Brainstorming about offense vs defense, and refleeting speeds.

[keith.lamothe]

I think the main thing that needs to be true is: if you play your defenses right, you don't have to "come home" for waves, or at least not in a hurry.

The things to make that true are:
- Tractors and turrets need to be strong enough.
- There needs to be some defensive way of dealing with stuff out of range of your turrets (i.e. drone launchers or "carrier turrets" as Badger said).
- Waves need to give enough warning that, if you've misplayed your defenses, you can reasonably respond.

Ultimately if someone wants to scrap all their fleet ships to have them rebuild back home more quickly than they could move... I don't see that as an issue worth upheaving the design for. Some of those upheavals I'd like to see (fleet ships tied to the things that built them; starship-centric fleets, etc) but I don't think it would be good for the project.

I agree that increasing turret / tractor capabilities also solves the problem. It creates another one, though: once I put in place the "perfect" defense, that part of the game is "solved".

The point is that the "keep you on your toes" regular attacks (waves and residual threat) should not normally require your fleet's involvement.

CPAs, Exos, too-many-things-happening-at-once, etc can definitely require your fleet. And if your fleet isn't in the right place, that's one of the too-many-things.

Generally in AIWC the AI didn't win by waves. It won because of a combination of things you (and perhaps it) didn't anticipate would happen all at once.

[etheric42]

A part of the reason why the "scrap and rebuild elsewhere because it's faster" strategy works is because the economy on the whole feels stronger. If you can rebuild the fleet in two minutes and still have reserves do to that twice more then unless you can get your fleet there in two minutes it is a no brainer. If it took 5 minutes to build a fleet and left you with a balance of 0 metal (that in itself took 10 minutes to recover) then having your fleet take 5 minutes makes scrapping not as good of an option.

I think part of the reason why the economy feels stronger is that a) waiting for refleet is boring, b) we aren't spending huge amounts of our budget building a massive amount of per-planet turrets every time we take a planet.

In AIWC, when I took my fleet out and fought battles, I was constantly increasing my home defense; whenever I lost a unit in battle it would get rebuilt at home. So I always had some reserve of ships. In AIW2 though, all my losses are rebuilt to stay part of the offensive fleet, and those fleets will rebuild Everything. So my planets are always going to be weaker, since they don't have the (say) 10 bombers, 1 starship and 30 fighters that I lost in combat.

This is also my experience with AIWC.  If I lost half my ships in an attack, I suddenly had a half-fleet ready to move to intercept a wave.

4. With the most-scarce and non-replenishable resource in the game, Science, people are going to be particularly conservative and only pick things they know about because the stakes feel very high.  They will pick things that help them in the part of the game they feel is most important, aka attacking.  Who can stomach unlocking turrets when they're getting smacked around on offense already?  And by definition, if you're not getting smacked around on offense at least some, you aren't doing it right.

I agree.  I also didn't spend much science in AIWC on turrets except when I felt "guilty".  So maybe one turret pick for every 5 or so offense picks, which probably wasn't enough but I didn't really have any way of knowing.  And AIWC gave you so many turrets to start off with it was crazy building those per-planet caps, I had no idea when I needed more turrets (particularly when the AI is sometimes "manageable, manageable, manageable, ooops that's game over".  Thankfully since your Ark can be by the front line, you always can defend your Ark with your fleet.

But of course if beachheading were a more common tactic, then you'd be spending points on turrets because that improves your offense.  That's also the problem with buffing turrets too much: suddenly beachheading becomes hugely powerful.

Is there a need to actually have "Offense Knowledge" and "Defense knowledge" as seperate things? What about just having more cool things to spend the Knowledge on? How about a Carrier Turret that acts like a carrier starship in spawning drones on a planet when an attack comes? 

Keith was talking about giving power distribution nodes the ability to spawn drones.  We also had some back and forth about simplifying the defensive techs, which I think would help a lot.

2. MAYBE: All power-using units have a per-planet cap (I keep pushing for this, mainly for simplicity).

This was also discussed (or removing turret caps in favor of just using energy).  But this still has the science problem of "How do I know when to spend science on defenses?"  Especially since they are automated and you probably aren't watching them for many attacks.

