Brainstorming about offense vs defense, and refleeting speeds.

[etheric42]

Blah.  Yeah, that pokes huge holes in it right there.  It also points out a probable flaw in the late-game as it exists already: once you're strong enough, you don't need to do anything special to make for supply lines to the AI homeworld, unless your flagships die.  That whole "ships being rebuilt on the front lines" thing biting us again.

Wasn't that just how it was in AIWC?  You got to where you were strong enough to kill the AI.  Location data on the AI.  And maybe a world 2 jumps out to reinforce from.  Then checked to see if your math worked out and either won or woke the beast. 

I don't see how that's different than: get where you are strong enough to kill the AI.  Kill jammers to get location data on the AI.  Find a staging ground close enough so your flagship(s) aren't beat up too much when you show up to fight.  Then check to see if your math worked out and either win or wake the beast (with your flagships crushed and on the AI homeworld for you to find some way to rescue).

There were two AI controllers before, so you either had to do both simultaneously or deal with a second one that was stronger.  There are 1+ overlords cores now, so you need to do all of them simultaneously, jump your non-flagship fleet from one to the next, while somehow keeping the secondary flagships alive without defense fleets, or deal with the later ones that are stronger after finishing the earlier ones.

And:

Spoiler for Hiden:

Heck, the enemy king unit has two forms now.  You could have all AI kings switch to their second form the moment any one of them does.  That's scary.

Remember, the player(s) will likely have a limited number of flagships.  They get prototype+Ark to start with, and will likely fix up one or two more... but a lot of their AIP would probably be better spent on golems and new ship types.  So it shouldn't get too overwhelming.  And by the time you have multiple new flagships, your area to defend is larger, so you'll need to commit more of them to defense.

[x4000]

It's true... and losing a flagship on the doorstep of the AI and then having to do a deep strike to get it back (either with another flagship, or -- shudder -- your own ARK) would be pretty terrifying.  So nevermind on that.  I still think you have to dig in less in order to establish a base there, but that's just because it's a different game and there's plenty of things to compensate.

[Magnus]

My main problem with "I can teleport a fleet to enemy planet X" is: unreliable A.I. tactical choices + savescumming = easy cheese victory.

Let's say the A.I. picks the optimal tactical choice 60% of the time, as was mentioned a few days ago in another thread.
So I send a bunch of scouting units around to find where the A.I. core is. Let's say I find it 10 planets away from me.

I move 3 flagships to my border. Save game. Jump to first planet in path. Does the A.I. pick the "right" choice i.e. charge at me? --> reload. Does it pick the "wrong choice"? Whohoo. All my 3 flagships get safely to the next planet. Save game. Repeat 10 times.
If the randomizer for the tactical A.I. choices is preserved in the savegame (say, a numeric seed) to avoid this problem, I can then use savescumming to foresee A.I. actions.

This doesn't mean shutting down the whole thing, though. Just make it so you cannot warp in small ships if you're more than X hops away from your nearest planet/dock/whatever you want to pick as logistical anchor.

Avoiding the "let's build 100 docks so I can instantly rebuild my whole small fleet": every dock costs fuel. So, enjoy your balancing act between max sustainable fleet numbers and how fast you rebuild your losses.

[BadgerBadger]

Right now I honestly feel like there isn't enough variety in things to spend Science on. I basically have 3 trees right now; Starships, Fleetships and Turrets. I actually think this is too few; in AIWC I had way more categories spend my Science on. The major categories that stood out to me were Turrets, Fleetships, Starships, Income, Production Speed, Utility, Command Centers. Now we just have Fleetships, Starships and Turrets; I miss having all of the variety. And now proposing to drop that even more feels like it's taking a lot of the replayability away. Every game is going to be more similar to every other game; I'm probably just going to unlock the same things every game. Also it really diminishes the role of specialization in multiplayer. There's no way to have someone focus on (say) defense while another player goes glass cannon.

