Fleetball woes; strategic choice in AIW2

[BadgerBadger]

I was playing a bit of AIWC, and I realized that AIW2 just feels really really simple by comparison, strategically. There just aren't anywhere near as many interesting choices to make. Or at least, it doesn't feel like it. Maybe I'm missing something?

So far here's how I play AIW2. I start by spending all my Science unlocking ships. I then make a giant fleetball with my ark and start conquering things. And then I spend most of the rest of the game with just a giant fleetball. I eventually lose when my giant fleetball runs into a larger fleetball. And that fleetball is pretty much always "Every single ship I can build." One thing that contributes to this is that Flagships feel like they are made of paper compared to the Ark (a sniper flagship has 50K health to the Ark's 600K). There's no element of risk to bringing the Ark most of the time, and you'll wind up actually needing it to rebuild the flagships more often than not. I feel like the Ark needs some weaknesses because it really doesn't have any right now.

And you pretty much always need that giant fleetball. If you bring a fleet that's not Very Large onto most planets, your ships will all die without really accomplishing much. This also means that Metal isn't an interesting resource anymore since there are no choices involved (you just always need to wait for everything to build. You don't need to worry about what to build, you just always need all of it).

An example difference in early game. In AIW2, you start by building a fleet, then going off and taking a few planets. It takes about 30 minutes to take the ARS, Golem and Flagship near you, and there really aren't many strategic choices to make in there. And you totally have the initiative the whole time. You attack, if you don't win you retreat to your planets, rebuild and then attack again. In AIWC on the other hand, the AI has made two powerful attacks against me in the first 10 minutes. I've only been able to build defenses and start amassing my fleet, and if I wasn't very diligent about doing both then I would have died. There's no weak early game in AIW2.

I also feel like I never want to spend Science on turrets. In AIWC, any hole in your defenses might just let the AI kill you. So you needed to carefully plan and build turrets, which gave the game a strong Tower Defense vibe. In AIW2, turrets are much less interesting. You can't use turrets offensively, and there aren't actually any strategic choices to make with them. Since you only need to keep the Ark alive, and the Ark is generally protected by the fleetball, Turrets only exist to prevent the AI from recapturing planets. So the only thing you do with turrets is say "Okay, I captured a planet, now build all the turrets I've unlocked and send my fleet elsewhere". And you never need Shield Generators (what do you have that's worth putting under a shield generator?)



TL;DR: Fleetballs are boring, metal isn't an interesting resource anymore and I like tower defense games.

[keith.lamothe]

Thanks for writing this up. It's hard to know much about the balance and gameplay because there aren't many people who get past the UI (currently under overhaul, as you know). My thoughts on how to address what you bring up:


1) Investigate why wave frequency is apparently less than specified. The '<budget_item type="Wave" seconds_between_attempts_to_spend="600" ...' in the xml means "launch a wave about every 10 minutes" but you're not seeing them until after 30 minutes?

Whatever's going on, we should adjust it to send waves about every 5 minutes, and to (by default) announce them 5 minutes out. So the player always sees that something's coming.

We may need to shift more AI budget into waves, too, to make them larger.


2) Demotivate "bring your Ark to every fight".

The idea I've had for a while is to add an "alert" guage next to AIP that goes up every time your Ark enters an AI planet. 3 such hops and it launches an immediate wave to hunt you. Then there are other things that also increase the gauge, perhaps including simply "time your Ark is on an AI planet".

But maybe that's too complex for a simple problem. Just have a red "Your Ark is on an AI planet. The AI is sending hunters to get you!" message whenever your Ark is on an AI planet, and have it spawn stuff a few hops out with "bumrush the Ark" orders. That:

- gives you the chance to capture initial planets
- but then you have to defend yourself, and you'll feel weak and hunted
- and you get the idea that you want to keep your Ark in a safe place
- and you'll be really glad to get a flagship that can handle spearhead actions without the Ark


3) Increase Flagship survivability. We can buff these, but that won't solve the issue as generally individual units in AIWC or AIW2 just die in heavy combat unless they're (i) covered by tons of shields or (ii) less likely to be targeted than other things.

So the best idea I can come up with is to add a simple rule whereby autotargeting always prefers a non-flagship to a flagship. So if you get wiped, you lose the flagship, but otherwise you can retreat in good order.


4) With the above, I think turrets will be more attractive as you'll need something to protect your Ark while your fleet is elsewhere. In addition:
- Turrets have 3x the strength of ships of the same mark and science-cost.
- When you spend science on ships you increase your force size by one cap. When you spend science on turrets you gain one cap per planet. So if your Ark is two hops deep, researching a turret type lets you put two caps of that turret in the AI's way. And that doesn't include moving your Ark around within your territory during the defense.


5) On Metal not being an interesting resource, I'm not sure what to say. If you run out of metal your fleet's effectiveness drops through the floor because you stop being able to replace ships within squads (or build replacement squads). It matters a lot more during combat than it did in AIWC.

I get what you're saying about "I don't choose, I just build everything", but did serious play in AIWC ever not involve building all the mobile fleet you could? I mean, that's why refleeting was such a bear: you had to do it. All of it.


6) Maybe we add a rule where "if Warden Fleet ships see your Ark, they immediately go threat and beeline for your Ark", so you can bait them through your 10 planets of turrets.