I have been advocating for split knowledge between offense and defense for a while now, but keep getting shot down on that.  How much do you actually like that idea?

I think this is something that could be tried, but it feels a bit forced.  What if my defensive science isn't enough to defend still?  What if I desperately need offensive science because if I don't take another planet I'm stuck defending for the rest of the game?  What if I feel forced to beachhead on every attack because that is the only way to bring all my science upgrades with me?

(By the way Chris, thanks for walking down the fleet-based idea.  Let's talk about it again for an expansion or a spin-off game.  While I think it might work for AIW, and might even solve some of the problems, I'm a little shaken by the timeline.  If this was something easy to throw together as a prototype, it might be fun to play with, but there are enough other unfinished items that I wouldn't want to spend a ton of effort on the prototype.)

So, within the existing framework, here are three ideas on the defense issue that should help increase readability (I'm not suggesting all three at once):

• Give the player feedback.  Put a gauge on the science screen showing the ratio between how much has been spent on offense versus defense.  Color gradiate the bar to give some recommended guidelines.  The center of the bar is a positive color, the edges turn towards a negative color.  If you know what you are doing (or are desperate) feel free to skew your science spend.
• Play up the bend-don't-break.  When the AI takes a planet, have a chunk of that fleet "settle down".  The reduced next fleet should be able to be stopped by your next planet's defenses.  Have the science officer send a "helpful tip" message (voiced or unvoiced), that more science should be spent on defenses before the next wave.
• Merge defense and offense tech.  Pair up each turret with a type of mobile ship (or just fleet ship) (for example, bombers->needlers) and get rid of defensive tech altogether.  Whenever bombers are upgraded, needlers are upgraded as well.  Fluff it as the turrets are basically stripped down, weapon-focused versions of the fleet ship they are paired with.  Starships can either be paired with mega-turrets or some other kind of defense/economy structure.

If you had two different types of science then at minimum we'd need to get Eric to rethink the resource bar.

I'm up for sending an updated wireframe of the necessary screens if this is the way we go.  Not a fan of this way, but I can solve it.

[Magnus]

I still don't understand why the "per-constructor" cap is a bad idea.

My general problem is: I have a space empire. It starts small and becomes bigger. Sooner or later (and usually it's sooner) the war is going to become a multi-front affair, simply because you will have more than one border planet.
I want to be able to enact this simple decision: "I have 3 fronts right now. I want to allocate 15% of my empire resources to front 1, 25% to front 2, and 60% to front 3".

Right now, I can't. Not for the small ships. Which are your dime-a-dozen, most plentiful, most mobile resource, and the one which keeps getting continuosly recycled, and it's thus by far the most dynamic part of the game.

Why I can't? It's due to how constructor queues work.
Let's say I have a total galaxy-wide cap of 100 fighters. I want fleet 1, my attack fleet, to have 70 fighters, and fleet 2, the one I leave behind to help the turrets defending my border, to have the other 30.
I start the constructor on fleet 1 until it has built 70 fighters, then I pause it. I start the constructor on fleet 2, wait till it has built the other 30 fighters, then pause it. So far so good.
If fleet 1 enters combat, I unpause the constructor's queue. Every fleet 1 fighter which dies is replaced by the fleet 1 constructor. Good. Same goes for fleet 2.

Now let's say both enter combat at the same time. I unpause both constructors. Fleet 2 is defending well and not losing any fighter. Fleet 1 is in hard combat, and loses two fighters. Given how the queues work, one fighter will be rebuilt by fleet 1 constructor, one by fleet 2 constructor.

My ratio has just changed from 70-30 to 69-31 and what's worse, it has changed in the opposite way of what I'd have liked. The fleet in the hardest fight just got weaker.
The only way to "solve" this would be to jump back and forth between the two fights pausing and unpausing the queues as needed. That'd be insane.
Now imagine the same situation with 3 or 4 simultaneous fights.

Now if the game philosophy is "let's just give one big fleet to the human player, it's the way we did it in AI War 1 and going further would complicate things too much / make the game too different" I'll understand. But I think the decision has to be explicitly made in these terms, i.e.: do you want the human player to be able to efficiently manage more than one fleet? The rest comes from this: if you choose to give him more fleets, then turrets will only need to be used as complement since some of those fleets will be naturally used in a defensive capability.