Incidentally, re: turret balance. I build 5000 strength of turrets (full cap of MLRS, Laser, Tractor Mark 1, 1/2 cap of Needler Mark 1), and it gets wiped out by 4K strength of attacking AI ships. However, the turrets would immediately start to rebuild, so the AI ships never actually moved on from trying to fight those turrets. They would just wait around for the turrets to rebuild then blow them up again. I tried changing the balance_seconds_after_death_before_rebuilding value but that didn't really change the AI ships behaviour.

[ptarth]

re: Forum Scroll

• Amusingly I started writing my original post when there were just 2 comments in the thread. I was confused no one had brought up anything yet.

re:Divergence from AIWar

• On the surface, maybe it sounds like it diverges from AIWars, but I think the core is still the same if you get down to what you are actually doing.

re:Edge cases & Gui issues

• No doubt. No argument there.

re: Who each Fleet Ship Belongs to

• Each Ship would have a Owner Tag somewhere in the Gui (probably next to ship class).
• Transfer of ships between Fleets would happen from the FlagShip/Planet controls level.

re: Ships are stored in a Limitless Inventory (aka Magic Bag)

• I sounds corny, but effectivley its the same local population caps and the ability to instantly transport your ships wherever you need them. I'd argue that you shouldn't track health and such, but just raw ship counts. Transfer/Created ships would just be new entities.
  

re: Problem with Fast Travel out of the Magic Bag

• As noted, unlimited instantaneous fast travel has infinite cheese. So just don't allow it.
• Ships can fast travel to the FleetShip when it is in an owned or neutral terroritory.
• Ships are highly limited when it comes to fast travel to or through enemy territory.
• In addition, you might have a cap on just how many ships can be using Fast Travel and it takes time/energy to send them.
• This is actually somewhat similar to how Transport Ships worked in AI War 1, just a little bit abstracted and you don't have to actually build or micro the transport ship at all.

re: The Problem with Fast Travel into the Magic Bag

• Ugh, micro cheese (retreat under attack and heal to full instantly).
• Logistic distance also is completely lost.
• The solution seems to be to restrict it like Fast Travel out of the Magic Bag.
• Although there isn't any need to have the Fast Travel rules be the same.

Re: Does Limiting Fast Travel Save Chris's Magic Bag?

• I think we'd also need off-map existences for ships, but otherwise.... Maybe?
• I think it at least would address the initial concerns.

Re:Why have Factories at Planets

• Give Factories weapons or have them be the best way to reinforce when under attack?
• Otherwise, I don't really see a good reason to worry about it either way.

Re: Why is logistics fun?

• I think the answer here is because it adds an interests choice and has a cost.
  
• Do I retreat with my main fleet, do I allow my defenses to handle it, or do I bring in another reserve fleet.
• The dull part is refleeting, so being able to replenish your fleet sans micro when in safety, is highly desireable.

Re: Why so many flagships in Ptarth's giant bulleted list

• The number of planets you have would be the upper limit on FlagShips, not the forced lower limit. In early game, I don't think you'd really want more than 1 or 2. Chris's Magic Bag Idea is a reasonable variant or extension of this. You just shift the X:1 link from planets to flagships to a three part system: Planet: Magic Bag: FlagShips. I think that would simplify a lot of the headaches
• I was trying to resolve this in my originally very short response. I think restricting fast travel as above would work.

Re: Technology Tying

• A terrible horrible hilarious interim idea for tying technology together. When you pick a technology, a similarly priced technology in the opposite tree is picked for you, at random. So you can choose a defense and get a free corvette upgrade. Huzzah! Obviously terrible for the long run, but would require very little work.

[keith.lamothe]

That whole "ships being rebuilt on the front lines" thing biting us again.

Starships would not be rebuilt on the front line. If you're strong enough to operate deep in AI territory without losing any starships then waiting for fleet ship reinforcements would probably be more of an inconvenience than a challenge.

That said, it was possible to do this in AIWC as well, it just involved the Mobile Space Docks, which were great once you got past the fiddliness of using civilian units in a fleet.