Ok, maybe that's not a good idea

[x4000]

That's a depressing post, but in line with what others say when they drop by and then quickly leave. The lack of incentive to split your fleet, and the lack of tug of war on turtling yourself versus attacking the AIs feels present from the first moment of play, though, I have to say.

What is the first thing you do? Take your Ark, build all ships to cap, and go capture some planet or planets by brute force. I think that sends a really bad message, just at a gut level from the start, and players will then have an uphill battle to retrain themselves or overcome that initial impression.

In Classic, I had a lot of things you had to capture and hold in order to make you spread yourself out. How little could you defend x planet in order to make your offense stronger? Things like per-planet caps of super powerful turrets negates any need for that. And gating all turrets behind science unlocks makes it unclear that there is a strong defensive element to this, too.

My thought, at the moment, is that the Ark should be made pretty low health. Maybe give it really good guns but that's it. And maybe have special Ark-killer big guns with sniper-like range on all AI planets, ala ion cannons. Aka you have to send in ships without your Ark in order to take that out before you can move the Ark onto the planet to even help. Don't let shields block those shots.

That trains players to send ships unsupported by a true capital ship. Maybe give players some turrets from the start, maybe even give them a flagship, so that the game gets off to a quick start and they are immediately thinking about how to divide up their forces.

They have to immediately think:

Where does the Ark hide? Do I send in ships without my flagship, or with it? If my flagship dies, then I have to take my Ark out of hiding. How much metal do a pour into turrets instead of offensive forces?

How else do we get players to divide their forces? In the past, it was always a need to defend multiple places at once, but having to attack multiple places at once works just as well.

A new mechanic that is ever present could be that in order to capture one planet, you have to be suppressing all AI network nodes on adjacent AI planets. Basically having some regenerating health AI nodes that prevent capture on adjacent planets, so you have to attack multiple of them at once, anyway. Boy that would make territory capture a lot more complicated, possibly in a good way. Seems like a good AI option to try out as a thing, even if it doesn't become the new default. That way high-connection planets are not just harder to defend, they are wicked hard to actually capture, even if you can mostly clear the off.

The mechanics of the game seem solid. The objectives early on, and the early pacing, less so. I've never played into the game much because it hasn't caught my interest yet; there was never a multifaceted problem that made me go "I wonder if I can pull that off?" It was always "I'll grind away at it and win, or maybe lose."

I just haven't had any sense of attachment to early territory, either. No sense that losing it is bad, or that I'm trapped or in any serious danger until the wardens show up. The Ark being a glass cannon makes sense to me from a tension perspective. And from a tower defense perspective.

And yeah, I've actually never played long enough to see a wave, even on fast forward.

Odd as this is to say, I think these are problems that have a variety of pretty simple solutions available to them. Which will be best, I don't yet know. I like your ideas about more AI waves, and I think that the AI being aggroed by the presence of the Ark on its planets would be an interesting option x whether that's the default or not. If you don't give me some turrets from the start, I'm unlikely to build any. And it makes them seem less critical than they should be.

The per planet cap for turrets makes me iffy. What if the cap was increased for every planet you hold, but you can build those turrets wherever? I might want turrets more clustered up, but requiring expensive metal to power converters to work on a front line planet. It would be a lot more interesting than a cap of turrets on a backwater I don't care about that I'd never use; but it has the same effect of solving the "I have too many planets and not enough turrets" problem.

Thoughts?

[Draco18s]

I think what we need is a way to make sub-fleets valuable and viable.

In classic RTS games not sending everything meant you could split your forces in two: one to get the enemy to overcommit on one side while half your troops snuck around back and did some real damage.

AI War...doesn't allow for those sorts of tactics because the AI just straight up outnumbers you all the time. Heck in AIW2 I'm finding that even clearing out a single guard post requires 3 caps worth of fleet ships (at game start) just due to the sheer numbers and durability of what the AI has on their capture points. But it's not a problem to throw three caps worth of ships at it because they build so fast and I can retreat back to a "safe place" for a moment to rebuild:

Every AI unit I kill is Gone Forever and every unit I lose is Gone For Two Minutes. The AI has advantage in numbers, but I have the advantage in reinforcements: I will eventually whittle the AI's forces down and I can take the planet.

But splitting my forces means I'm trying to do this on two fronts, and with fewer numbers on my side, it means I need to refleet more times: so rather than 6 total refleets if I were to go after both objectives one after the other, I'd need (say) 9. And because unit creation is measured in time, that means its less efficient to split my forces: It will take me 50% longer to conquer both planets if I go after both of them at the same time.

[BadgerBadger]

@Keith: wave frequency is working as coded. You get a wave every 10 minutes. There's 1 AI, so every 10 minutes you get a wave. I do think that the AI should devote more resources to waves; waves are pretty unintimidating at the moment. Unfortunately, the first wave will only get queued 10 minutes in, so it doesn't show up till 20 minutes into the game. Perhaps we should start the AI with a wave queued up already?

I do like the notion of some penalty for using the Ark offensively. The Ark isn't just going to die in a battle unless you are playing very badly, or your fleet is going to be entirely wiped out. So some mechanic to make using the Ark riskier seems useful.

[keith.lamothe]

Things like per-planet caps of super powerful turrets negates any need for that.

AIWC has had per-planet turret caps for quite a while.

My thought, at the moment, is that the Ark should be made pretty low health. Maybe give it really good guns but that's it.