[etheric42]

One of the proposed UI tools (and it may still be implemented) was building control groups to specs.  For example, control group 1 is 50 fighters, 30 bombers, control group 2 is 30 fighters, 20 bombers, etc. and then ships are replaced for the control group that gets damaged.

Some of the concern with that is it is an effort to put that together and people may still just cluster all their fleets into a single ball anyway.

Having per-constructor caps is kind of an enforced way of doing that control group rebuilding where you don't get to decide who has the big fleet and who has the small fleet.  That works too, but if once the ships are built they are independant, then they can still leave their mothership and join another fleet.  Also has the same "fleet going back to defend" issue where if you need all your fleet to attack, it's frustrating to send a fleet back to defend, and if you can attack with a partial fleet, then you could be overwhelming when you attack with all your fleet.  (Of course AIW has a bit of a fix for that in variable Mk targets... midgame you can attack Mk1-2 planets with partial fleets and Mk3-4 planets with everything you've got, leaving the homeland less defended.)

[x4000]

Merge defense and offense tech.  Pair up each turret with a type of mobile ship (or just fleet ship) (for example, bombers->needlers) and get rid of defensive tech altogether.  Whenever bombers are upgraded, needlers are upgraded as well.  Fluff it as the turrets are basically stripped down, weapon-focused versions of the fleet ship they are paired with.  Starships can either be paired with mega-turrets or some other kind of defense/economy structure.

I LOVE this idea.  It will probably require yet another revised Science screen, but probably a simpler one.  Instead of "unlock mark II bombers," it's "unlock bombs mark II" and hovering over that shows that it benefits ship types x, y, and z.  The underlying data structures already supports this, I am pretty sure.

I still don't understand why the "per-constructor" cap is a bad idea.

...

Now if the game philosophy is "let's just give one big fleet to the human player, it's the way we did it in AI War 1 and going further would complicate things too much / make the game too different" I'll understand. But I think the decision has to be explicitly made in these terms, i.e.: do you want the human player to be able to efficiently manage more than one fleet? The rest comes from this: if you choose to give him more fleets, then turrets will only need to be used as complement since some of those fleets will be naturally used in a defensive capability.

We have two main problems.

1. If we stray too far from the original game, then this ceases to "be AI War" in the minds of a lot of people.  For me, this happened between SupCom 1 and 2.  That said, we're certainly not shy about changing a lot of mechanics, so it all comes down to capturing the feel of the first game, not the exact mechanics.  This isn't a re-skin.

2. Time.  We're headed for Early Access on the 26th of this month come hell or high water, and we'd like it to be in a solid state for that.  One of the big things that killed Stars Beyond Reach was endless complicated re-prototyping, and I don't want to go down that route again.  When things seem to risky with this game, then for now we're falling back on the first game for inspiration wherever possible.  We always have room for expansions and other post-release stuff that makes this more involved in other ways later on.  Right now we just need a rock-solid core.


Personally I am still a fan of per-constructor caps, or something to that effect.  The goals you mention are things I, too, want to see happen.  But given the two constraints about, none of the ideas that I or Eric have advanced are yet fitting within those constraints.  It's entirely possible that there's a revision -- or entirely different way of going about it -- that would satisfy the two conditions as well as getting the per-constructor stuff.  We just haven't thought of it yet, if that's the case.

[Magnus]

I totally understand where you're coming from.

Possible idea: a queue setting where constructors won't rebuild ships destroyed outside their planets?

[ptarth]

Perhaps shift the balance more towards how the AI deals with the game in terms of Fleet Management?

