names, those are all randomly set by a collection of syllables that are randomly combined and then scrubbed to be a little bit more readable.  However, what I did was I added syllables for everyone from the current 1.011 release notes, so there's a slight chance that it will produce your various names for some planets.  Actually, it will produce your names at some point, it's just that the likelihood of any of you seeing those is really low.  Same with my name and the names of my alpha testers, incidentally, but I've only ever seen it come kind of close to one of our names (and it got the nickname Desi for one of our alpha testers at one point, I think), so it's really really rare.  But in there.

I'm currently leaning to the opinion that EtherJets are unbalancing the game and will have to be nerfed. They give me the ability to choose the time and place of battle, both in offense and in defense, and that means that I can handle almost anything.

I believe that with Etherjets and a reasonable starting position, that if I can survive the first hour, almost any AI-7 can be easily defeated (possibly with the exception of the AI that cloaks everything even when firing and maybe the Attritioner as Etherjets are not very sturdy). What I've seen of AI-8, the same probably holds as well. Don't know about 9 but the premise (...that if I can survive the first hour...) is certainly false for most AI-10 that do serious raiding.

I think that perhaps Etherjets should not be able to use their tractor beams while cloaked (but they can recloak and keep holding their current ship once they do capture a target).  I've added that to my to-do list, I think you are right that it is unbalancing.

BTW: To save time on experimenting: does anyone know whether Ion Cannons will penetrate Mobile Forcefields?

Currently they are not, but in the next prerelease they will be able to.  Thanks for bringing it up.

[Haagenti]

That's a lot of ships for just a couple of planets, so I'm thinking the energy is seeming okay for the moment with that.  We'll see how it goes, but right now I'm feeling pretty okay about that based on your description.  Thanks for the notes!

If you change the Energy formula, nerf it downwards: I think I generate too much energy from my few planets. What I consider to be "sufficient for my needs" is something rather formidable. 

Same with my name and the names of my alpha testers, incidentally, but I've only ever seen it come kind of close to one of our names (and it got the nickname Desi for one of our alpha testers at one point, I think), so it's really really rare.  But in there.

Fair enough

Currently they are not, but in the next prerelease they will be able to.  Thanks for bringing it up.

Don't mention it. You may want to consider toning the higher AIs down a little now that many strategies have been nerfed: keep the insane raiding, but lower the insane rate at which it fills up its planets.

[darke]

Currently they are not, but in the next prerelease they will be able to.  Thanks for bringing it up.

Don't mention it. You may want to consider toning the higher AIs down a little now that many strategies have been nerfed: keep the insane raiding, but lower the insane rate at which it fills up its planets.

The planet filling doesn't actually seem too bad for me. I only have one planet at insane levels (Mark II planet, with 4V 152IV 490III 2964II 396I), which has been sitting next to my home planet for the last 3 hours. This is what Nuclear Missiles were designed for. Though I think I shall take the hard way out.

On a random note, +40 AI progress for an Armoured Missile vs +50 AI progress for a Nuclear Missile seems... somewhat out of proportion given the Nuke will take the entire planet out for you (granted you lose resources, but the addition by most planets won't be much by the time you want to worry about these).

[x4000]

If you change the Energy formula, nerf it downwards: I think I generate too much energy from my few planets. What I consider to be "sufficient for my needs" is something rather formidable.

Good to know -- thanks.

Same with my name and the names of my alpha testers, incidentally, but I've only ever seen it come kind of close to one of our names (and it got the nickname Desi for one of our alpha testers at one point, I think), so it's really really rare.  But in there.

Fair enough

Sorry!  There's just way too many possible planets for me to be able to have full names specified.   

You may want to consider toning the higher AIs down a little now that many strategies have been nerfed: keep the insane raiding, but lower the insane rate at which it fills up its planets.

Maybe...  I essentially balance for difficulty 7 when it comes to the higher ones, and then they are a certain amount harder than that.  So with difficulty 7 getting harder now with a smarter AI, that makes the higher ones even tougher.  Difficulty 10 is not meant to be generally winnable except by some of the best players in the world, and/or the biggest experts at this game.  So the fact that I'm plugging some holes there that make it harder again is not necessarily a bad thing as far as I am concerned.

[Haagenti]

Don't you play against teleport turtles? When I started on AI-10, I considered playing against turtles or teleport turtles, but my strategy would be fairly trivial:
- Get a few planets (this is harder than it sounds)
- Defend the wormholes
- Nuke ALL other planets one by one.

If the enemy doesn't raid, it *was* quite irrelevant what their AI level is (the new fixes with raiding AIs may have changed this). Or am I, once again, totally mistaken?

[x4000]

Don't you play against teleport turtles? When I started on AI-10, I considered playing against turtles or teleport turtles, but my strategy would be fairly trivial:
- Get a few planets (this is harder than it sounds)
- Defend the wormholes
- Nuke ALL other planets one by one.

If the enemy doesn't raid, it *was* quite irrelevant what their AI level is (the new fixes with raiding AIs may have changed this). Or am I, once again, totally mistaken?

This is a really good point, and not something that I considered since adding the nukes.  However, a few points:

1.  You can clear a ton of planets like this, but you won't get the resources or the special ships from it, so at the end of the game you'll be in a really weak position.

2.  When you get to the core/home planets, nuking those won't hurt the core/home ships.

3.  On any planet, it won't hurt the starships, so they are likely to come kick your ass if you aren't careful.

4.  When you have a monstrously high AI progress and then get to the final AI planets, they will get so strong that you will have almost no hope of destroying them.

5.  They will also then launch huge cross-planet raids against you with core ships (since those are some of the only remaining planets), which will likely sweep all your planets since your resources and ship caps are so low.


I don't see being too nuke-happy as a sure-fire way to win.  In 1.010 and back it was a sure-fire way to stalemate at best, but now that stalemate is likely to turn into a loss with the latest versions.