It used to have very powerful guns. Giving them back encourages offensive use.

Nerfing the health may be good, but I think the main thing is making it too dangerous to use in enemy territory unless you have no choice.

And maybe have special Ark-killer big guns with sniper-like range on all AI planets, ala ion cannons. Aka you have to send in ships without your Ark in order to take that out before you can move the Ark onto the planet to even help. Don't let shields block those shots.

"Game-Over Guns" I think those would feel very ham-handed to the player because it feels less like a normal consequence to "exposing your king" and more like a "the developer will point a gun at you if you do what he doesn't want you to do".

A normal consequence, that has a similar motivating factor, would be: while the Ark is on an AI planet, all AI budget is shifted into spawning threat (a few hops out) that comes after the player. This is accompanied with a very angry-looking warning message on the UI.

It's like if someone playing GTA starts punching people on the sidewalk in front of the police station. They're going to learn pretty quickly that this isn't the thing to do. You don't want to be seen, so don't stand up.

So here, in the early game, it would mean a fairly quick death unless you immediately reverse. Later-game, it might be a useful way of preventing AI defensive growth, in exchange for taking a heavy hit against your defenses, but if you do it for any length of time you're going to die.

This would change the start-of-game experience to:
1) Build your triangle ships to cap (generally done in AIWC)
2) Pick your initial unlocks, and build those to cap too (also generally done in AIWC, though sometimes you lacked the energy)
3) Take that fleet and clean out a neighbor (also generally done in AIWC)
4) Now that the target planet is not AI controlled, bring in the Ark to capture it (similar to the colony ship in AIWC, but reusable and without as many clicks)
5) Repeat until you have a decent hinterland for resources and basic defensive depth (again, very similar to AIWC, but you're more likely to pick a flagship and/or golem along the way)
6) Enter the mid-game and start taking objectives further out and searching for the AI. Since you'll have flagship(s) now, you can bring a planet-capturing unit directly into battle without triggering the AI-drops-everything-to-come-kill-you repsonse. But by then you'll have learned how to operate offensively without such a unit, so risking a flagship is not an automatic decision.

Maybe give players some turrets from the start, maybe even give them a flagship, so that the game gets off to a quick start and they are immediately thinking about how to divide up their forces.

Dividing offensive forces is not generally done in AIWC unless resistance is very thin. The player does not usually have the attention for controlling two separate battles that can't be on the same screen, and concentration-of-force is always a dominant tactical consideration.

Dividing between offensive and defensive, sure. The key to that here is making sure they have to leave the Ark "home". And they have to hold territory to maintain fuel (and thus fleet size), and secondarily to maintain metal income.

A new mechanic that is ever present could be that in order to capture one planet, you have to be suppressing all AI network nodes on adjacent AI planets.

An option that put those in some places would be interesting, but if that was the everywhere default many folks would just quit from exhaustion after a few planets. Fighting multiple offensive battles at once is very tricky.

If you don't give me some turrets from the start, I'm unlikely to build any. And it makes them seem less critical than they should be.

To a certain extent, if a player doesn't do something because it's not immediately doable, they'll just keep losing until they quit or learn to use the tools available. You can't really win either game without turrets.

That said, We can give them Mk1 Needler Turrets to start, and they'll be motivated to unlock more to get turrets with longer-range and/or are good against defenses other than Armor.

I just want to avoid the situation in AIWC where we added a few things to the "Starting set" and then a few more, and then a few more, until they wound up starting with (nearly?) every turret type and (nearly?) every starship type and 10,000 knowledge.

The per planet cap for turrets makes me iffy.

Again, it's worked in AIWC since 7.032 (nearly four years ago).

The difference is that they cost fuel (called energy in AIWC), where here they cost power. And actually they haven't even cost power for a while, because there was no motive to spend science upgrading something that was already bound by two separate per-planet resources. With the mark-level-upgrade simplification I could reintroduce the power cost, and each upgrade would simply increase their strength without increasing their power cost.

The hope with power was to avoid the player feeling that they couldn't build defenses without cutting into the size of their fleet. It also allows for per-planet customization because your tachyon coverage, shields, gravity generators, etc all cost power too.

Maybe we just need to drop power and fuel and go back to energy as a single resource (we could keep calling it fuel). I'll miss the "power as a tachyon/gravity/etc shared budget" effect, but oh well.

[TheVampire100]

Just an idea from my head, can you make it that the Ai gets additional Ai progress per increment when the Ark is on an enemy planet?
So for ecample instead of having 10 ai progress each 30 minutes, make it 15 AI progress. Maybe you could increase this the higher the planet rank is.

[keith.lamothe]

Just an idea from my head, can you make it that the Ai gets additional Ai progress per increment when the Ark is on an enemy planet?
So for ecample instead of having 10 ai progress each 30 minutes, make it 15 AI progress. Maybe you could increase this the higher the planet rank is.

There's actually no auto-increment in AIW2; it's implemented but not used because the norm of actual play in AIWC was that players never wanted it on, so they left it off.

Either way, we could make Ark-on-AI-planet cause AIP, but I'm thinking that it's better that the player only receive a temporary penalty (the AI sending tons of ships) that will kill the player quickly if it's going to kill them at all, rather than a permanent penalty that probably won't kill them soon (unless they really overdo it) but has a good chance of ruining their chance of winning long-term.