• All Ships are owned by a Planet, except the Ark, which is its own owner/planet.
• Each Planet has one Fleet flagship.
• The Fleet Flagship and all its assigned Fleet Ships is a Active Fleet.
• Fleet Ships in a gravity-well, and not assigned to an Active Fleet, are the Defensive Fleet
• A Defensive Fleet can never leave its gravity-well, but the Fleet Ships can be transfered to a different Planet (becoming part of that Planet's Defensive Fleet.
• Gravity well-bound Constructors are owned by the Planet in their gravity well and all Fleet Ships built by that constructor are also owned by that Planet.
• Non-Gravity well-bound Mobile Constructors are owned by a Fleet and therefore a Planet and all Fleet Ships built by that constructor are also owned by that Planet.
• Active Fleets have a desired Ship Type and Count. The Planet will build these ships (priority first in/first out) to reach this cap.
• Defense Fleets have a desired Ship Type and Count. The Planet will build these ships (priority first in/first out, higher priority that Active Fleet. (maybe a setting)).
• Fleet Ship capacity limits are by Planet (possibly separate caps for Active and Defensive Fleets).
• Active and Defensive Fleets have capacity limits. Extra ships built over fleet capacity remain at Planet in the Defensive Fleet.
• Active Fleets that have exceeded the capacity limit do something to limit cheese.
• The Ark has a greater Fleet Capacity than a regular Fleet Flagship.
• Starships grant extra fleet Capacity to the Active Fleet it is assigned to.
• Gravity-well bound Active Fleet Ships do not count against the Defense Fleet Capacity limits.
• Each Planet has a fast-travel link to its respective Fleet Flagship, and vice versa. (e.g., Sins of a Solar Empire - Antorak Marauder
• The fast-travel link has some limitations especially in hostile gravity wells.
• The Fleet Flagships and Arks cannot Fast Travel themselves, they must always take the long way around.
• Starships might get to Fast Travel.
• Each Active Fleet Flagship, Ark, and Planet have a setting for Reinforcement Target (Active Fleet, Ark Fleet, Defense Fleet) and a Reinforcement Travel Method, by direct wormhole travel, fast travel, or never (Fleet ships stay in the gravity well they are built in).
• Two Flagships in the same system can swap Active Fleet Ships.
• Two Planets can swap Defensive Fleet Ships.
• Individual Fleet ships can be sent manually to their Active Fleet Flagship or back to their Defensive Fleet at the Planet.
• StarShips can be assigned to Fleet Flagships or split off independently. When assigned to a flagship they get all of the Flagship bonuses (e.g., Fast Travel).
• When an Active Fleet Flagship is destroyed, Fleet Ships cannot be sent to the Active Fleet.
• When an Active Fleet Flagship is destroyed, Fleet Ships in the Active Fleet are rooted to that planetary well.
• When an Active Fleet Flagship is destroyed, Fleet Ships in transit either go back to their Planet or continue on to the Active Fleet.
• The wreck of an Active Fleet Flagships can be rebuilt by another Active Fleet Flagship.
• The core of an Active Fleet Flagship can be recovered via something (Ocean's 11 Style Heist?), so you won't ever permanently lose one.
• If you lose a planet, that fleet has X minutes to be transfered to another Active or Defensive Fleet before all of the ships join the AI.

Strengths

• A clear separation between Offensive and Defense.
• Clear methods to reinforce and transfer ships between Offense and Defense.
• Transfering ships between Offense and Defense still requires Time and Resources.
• Removes the advantage of Scrap and Rebuild, its faster just to shift existing ships to where they need to be.
• Main Offensive Fleet will be the Ark, but additional Active Fleets can also be used up to the number of currently held Planets.
• If you retreat your entire Main Offensive Fleet to defend, you have to leave your Ark Behind and Alone.
• Extra Fun if you lose a planet.

The Research Problem

• Either link offense and defense research together, or separate them so there is no competition for resources.
• Example of link: Space Fighter 1 and Basic Turret 1 are the same technology.
• Example of Separation: When getting science points you get X offense points and the same X defense points. You spend your offensive points on offense tech and your defense on defensive tech.
• I favor linking. A Fighter is just like a tiny turret with a smaller engine. So it is natural you'd improve both offense and defense with the same technology upgrade.

[x4000]

I totally understand where you're coming from.

Possible idea: a queue setting where constructors won't rebuild ships destroyed outside their planets?

That... has a lot of merit.  I'm a bit stuck on what happens when your population cap increases or you want to scrap some units and have them build elsewhere, though.  Hmmm.

[Magnus]

I totally understand where you're coming from.

Possible idea: a queue setting where constructors won't rebuild ships destroyed outside their planets?