The goal of the flagships is to deliver that experience (which was awesome) without the fiddliness.

Just as the Shield Starship/Flagship was there to deliver the awesome experience normally only possible with a specific bonus ship type (or certain specific balance-altering minor factions).

The truth is, both of those tactics are OP in AIWC, and they're OP here. I think they're a fun kind of OP, but Shields already bit it, and I haven't heard anything that indicates that was a bad decision. Perhaps it's time to try killing off mobile fleet construction as well.

[chemical_art]

Right now I honestly feel like there isn't enough variety in things to spend Science on. I basically have 3 trees right now; Starships, Fleetships and Turrets. I actually think this is too few; in AIWC I had way more categories spend my Science on. The major categories that stood out to me were Turrets, Fleetships, Starships, Income, Production Speed, Utility, Command Centers. Now we just have Fleetships, Starships and Turrets; I miss having all of the variety. And now proposing to drop that even more feels like it's taking a lot of the replayability away. Every game is going to be more similar to every other game; I'm probably just going to unlock the same things every game. Also it really diminishes the role of specialization in multiplayer. There's no way to have someone focus on (say) defense while another player goes glass cannon.

Incidentally, re: turret balance. I build 5000 strength of turrets (full cap of MLRS, Laser, Tractor Mark 1, 1/2 cap of Needler Mark 1), and it gets wiped out by 4K strength of attacking AI ships. However, the turrets would immediately start to rebuild, so the AI ships never actually moved on from trying to fight those turrets. They would just wait around for the turrets to rebuild then blow them up again. I tried changing the balance_seconds_after_death_before_rebuilding value but that didn't really change the AI ships behaviour.

This is all very true, and goes with my observations as a player. It is like pulling teeth for me to get extra turrets because I never get the "wow" factor out of turrets to begin with. Lumping them with ships is like my parents forcing me to eat my vegetables to get my dessert: It's best for me long term but I wouldn't be forced to get them if the vegetables were delicious to begin with and I am not going to like it. The further axing of total research options is concerning too, games are samey enough.

I feel like turrets could get a passive synergy with fleetships, and it has been done before in AIW1. So higher levels of fighters can boost needler turrets, etc. But this is still band aids from turrets not feeling like they pull their weight. I never got that vibe in AIW 1, so when the waves starting getting too strong I didn't feel bad spending some K on them to keep my fleets free. I feel like we keep talking around that fact. If turrets could do their job we wouldn't need magic bags of holding or split tech trees or other options to compensate for this.

Basic human defenses should feel like they can hold until a set amount of AIP, I'd say 100 but no less then 80. If players feel like the initial turrets can do their job, then they would feel much more incentive to increase them down the road. Otherwise it feels like dumping tech into a failed research branch which logically no one does.

[keith.lamothe]

This brainstorming has been helpful, but I think at this point in the project we should generally stick to either:

A) Small changes within the framework of what's there. Maybe these problems aren't as complex as we think.

or

B) If large changes are necessary, fall back to "what AIWC did", rather than trying to chart new territory.

For purely pragmatic purposes; not because this will produce the best design, but because it is much more likely to produce a good design within our current constraints.


So overall I suggest our approach to the defensive game be:

1) See how 0.719 (now out) does. In Badger's testing it was a step in the right direction, at least.

2) If tractors+turrets are still insufficient for defense-against-waves, investigate what's going on at a mathematical level, because maybe there's some kind of bug. 5x strength should have been enough; 10x really should be.

3) If the "I don't have enough turrets to cover everything" issue is just not getting resolved, unlock Mark 1 of all turret types from the start.
- If that still doesn't work, just fall back on the AIWC model of having turrets cost Fuel (AIWC's energy) and have a per-planet cap. Then have the sharp distinction be "Fuel is for everything with a gun" and "Power is for everything without a gun" (tractors, tachyon, gravity, construction facilities).
- Alternatively, do the tech-simplification Eric and I discussed: remove turret-ship-cap entirely, keep turret-power-cap, and have the techs be "upgrade all turret types to Mark 2", 3, and 4. I'm wary of that sending a "we're sidelining turrets compared to AIWC" message, though.