But it is true that if there's any one way we can get players to stop doing something, it's to assign an AIP cost to that thing. Unless it's warheads, of course, in which case players laugh at the cost (mostly; I think I can count the number of intentional nukes in serious play on one hand).

[etheric42]

2) Demotivate "bring your Ark to every fight".

The idea I've had for a while is to add an "alert" guage next to AIP that goes up every time your Ark enters an AI planet. 3 such hops and it launches an immediate wave to hunt you. Then there are other things that also increase the gauge, perhaps including simply "time your Ark is on an AI planet".

But maybe that's too complex for a simple problem. Just have a red "Your Ark is on an AI planet. The AI is sending hunters to get you!" message whenever your Ark is on an AI planet, and have it spawn stuff a few hops out with "bumrush the Ark" orders. That:

- gives you the chance to capture initial planets
- but then you have to defend yourself, and you'll feel weak and hunted
- and you get the idea that you want to keep your Ark in a safe place
- and you'll be really glad to get a flagship that can handle spearhead actions without the Ark

You just said the word hunt and hunters.  Maybe this is what the hunter fleets should really be doing.  Not attacking human planets (unless doing so helps them on their hunt).  Maybe the Ark needs to play a sensor game with the hunter fleets and hide its presence until it needs to commit.  I've frequently described AIWC as a "stealth RTS" because you had to be careful and strategic about what you take.  Wouldn't this be an evolution of that concept?

3) Increase Flagship survivability. We can buff these, but that won't solve the issue as generally individual units in AIWC or AIW2 just die in heavy combat unless they're (i) covered by tons of shields or (ii) less likely to be targeted than other things.

So the best idea I can come up with is to add a simple rule whereby autotargeting always prefers a non-flagship to a flagship. So if you get wiped, you lose the flagship, but otherwise you can retreat in good order.

That's simple if a bit weird about forcing that decision in that way.  What about increasing flagship range so it can sit in back?  Or do some naval equivalence and make them functionally immune to attack except by equivalently sized weapons, meaning you need a fighter/AA screen to take out bombers and if the warden fleet shows up with its group of battleships you need to book it until you can bomb them (and in general you don't bring your flagship on-planet until you run a bombing strike to take out the flagship kilers).

This solution works with the Ark too.

6) Maybe we add a rule where "if Warden Fleet ships see your Ark, they immediately go threat and beeline for your Ark", so you can bait them through your 10 planets of turrets.

Ok, maybe that's not a good idea

Cheesing the AI is part of the game, but that may be too cheesy?  What if the Warden and Hunter fleets just carry battleships that can tag the Ark/Flagships at range, but planets generally don't (just a few fixed structures that you can bomb out with squads first), so you need to distract the wardens and commit your heavy ships where they are not.

In summary, increase key ship HP (flagships/ark/dunno about golems) or maybe give them unique armor that only ships of that class have.  Create classes for the AI that prey on the big ships (they were the ones that cracked the golems in the first place) (give them armor-piercing ammo that works differently than the traditional good/bad triangle, or is "super effective").  Give some stationary versions to the planets that can be removed by sorties.  Give bombers armor-piercing ammo too, but make them something you can take down en route to your capitals.

So let's look at some inspirations:

WWII
The Ark is kind of like a carrier (it's also a mobile base, but we don't really have those to compare to). It releases swarms of smaller craft that can strike at a longer distance than a battleship's cannons.  Carriers control the sea but are also vulnerable and rely on other support ships.  What do the support ships do?  They scout, deal with submarines (stealth ships), help with AA, bombard shorelines protected by AA, escort transports.  Theoretically carriers were vulnerable to anything that got too close, the range and visibility of aircraft meant that didn't happen much.

Battlestar Galactica (new series)
Galactica was a kind of combination battleship and carrier.  When we saw it engage with an equivalent enemy it exchanged long range cannon fire while the vipers tried to control airspace.  I am not sure if the vipers were just there as a defensive measure or if they could have delivered some kind of anti capital weapon once airspace was controlled.  The long range cannon fire seemed to be something that could be received for a long period, but protracted conflict was still avoided (due to there being only one Battlestar, just like there's only one Ark).  For intentional conflicts, most of the time it just scouted with raptors and popped in long enough to deploy vipers within range of a target (unless it was a critical conflict, which was dangerous to risk the ship on).

Star Wars
This was asymmetrical since the imperials had shorter-range fighters, but more battleship/carrier hybrid capital ships.  (It was always a pet peeve of mine you rarely ever saw rebel capital ships ever shoot even so much as an anti-fighter battery.)  Rebel fighters generally deployed at long range (a jump away or more) from the capitals and thee capitals mostly served to try to support the fighters (and stay away from the enemy capital ships).  Imperial capital ships would sometimes do jump-drop-jump maneuvers to deploy fighters, but they generally were on the hunt for high-value targets (rebel capitals).

Also wanted to talk for a bit about the "core loop" of the game.  To me the core loop of AIWC was:

• build defenses (to free resources for offense)
• scout targets (to give a target for offense)

[li\select target and payload (sometimes murderball, sometimes raid ships, etc.)[/li]
[li]strike[/li]
[li]extract gains (capturing planet, mining science)[/li]
[li]back to build defenses[/li][/list]

Waves give a pattern to that loop where you have windows to strike and windows where you need to defend (and also serve as a cue to continue on the loop).  Other strikes (like threatfleet) are major wrenches in your plans and keep the loop spicy.