That... has a lot of merit.  I'm a bit stuck on what happens when your population cap increases or you want to scrap some units and have them build elsewhere, though.  Hmmm.

Use it as an additional "pause" setting. If the setting is active, that constructor queueing function is only going to be called from the small_ship.cleanup_after_getting_destroyed() method of a small ship located in the same planet (assuming you're using OO programming about which, of course, I have no clue ).

Basically, if you want to build "new" units, you need to have it unchecked. Just like the pause button.

[x4000]

@ptarth:  That is... wow. 

I have to say, it looks more complicated than it is, in the main, but at the same time it has a LOT of divergences from the core flow of the game that this is a sequel to.  On the research problem, I think we're agreed on those being joined up.

When it comes to your suggestion, I feel like there are too many little edge cases and potential GUI snafus in there.  How do I know which fleet a ship is assigned to, for instance?  And a whole host of other display problems that I see going along with that.

But there's a lot that I like, so let's see if we can distill some elements out of that which are less radical but get at the same ideas:

------------

1. Okay, so first of all, we're all lately talking about "techs don't upgrade one ship, they upgrade an offensive and a defensive one."

2. Secondly, this thing of fast-travel keeps coming up.  Overall I really don't like it... EXCEPT I really liked how you mentioned that basically there was a conduit to/from a planet and its flagship (in your design where every planet has a flagship), and that conduit was only good for the fleet ships and maybe starships.  This solves the Small World problem I was concerned about, while allowing for reasonable logistics.  Having it feel like your flagships are stuck way out behind enemy lines, but can shuttle smaller guys back and forth is... cool. 

3. The idea of having flagships that are associated with planets is something that appeals to me.  Not... exactly that thing, but something close.  I'm not sure what feels wrong about one flagship per planet, but it rubs me slightly wrong at the moment.  That seems too overpowered, for some reason.  I love the idea of these mobile conduit ships, though.

4. For me... if we had conduits along the lines of -- hang on, I just had a brainwave.

---

The AI is always warping stuff in from outside the galaxy, right?  Or at least from "places unseen."  What if the players could do that, too, only better?

1. Make it so that only space docks can build fleet ships.

2. Make it so that building fleet ships is a lot slower.

3. Let players build however many space docks per planet, like piling on more engineers in AIWC.  But have it cost a lot of power there.

4. Have constructed ships either be set to "pop out here" or "pop out at x flagship or the Ark" or "wait in storage."

5. At any flagship or your Ark, you can immediately -- from the Ops tab -- dispense any ships from storage, at that location.  Boom, fast travel.  Dispense exactly the number you want.

6. The ability to scrap ships is removed.  Instead, when a ship would be scrapped, it is simply sent to storage and you get nothing back.  It's ready to be redeployed somewhere else, instantly if need be, but they're in the same state they were before (if they're shot-up, or engines-out, fix that before sending them to storage).

7. Pulling more from Magnus than anyone else: On the build tab, or something of that nature, on each planet have a thing that says "if there are fewer than n of unit x on this planet, dispense from storage until the quota is met."  And another rule, optional, that says "if there are more than m of unit x on this planet, put back in storage until we're no longer over-filled."

Something like that?

[x4000]

Possible idea: a queue setting where constructors won't rebuild ships destroyed outside their planets?

That... has a lot of merit.  I'm a bit stuck on what happens when your population cap increases or you want to scrap some units and have them build elsewhere, though.  Hmmm.

Use it as an additional "pause" setting. If the setting is active, that constructor queueing function is only going to be called from the small_ship.cleanup_after_getting_destroyed() method of a small ship located in the same planet (assuming you're using OO programming about which, of course, I have no clue ).

Basically, if you want to build "new" units, you need to have it unchecked. Just like the pause button.

Huh!  That... would be easy to implement, interface-wise, and easy to understand I think.  New players might get themselves into a bit of trouble with it, but not too badly.

[etheric42]

I mentioned something like this before, with the Ark building ships and then warping them to your flagships and creating temporary wormholes for your smaller-than-flagships to use.  I didn't think about just putting ships into storage and leaving them there on call.  I'd rebrand "scrap" into "dock" and have it not instantly destroy/store the ships but cause them to path to the nearest flagship/ark for docking.