4) If the "building forces on the front-line" thing just isn't working as the common-case instead of the uncommon-case it was in AIWC, remove the fleet ship build menu from Flagships.
- They would still have the ability to repair anything (if it's not under fire), and to reinforce squads that had taken losses (even if that squad is under fire), so fleet ships would still be way more convenient than they were in AIWC.

5) If the "scrap my waiting-to-rebuild dead turrets and plant new ones" tactic proves unacceptable, make turrets-under-construction die in one hit and have the AI assign a reasonable priority to shooting them.
- This would still allow building a second line of turrets outside the AI's current range, but that's legit if you can hold the enemy back long enough.
- If even the 1-hit-kill is too exploitable in distracting AI targeting, we could make placing a turret come with an upfront metal cost or something like that.

6) If there's really a strong need for defenses-that-can-maneuver (since we don't want pulling-the-fleet-back to be common), make the distribution nodes drone-carriers.
- One possibility is to have them deploy drone-variants of each fleet ship type you have (highest mark of each); this would kind of make Starships the "offense-heavy" tree, Turrets the "defense-heavy" tree, and Fleet Ships the "hybrid" tree. This is pushing it in terms of design change scope, though.

[MaxAstro]

Been in a little while since I've weighed in on anything here, but I've still been watching, and this discussion hits on a core complaint of mine from the first game, which was needing to pull back my fleet for defenses or needing to split my fleet between the two - which of course makes my offense slower and less effective and the game drags out.

The big thing is, static defenses are awesome, until something goes wrong and there's a punch through, and then you absolutely have to have mobile defenses.  A lot of the time, I wouldn't even engage my fleet when a big wave came; I'd just sit them at my home base so that they could field out in response to a punch-through if one happened.  If the goal is for players to not recall their attack fleet to defend (an idea I love) mobile defenses are a must.  But mobile defenses really need to be able to respond on multiple planets or they are just turrets that can reposition themselves.

Ideally, I'd like to see the defense fleet that does exactly what I'd do with my offense fleet when defending: Sit together in a ball ready to leap to action if there's a break in the line.  Sorta like a player version of the Special Forces fleet.

Instead of being locked to one planet, what if the mobile defenders simply only functioned on player-owned planets?  So they can still go from planet to planet to respond to moving threats as needed, but they can't be used to attack because their weapons turn off on a planet without power.

[MaxAstro]

Sorta like a player version of the Special Forces fleet.

Quoting myself and double posting because I realize this is really what I'm trying to say.  A defensive response fleet like the AI Special Forces is exactly what I want.

Back in AIWC I would often imitate this by keeping the nebula fleet riff raff in reserve to respond to threats basically exactly like this.

[Toranth]

Late to the thread, but my .02:

The concern is that AI attacks will sometimes overwhelm fixed defenses, and so something "extra" needs to be available for when that happens.
But, we
1) don't like the idea of scrapping stuff to quickly rebuild it half a galaxy away...
2) don't like the idea of being forced to keep half the mobile fleet at home...
3) don't like the idea of being forced to run back home every wave or two.

I'm assuming that it isn't considered bad to pull back the fleet for unusual, major events - pulling back for CPAs and Exowaves was standard behavior in AIWC.  In which case, I see people have already posted many ideas, some quite complicated.  While I've read all of them, that doesn't mean I understand everything I read yet.  But what stands out to me is this one:

Instead of being locked to one planet, what if the mobile defenders simply only functioned on player-owned planets?  So they can still go from planet to planet to respond to moving threats as needed, but they can't be used to attack because their weapons turn off on a planet without power.