In AIWC (post per-planet caps on turrets), losing planets had a few effects:

• it slowed you down while you rebuilt what you lost (costing you time and annoyance, although rebuilding turret wreckage eases that)
• it concerned you that the enemy was one step closer to your homeworld
• you might have to stop/delay an attack to deal with it
• it messes with your energy supply (possibly causing a chain reaction of other problems)

What translates to AIW2?  Not that they were one step closer to your homeworld unless your Ark was on the lam.  It might slow you down more if they recapture the planet (does it cost AIP to retake a planet that has been recaptured?).  Your fuel doesn't change too much unless you had a lot of power->fuel or metal->fuel converters on the planet, and even if you did it doesn't mess with your turrets on other planets so you don't have to worry about suddenly becoming vulnerable due to a power outage.

So yes, it seems like losing planets seems to be less of a concern in 2 than it was in classic, but that just means it is incumbant on the recapture feature (or some other new feature, like some way to reveal where your Ark is or increasing AIP) to increase the danger factor.

[etheric42]

Sorry for the doublepost, but I guess it took me a long time to write that and I wanted to respond to something Keith said:

"Game-Over Guns"  I think those would feel very ham-handed to the player because it feels less like a normal consequence to "exposing your king" and more like a "the developer will point a gun at you if you do what he doesn't want you to do".

A normal consequence, that has a similar motivating factor, would be: while the Ark is on an AI planet, all AI budget is shifted into spawning threat (a few hops out) that comes after the player. This is accompanied with a very angry-looking warning message on the UI.

It's like if someone playing GTA starts punching people on the sidewalk in front of the police station. They're going to learn pretty quickly that this isn't the thing to do. You don't want to be seen, so don't stand up.

So here, in the early game, it would mean a fairly quick death unless you immediately reverse. Later-game, it might be a useful way of preventing AI defensive growth, in exchange for taking a heavy hit against your defenses, but if you do it for any length of time you're going to die.

The game over guns don't have to be one-shot kills.  They could hurt enough that the Ark has a lifespan and the player either needs to silence them or bug out.  It should be slow enough that the player has time to think and try things out, but fast enough that their Ark shouldn't just facetank every planet (again, bomber raids to silence the big guns, maybe the Ark could have survived against them but not if the warden/hunter fleet had shown up).  Otherwise the Ark could have no guns and enough armor to ignore most attacks (so the attackers ignore it anyway unless it is the only target).

This also plays into loss conditions somewhat.  AIWC had a fair loss condition.  If you did not have enough defenses prepped to deal with the final attack, that save file was dead (you would need a significantly earlier save to deal with it) (except for certain circumstances like one-off surprise raiders).  In AIW2 many times losing an Ark just means loading a save game from 5 minutes earlier (which is really annoying in multiplayer).  Cheap Ark kills should be avoided because of the power of saves.  Other than overexciting the AIP, is there a legitimate "that save file is dead" condition in AIW2?

I want there to be a reason to deploy the Ark on an offensive, when it makes sense and would be impactful.

Dividing offensive forces is not generally done in AIWC unless resistance is very thin. The player does not usually have the attention for controlling two separate battles that can't be on the same screen, and concentration-of-force is always a dominant tactical consideration.

I agree with this, I like the idea of splitting offenses, but the tools for doing so are difficult unless you have multiple humans playing or are using one of the offenses as just a sacrifice to distract warden/hunters (and even then it is hard to make enough of a stir to assume they are drawn in while making sure enough units of the one they don't respond to are alive).  Some extra automation of your fleet ships would help in that aspect, but that could be a resource sink to develop and doesn't mean it will be done the way the player wants.

What about doing picture-in-picture to monitor the second planet and swap quickly?  Probably would be a clunky solution.

[keith.lamothe]

You just said the word hunt and hunters.  Maybe this is what the hunter fleets should really be doing.

As in AIWC, the threatfleet (hunter fleet) strongly prioritizes any opening it sees to get at the king unit. But generally the human will not leave themselves that open, and an AI victory is achieved through hitting the human wherever it can. So the hunters do that if they can't just go for the kill.

But yes, the "shift AI budget to hunt the Ark" mechanic would be most appropriate if it shifted the budget into donations to the hunter fleet. I don't know if it will be more or less responsive/effective than just dumping it into generating normal threat ships, so I didn't specify earlier.

Maybe the Ark needs to play a sensor game with the hunter fleets and hide its presence until it needs to commit.

I know it would be nice to be able to have an offensive use of the Ark, but for now I think it's going to have to just be the king unit and the unit that does the hacking (which is an offensive use of sorts). Maybe post-1.0 we can add late-game ways to upgrade your Ark into a main combatant, and then a kick-down-the-door combatant, etc. But in AIWC the King didn't need to be an offensive asset, and it doesn't need to be one here.

The "sensor game" would just be the basic "stay off the AI's planets". And perhaps it can also do a much-reduced version of the dump-to-hunters effect if your Ark is on a non-friendly (but non-AI) planet. One side-effect of this is that if the AI ever destroys the controller on your Ark's current planet, you'll start getting dogpiled by the hunters to turn a probable-loss into a quick-loss.