The construction yards gets around the problem of how do you scale up/down your build speeds as the game progresses, but people might ask why you can't warp ships to the construction yards.  You could also tie construction speed to a research tech.

This also does three things: gives something for the Ark to do (it's how you respond defensively), reinforces the AIW2 idea that the player is mobile guerrillas (as opposed to AIWC where you had to defend the home planet and various permanently destroyed things), and flagships now also fulfill the AIWC role of transports for deep striking (just not disposable).

I like this, but I think doing some digesting/sanity-checking to make sure this isn't as different as the fleet-focus or per-builder caps should be done.

Edit: if we decide to go through with this, I have some ideas for some UI improvements to facilitate it (simple stuff using the styles we already have, nothing major), let me know and I can mock up the design.

[x4000]

I want feedback from Keith, first, and to give chemical_art and Magnus and ptarth and Badger and so on time to pour water on it if need be, too.

But so far, so good!

[Magnus]

A lot of things to digest from ptarth's proposal. It would certainly change a lot of things.

A couple of problems I can immediately see:

- one flagship per planet = way too many especially on big maps. Since every flagship is basically a fleet by itself due to the warp-in ability, you would gain the ability to warp your whole small fleet to any planet of your liking as long as it's within your empire.
- one (or more) docks per planet = easy way around the "building small ships is slow" limitation. And after conquering 10 or so planets, why would I ever want to build more than one dock on a planet? Better to keep the energy for turrets, since I can insta-teleport the whole fleet wherever I want anyway as long as there's a flagship there, so it's not like I really need the ships to be built in a specific location.
- general problem: "whohoo I can insta-warp my whole fleet at enemy planet X" = "hey let's develop some cheese tactic where I manage to reach the core A.I. planet with at least a flagship then boom --> instakill A.I.!"


For flagship caps: in practice, 99% of the time you don't really need to move your fleet to any arbitrary point. You need your fleet(s):

 - on the planet you're attacking
 - on the planet where the A.I. is attacking you

Let's assume you want two simultaneous attacks ("take this, Warden fleet!") and you're forced to defend on two planets either because the A.I. blitzed past your outer border and/or attacked in two spots at once. That's still 4 fleets and thus, 4 locations. Any more than that and it's probably overkill for the capability of your average player. You can probably get away with increasing the maximum theoretical number to 8, but above that it would be something that very few would ever use.

[x4000]

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.

At the moment, the things I'd prefer to focus on the most are making things more usable for players, and making defenses more valid in general.  To that end, the minimum amount of work we could do to get in that direction would probably be:

1. The change to having techs tied together to unlock one offensive and one defensive thing together, rather than having to choose.  That's super simple to understand as a player, and nobody seems to have any huge beef with it.

2. Working on balance in general with defense, partly under the new "now more turrets are available" banner but also just in general so that the fleet doesn't have to return from the front lines all the time.

3. Then there's an open question about whether flagships should actually be able to produce fleetships or not.  Maybe it should just be space docks.  And on the space dock, maybe there's a little button that is something like "spend speed" that you click to be paused, 1x, 2x, 3x, 4x, etc.  So how fast that dock is going to work to make things.  It's basically just like stacking engineers in the first game.  With a system like this, you could have just a couple of space docks ever running in the galaxy, if you can afford that sort of throughput.  And then pause the rest when you don't need any more of whatever.

4. It has been mentioned several times that a lack of speed bonus on friendly planets is sad-making.  I kind of agree with this.  If ships on our own planets got a speed boost, that would help both in defense as well as in getting ships logistically from my space docks to the front lines or whatever.

5. Some sort of mechanism for handling rebuilding of local losses does need to happen, but I don't know exactly how that would work.  Either that, or we need some power-based mobile ships that are bound to the planet they are on (like all other power-based stuff).  Diesel truck, meet electric car.  Fueled bomber, meet powered bomber.  That would basically help to provide some of that stated variety that chemical_art was requesting in defenses, it would let us customize little mini-fleets on our planets along with turrets as desired on our planets, and it wouldn't interfere with offense.  Maybe these things are only drones, I don't know.  Drone turrets, after all?