Hear hear!
Long ago in an AIWC discussion, I suggested something similar.  At the time, I was suggesting that Fortresses being allowed to use warps, but require supply.  That would allow it to reinforce chokepoints, but not to go on the offensive.  That issue went away with galaxy-wide turret caps, but it may work here.  Something similar to mobile Fortresses - an obvious defensive unit, powerful but limited in count, that can be moved amongst your planets.  At the same time, it shouldn't be able to go on the attack.  I'm not sure how to restrict it to Human systems, but I have no doubt other people can come up simple methods.


As an alternate, a drone-based "Strategic Reserve" - You build into the reserve using resources (and perhaps build structures that require power/fuel to increase the size-of) and those reserves auto-deploy based on AI strength when a Human system is attacked.  Humans cannot control the units, and they won't ever go on the attack, but it is a global "help on defense" while at the same time not being too fidgety or micromanaged.


Finally, one comment about turrets in AIWC.  While at Diff 7, it was easy to ignore turrets, by Diff 9 a few unlocks and/or Mk V captures were essential.  As the difficulty went up further, turrets got even more important.  By diff 10, much of your K needed to be spent on turrets - up to 50% - just to not die quickly.

[keith.lamothe]

At the time, I was suggesting that Fortresses being allowed to use warps, but require supply.

That was the first idea that came to my mind in response to MaxAstro's point. I'm concerned about units that magically don't work based on where they are, but "units that only work on your planets" makes a lot more sense than "units that only work when not on your planets", so I think it's doable. It can fly anywhere, but the guns don't work unless the controller is friendly. If players abuse them as HP sponges on AI planets we'll deal with that later.

The fortresses would need to have a galaxy-wide-cap, and not cost power (because they can hop planets, and no unit that uses power can do that as it would require a blackout mechanic). I'd prefer that they not cost fuel either, as that would compete with offensive units for that resource (the whole idea of the fuel/power split is to avoid that).

So they would be a straight up exchange of permanent science and temporary metal (perhaps a lot of metal) for a mobile defensive bruiser.

But they wouldn't exactly be a rapid-response force. You could still get caught out of position.

As an alternate, a drone-based "Strategic Reserve" - You build into the reserve using resources (and perhaps build structures that require power/fuel to increase the size-of) and those reserves auto-deploy based on AI strength when a Human system is attacked.  Humans cannot control the units, and they won't ever go on the attack, but it is a global "help on defense" while at the same time not being too fidgety or micromanaged.

I like this a lot, because I want something that doesn't require player attention to do a decent (if not optimal) job. Even if you don't have to bring your fleet back, bringing your tactical attention back is almost as bad.

So:

1) Add "Warp Drone Cluster" to the Infrastructure build menu.
- costs significant power.

2) Give the Power Distribution Nodes (or whatever guard-post-like thing they become) the same mechanic the Carrier Starship uses to automatically produce, store, and launch its drones.

3) Each of your Warp Drone Clusters, anywhere in the galaxy, contributes evenly to the drone production of all your non-full Power Distribution Nodes.
- so if only a few nodes are not at their drone capacity, all the clusters will focus on those few nodes
- unlike normal carrier-drone production, this extra production should cost metal

Done Well, except for selecting which drone types are available and whether it's tied to fleet ship tech, etc.

One possible simplification is to not add "Warp Drone Cluster" and instead simply give that functionality to the existing Space Dock (the downside is that it's not obvious the space dock does this). It would have the same effect: you probably aren't building these on chokepoint worlds (too much power), but you can put them on interior planets to support your defenses anywhere. It weakens the defenses you can put in that interior, though.


So fortresses would be the big-doorstopper that you have to position manually that costs lots of science and up-front metal, and warp-drones would be the "fire and forget" option that costs lots of territory and during-battle metal. The latter would be less likely to actually stop an AI attack, but they'd slow it down.

Players could specialize in one, both, or neither.

Worth thinking about, anyway.

[chemical_art]

1) Add "Warp Drone Cluster" to the Infrastructure build menu.
- costs significant power.

2) Give the Power Distribution Nodes (or whatever guard-post-like thing they become) the same mechanic the Carrier Starship uses to automatically produce, store, and launch its drones. Other option is to have the drones get higher marks through research.