That's simple if a bit weird about forcing that decision in that way.

We could also have nearby friendlies take the damage that would have gone to the flagship, kind of like an inverted shield, but I figured the targeting logic would achieve a similar effect with way less cpu resources. 

What about increasing flagship range so it can sit in back?

The problem is not generally that your flagships get killed while attacking. They can sit out of the fight already to a large extent. But they'll still die pretty easily because the AI can get at them either through long-range stuff or pursuing forces. 

Or do some naval equivalence and make them functionally immune to attack except by equivalently sized weapons

That's basically what Shield vs Plasma is currently. Small stuff with anti-shield is pretty rare and fragile (Eyebots, etc), but Plasma Turrets and Plasma Guardians are a big problem for shields. Any serious non-early-game human fleet is going to want Shield Starships (preferably a Shield Flagship) to protect the vulnerable long-range ships like missile corvettes and siege starships.

The Assault Flagship is tanky and uses Armor instead of Shield, and the only weapons that get bonuses vs Armor are short-ranged. But the Assault also has to get close since its own weapons are short-ranged. Needler Turrets and Needler Guardians are a significant threat to it if they get close.

The Stealth Flagship is cloaked and uses Evasion instead of Armor/Shield, and isn't very tanky. So if it gets decloaked it's in trouble, but especially so against enemy Missile Corvettes/Turrets/Guardians.

The Siege and Sniper Flagships both lack any special defense and are not tanky at all, but can attack from range. This leaves them very vulnerable to Sniper stuff (which has a bonus against units without special defense), to the point that if they don't have shield cover they're probably going to die pretty quick because the AI usually has a fair amount of sniper stuff. Aside from snipers, these are incredibly vulnerable to Bombers, which also get a bonus vs Structure-type units. To use these you definitely need a fighter screen or other way to kill lots of bombers before they get close.

The Ark currently has Armor, but we could take that off to emphasize that it's vulnerable.

In summary, increase key ship HP (flagships/ark/dunno about golems) or maybe give them unique armor that only ships of that class have. Create classes for the AI that prey on the big ships (they were the ones that cracked the golems in the first place) (give them armor-piercing ammo that works differently than the traditional good/bad triangle, or is "super effective").  Give some stationary versions to the planets that can be removed by sorties.  Give bombers armor-piercing ammo too, but make them something you can take down en route to your capitals.

I would summarize the above as "that's already there", except that it's more varied than just having one "anti-big-stuff" mechanic. It's kind of a combination of the examples you gave later.

Increasing Flagship HP is probably a good idea, though. I think the Golems are ok, and the Ark probably needs to lose HP instead of gain it.

In AIWC (post per-planet caps on turrets), losing planets had a few effects:

• it slowed you down while you rebuilt what you lost (costing you time and annoyance, although rebuilding turret wreckage eases that)
• it concerned you that the enemy was one step closer to your homeworld
• you might have to stop/delay an attack to deal with it
• it messes with your energy supply (possibly causing a chain reaction of other problems)

What translates to AIW2?  Not that they were one step closer to your homeworld unless your Ark was on the lam.  It might slow you down more if they recapture the planet (does it cost AIP to retake a planet that has been recaptured?).  Your fuel doesn't change too much unless you had a lot of power->fuel or metal->fuel converters on the planet, and even if you did it doesn't mess with your turrets on other planets so you don't have to worry about suddenly becoming vulnerable due to a power outage.

It does still reduce your defensive depth even with the Ark being mobile.

Fuel balance is up in the air, but generally losing a planet's worth of fuel should be something the player would notice. You can't build up your fleet without enough fuel, and you can't win without a big enough fleet. So losing a planet means having to take it back or take a different one or you won't win.

But point taken on it not interfering with turrets. I do think we need to reverse on the "make offensive fleet and turrets use separate different population-cap mechanics", i.e. make turrets cost fuel. I don't think we want a fuel-shortage to disable turrets but the implicit inability to replace losses would serve a similar purpose.

I'd like to keep Power for non-weapon per-planet stuff like shields and tachyons, but we'll see.

The game over guns don't have to be one-shot kills.  They could hurt enough that the Ark has a lifespan and the player either needs to silence them or bug out.  It should be slow enough that the player has time to think and try things out, but fast enough that their Ark shouldn't just facetank every planet

With that list of caveats we've already got those: Sniper Turrets and Sniper Guardians. If we make the Ark's defense type Structure instead of Armor it will take 4x damage from snipers. And then its HP could be tuned as needed.

Other than overexciting the AIP, is there a legitimate "that save file is dead" condition in AIW2?

It's functionally the same as in AIWC: if you've got more anger coming down on you than you can fend off, you lose. You can mount a longer last-stand by moving the Ark, but the AI will keep chasing you until you defeat the attack or it kills you.

What about doing picture-in-picture to monitor the second planet and swap quickly?  Probably would be a clunky solution.

That would be pretty intense on the GPU, because it basically means doing 2x as much work to track the visual objects on both planets, etc. But we could probably do minimaps for multiple planets, once we figure out how to do a minimap

[BadgerBadger]

I would consider making Sniper Guardians much more rare. It feels like many/most planets have them, and they contribute significantly to the inability of smaller human fleets to wear down a planet. In AIWC you could tackle a strong planet piecemeal more efficiently. Perhaps this is contributing to the feeling that you need to fleetball.

[keith.lamothe]

I would consider making Sniper Guardians much more rare. It feels like many/most planets have them, and they contribute significantly to the inability of smaller human fleets to wear down a planet. In AIWC you could tackle a strong planet piecemeal more efficiently. Perhaps this is contributing to the feeling that you need to fleetball.

I don't know if you want to try that locally by itself, but you can change /Configuration/BuildMenu/KDL_AIDrawBags.xml:BasicGuardianDrawBag to have weight="9" on every item in the list except the snipers. That would automatically reduce sniper guardian population to about 10% of its current level.

If we wanted to keep that change we'd need to modify the SpecializedAIDrawBags files to maintain the desired target ratios, but that's only for games with specific AI types.

[x4000]

"Game-Over Guns" I think those would feel very ham-handed to the player because it feels less like a normal consequence to "exposing your king" and more like a "the developer will point a gun at you if you do what he doesn't want you to do".

An excellent point.  And agreed on the GTA approach to not punching people in front of the police station.  And on not giving the Ark good guns.

A new mechanic that is ever present could be that in order to capture one planet, you have to be suppressing all AI network nodes on adjacent AI planets.

An option that put those in some places would be interesting, but if that was the everywhere default many folks would just quit from exhaustion after a few planets. Fighting multiple offensive battles at once is very tricky.

Could be, yes.

If you don't give me some turrets from the start, I'm unlikely to build any. And it makes them seem less critical than they should be.

To a certain extent, if a player doesn't do something because it's not immediately doable, they'll just keep losing until they quit or learn to use the tools available. You can't really win either game without turrets.

That said, We can give them Mk1 Needler Turrets to start, and they'll be motivated to unlock more to get turrets with longer-range and/or are good against defenses other than Armor.

I just want to avoid the situation in AIWC where we added a few things to the "Starting set" and then a few more, and then a few more, until they wound up starting with (nearly?) every turret type and (nearly?) every starship type and 10,000 knowledge.

Understood.  I think that just giving them the needlers would work, though.  They need enough of a taste to go "ooh, this is handy," versus it being a completely unknown thing.  In AI War Classic we were essentially going "here's your starter kit," whereas here we want to be drug dealers giving "just a taste" of everything.

The per planet cap for turrets makes me iffy.

Again, it's worked in AIWC since 7.032 (nearly four years ago).

That's fair enough, but I'm mainly going based on what I see in this game at the moment.  Some of the other changes around here have made the territory a bit different.  Since we already have per-planet Energy here, having the cap increase by planet count held, but be distributable on any planet, seems both simpler to understand as a player (I see the cap on anything, and I can build it anywhere, universally for anything I see), and more consistent (not some strange exception for turrets; especially if the turrets category is being collapsed into one Defensive category).

We're establishing a pattern, so far, that extra ship cap can be gotten in a variety of different ways, but so far turrets are the only exception to the cap being usable anywhere you see feet (local resources permitting).

The difference is that they cost fuel (called energy in AIWC), where here they cost power. And actually they haven't even cost power for a while, because there was no motive to spend science upgrading something that was already bound by two separate per-planet resources. With the mark-level-upgrade simplification I could reintroduce the power cost, and each upgrade would simply increase their strength without increasing their power cost.

The hope with power was to avoid the player feeling that they couldn't build defenses without cutting into the size of their fleet. It also allows for per-planet customization because your tachyon coverage, shields, gravity generators, etc all cost power too.

I like the idea that these are costing power, but not an increasing amount per mark level given the mark level simp-lifiction coming.  And I agree that per-planet customization is a good thing, and the converters to create power from other sources (fuel, metal) make that even more complex of a decision if you really want to create a choke point.

I just think we have little to lose here from a bit more flexibility (cap goes up by planet, can build anywhere, but costs power wherever they are built), and we lose some bad-type complexity (turrets are per-planet but nothing else is, cap-wise?), and planets should still be plenty diversified with a non-increasing mark level power cost for these.

Maybe we just need to drop power and fuel and go back to energy as a single resource (we could keep calling it fuel). I'll miss the "power as a tachyon/gravity/etc shared budget" effect, but oh well.

I really really don't think that is the case, but I suppose we'll see.  I think that the above system would have a high chance of working out well.

Either way, we could make Ark-on-AI-planet cause AIP, but I'm thinking that it's better that the player only receive a temporary penalty (the AI sending tons of ships) that will kill the player quickly if it's going to kill them at all, rather than a permanent penalty that probably won't kill them soon (unless they really overdo it) but has a good chance of ruining their chance of winning long-term.

But it is true that if there's any one way we can get players to stop doing something, it's to assign an AIP cost to that thing. Unless it's warheads, of course, in which case players laugh at the cost (mostly; I think I can count the number of intentional nukes in serious play on one hand).

Permanent costs seem okay to me, but having things be time-based is something that I don't care for if it has a permanent cost.  Something that is time-based but which has increasing intensity (ala dealing with Super Terminals in Classic) was great, because it was "how long can you hold out while doing this."  Something that has a more permanent cost but that is also time-based seems negative to me.

The only way I'd be all that happy with it is having something like a small AIP cost every time your Ark enters an AI world; something like +1 AIP just for entering an AI planet, but then you can stay there with it as long as you like.  Send the Ark out of that planet and come back with the Ark, though, and it's another +1 AIP.  Then the costs are discrete and understandable.