3) Each of your Warp Drone Clusters, anywhere in the galaxy, contributes evenly to the drone production of all your non-full Power Distribution Nodes.
- so if only a few nodes are not at their drone capacity, all the clusters will focus on those few nodes
- unlike normal carrier-drone production, this extra production should cost metal

Done Well, except for selecting which drone types are available and whether it's tied to fleet ship tech, etc.

This could just follow the standard triangle ships to start with: You get the MK I drones that the carrier starships start with and when you get fleet II's you get drone II's. It isn't ideal in that it incentives fleet ships over turrets but it is the most intuitive.

One possible simplification is to not add "Warp Drone Cluster" and instead simply give that functionality to the existing Space Dock (the downside is that it's not obvious the space dock does this). It would have the same effect: you probably aren't building these on chokepoint worlds (too much power), but you can put them on interior planets to support your defenses anywhere. It weakens the defenses you can put in that interior, though.

What is the purpose of the Warp Drone Cluster, exactly? I get that it is meant to be a power drain at the cost of turrets, but it feels unintuitve to have it suck power from one world to help aid another world. One some level it confuses the mechanic, for me. Do you need one or does each one add more total capacity? If the former it goes in the most backward planet I have and doesn't provide meaningful choice and if the latter it goes on every planet so I can get max cap which also isn't that much of a choice. Having it suck a lot of power so you don't deploy it on frontline worlds will result is scrapping schenagins that we already are worried about with turrets. In short, from my understanding, it is not the right way to go.

However, I could see the value of the building so as to act as an administrative role for the player, so they can decide parameters on how metal is spent (Don't spend if under 1/3rd cap, etc)


So fortresses would be the big-doorstopper that you have to position manually that costs lots of science and up-front metal, and warp-drones would be the "fire and forget" option that costs lots of territory and during-battle metal. The latter would be less likely to actually stop an AI attack, but they'd slow it down.

Players could specialize in one, both, or neither.

Worth thinking about, anyway.

I'm a fan of both mobile forts and the drone concept, and my critique of the Warp Drone cluster is just a questioning of the implementation. I don't want turrets to be crowded out due to power reasons due to competing defense builds as a first resort since there is already a struggle for players to use them to begin with. If, down the road, it turns out mixing turrets and drones are OP then that can be looked at but I feel like the two are much better when they work together rather then competing with each other.

[keith.lamothe]

The idea is that those behind-lines planets would be contributing to the defense of the other planets, yes. But if it sounds unintuitive that's a good reason not to do it. Probably just having the nodes be guard posts with drone-launchers (like the carrier starship), and having the power of those drones be based on your current fleet ship tech, would be quite sufficient.

Then the fortresses could have their guns be based on your current turret tech, so both fleet ship and turret research have their side effects in a reasonably intuitive fashion.

And the Ark, as we were talking about in the other thread, could get guns from turret tech. It could also get drone launchers from fleet ship tech. But it wouldn't do a lot of good unless you put a lot of Science into the Ark itself, as well.

So taking your concept of the "radar graph" the science categories would be like:

[Fortresses] <-- [Turrets] --> [Ark] <-- [Fleet Ships] --> [Drones]

And then there's Starships, which are all about offense, and Resource Production, which is in the middle supporting each of these.


Now, is any of this a good idea? Hard to say. I think it's probably better to rein it in closer to AIWC and the current state of AIW2, but it could deal with some of the current complaints.

[chemical_art]

Your idea of having the drones come from "guardposts" is the most intuitive and easiest to balance. Since their power comes from fleetships, and it already is a problem that a player is more tempted to pick fleetships over turrets, you don't want to put a player in the position of thinking "should I focus on fleetships or turrets due to limited power?"

This also eases worrying about potential shenagins. You don't want players use a drone carrier to pop its fleet, scrap It, then build turrets instead. Rather then build up a whole rule library to address this just make it so there is no opportunity cost.