The game over guns don't have to be one-shot kills.  They could hurt enough that the Ark has a lifespan and the player either needs to silence them or bug out.  It should be slow enough that the player has time to think and try things out, but fast enough that their Ark shouldn't just facetank every planet (again, bomber raids to silence the big guns, maybe the Ark could have survived against them but not if the warden/hunter fleet had shown up).  Otherwise the Ark could have no guns and enough armor to ignore most attacks (so the attackers ignore it anyway unless it is the only target).

Yeah, that would certainly improve the concept.

What about doing picture-in-picture to monitor the second planet and swap quickly?  Probably would be a clunky solution.

It would be a really really big overhaul under the hood to show more than one planet at a time in terms of true visuals.  We could show a tactical map of some variety of what's happening elsewhere, but not a true view very easily.

I know it would be nice to be able to have an offensive use of the Ark, but for now I think it's going to have to just be the king unit and the unit that does the hacking (which is an offensive use of sorts). Maybe post-1.0 we can add late-game ways to upgrade your Ark into a main combatant, and then a kick-down-the-door combatant, etc. But in AIWC the King didn't need to be an offensive asset, and it doesn't need to be one here.

The "sensor game" would just be the basic "stay off the AI's planets". And perhaps it can also do a much-reduced version of the dump-to-hunters effect if your Ark is on a non-friendly (but non-AI) planet. One side-effect of this is that if the AI ever destroys the controller on your Ark's current planet, you'll start getting dogpiled by the hunters to turn a probable-loss into a quick-loss.

I agree with this.

We could also have nearby friendlies take the damage that would have gone to the flagship, kind of like an inverted shield, but I figured the targeting logic would achieve a similar effect with way less cpu resources. 

I like that, too.  You could base it on the control group (ha!) rather than proximity, potentially.  That would be interesting.  Have to be on the same planet and in the same control group, and beyond that the damage spread is random.  Then targeting doesn't have to be cheesed.

Increasing Flagship HP is probably a good idea, though. I think the Golems are ok, and the Ark probably needs to lose HP instead of gain it.

I like these things.  I should feel afraid for my Ark, and powerful with my flagships.

But point taken on it not interfering with turrets. I do think we need to reverse on the "make offensive fleet and turrets use separate different population-cap mechanics", i.e. make turrets cost fuel. I don't think we want a fuel-shortage to disable turrets but the implicit inability to replace losses would serve a similar purpose.

I'd like to keep Power for non-weapon per-planet stuff like shields and tachyons, but we'll see.

I still have to say that I really like the separation of fuel and power.  I see the point of the other approach as well, and I could see it working, so I'm not married to the current idea.

But one of the things that was present in AIWC, that is not here, is the scattered things-to-defend all over your planets.  What if little "power collectors" were in various fixed locations on the planet?  Right now there are metal gather spots, but it's only like one per planet.  Having power collectors that would cause outages on the planet to all those defensive units would be great.  That way you have some choices at a planet:

1. Build up expensive metal or fuel to power converters to get some buffer power.
2. Just don't use at least one collector's worth of power in order to avoid losing your turrets and other defenses.
3. Protect those collectors very well, either by having point defenses or a very strong perimeter, and hope for the best.

We could make this more fair by having "rolling blackouts" of turrets and other forces based on how much of a power shortage there is.  Aka, don't completely disable turrets and shields, but have them potentially shut off some of them at random for 5-10 seconds, and then those pop back on and others pop off.  The ones that turn off are completely random and it just keeps turning them off until there's enough power to run whatever is left.  So you have turrets still active, but with a lot of chaos going on, increasing in intensity as more power generators are lost.

Or all the turrets just fire more slowly based on the percent shortage of power, or something like that.

The game over guns don't have to be one-shot kills.  They could hurt enough that the Ark has a lifespan and the player either needs to silence them or bug out.  It should be slow enough that the player has time to think and try things out, but fast enough that their Ark shouldn't just facetank every planet

With that list of caveats we've already got those: Sniper Turrets and Sniper Guardians. If we make the Ark's defense type Structure instead of Armor it will take 4x damage from snipers. And then its HP could be tuned as needed.

Ooh, then Ark should definitely be Structure.  That makes perfect sense anyhow.

I would consider making Sniper Guardians much more rare. It feels like many/most planets have them, and they contribute significantly to the inability of smaller human fleets to wear down a planet. In AIWC you could tackle a strong planet piecemeal more efficiently. Perhaps this is contributing to the feeling that you need to fleetball.

Going along with the above train of thought, what if we didn't make them any more rare at all, but we made them a lot less effective against non-structure, non-turret stuff?  So basically these things will WRECK your own planets if they get there, and do so from far range as well, but on AI planets they aren't that big a deal to the human fleet?

[BadgerBadger]

I also observe that the Special Forces get most of the AI resources (50% at all times). There's also very little given as reinforcements or waves by comparison. Keep in mind that most AIP is going to be in the >= 50 AIP range, which means that it's going to be spent mostly on Special Forces (50%) and CPA (37%). I think GetSpendingRatios() needs a balance pass (and I think the comment I put in there is incorrect). It's also a bit weird since you start at 20 AIP, so why bother with a 2.5 and 10 topOfStep? Those are going to be so inconsequential when you are at 200 AIP.