Salvage/Reprisal in the sequel

[keith.lamothe]

(Meant to do a post on the overall balance approach next, but someone else prompted me to think about Salvage and the ideas attacked...)

Salvage/Reprisal in Classic was aimed at one main goal: keeping the momentum going. Put more precisely, the point is to avoid periods of the game where all three of the below are true:
- The player is unable to effectively attack.
- The player is relatively safe.
- These are both true for a long time.

Typically this happened when the player fleetwiped:
- Having no fleet, the player could not put together enough mobile force to take a serious objective.
- Having lost zero defenses, the player could generally still weather any storms from the AI.
- Due to the cost of a full fleet, the refleeting process took a long time.

So I added a mechanic where the player losing ships on AI planets gave the AI "metal" (not really, but kinda) to launch reprisal waves that had to attack the player (couldn't wait like other threat), and the AI losing ships on player planets gave the player Metal. Kind of a ping-pong match, in theory, as a better alternative to staring at the table (or Netflix) for an hour.

And in practice it worked for some and not for others. In general it was a collection of gamey rules aimed at a particular result, rather than an element of the setting with intuitive consequences that help keep things moving.

So for the sequel here's what I'd like to do instead.

The Plan

1) When any unit dies, it leaves some kind of visible debris.

2) When you have an Ark, Flagship, or Controller (command station) on a planet with debris, it hoovers them up (infinite range, but starting with closest); resulting in the debris disappearing and you gaining metal.

3) Similarly, if the AI has a Controller or Usurper (the unit it uses to reconquer planets) or something like that on a planet with debris, it hoovers them up; giving the AI points to put towards one or more strength budgets. Probably the "Reconquest" budget so it will be used offensively, but not necessarily immediately thrown against your wall.

The Goal

When a side wins a battle, they gain momentum to press the opponent. That way, no matter who won, somebody has the momentum and things don't have to just sit still for a long time.

The Caveats

The AI being able to gain momentum from killing stuff in your territory is a balance challenge.
- But I think together we're up to it.

Uncollected debris may need to disappear at some point (time, or sheer count of debris objects on the planet) to avoid performance problems.
-  But I'm not sure about that. It can be tracked as a very different sort of thing than ships so that it basically only increases cpu time when being drawn (and our graphics engine can handle scads of totally inert objects) or when a hoover-er is on the planet looking for some (which means the amount is actively being reduced already).

"Remains" and the "Remains Rebuilding" function of Ark/Flagship/Controllers are still there, and your units will always rebuild rather than hoover your own rebuildable debris (so turrets, not ships), but if the AI has hoovered the actual debris it will cost more to rebuild the turret/whatever.

Some units like your Flagships are already designed to actually start as dead units that you repair into useful condition, and if they "die" they just go back to that state so you can lose them temporarily but still have a chance to rescue them. These units will not drop debris or be otherwise impacted by this mechanic.

I'm not sure if there will be any difference in the rate-of-hoovering between the Ark, Flagship, and Controller. My inclination is to have the Ark be somewhat faster than a Flagship, and a Flagship somewhat faster than a Controller. I like people having motives to involve the less-expendable (or completely non-expendable) units. But it would need to be balanced to avoid the feel that you're gimped if your Ark isn't there.
- This wouldn't be as bad since you'd generally be able to collect all the debris eventually either way (as opposed to just losing out on it in Classic), unless the AI was had a collector there to compete with you for it.

I don't think I'd have multiple collector units on a planet "stack" in gathering the debris faster, so as to not always motivate you to have all your flagships (and even the Ark) all on one planet during a big battle.

The amount of metal salvaged would not be the full price of the ship, for obvious thematic and balance reasons. But this "efficiency" would not vary based on what you used to collect it.

Further Opportunities

I think we could bring back the Parasite/Reclamation mechanic by building on top of this, specifically:
- have "parasite" ships tag their targets with "reclamation damage", as before
- but when the ship dies with enough reclamation damage to reclaim it, it doesn't respawn on your side as in classic; instead it becomes debris like normal but with a "can be reclaimed by the owner" flag on it
- and if you salvage that debris, instead of getting the metal it respawns the unit there for you to use. If you don't have the available ship cap for them you just get the metal like normal.

This is easier to balance because there's a limit on how _fast_ you can actually spawn those reclaimed ships, you have to have a collector present, and there's a tradeoff (you don't get the metal).

Other reclamation-like mechanics from classic could also piggy-back on this:
- The Botnet's Zombie-reclamation
- The Saboteur's "hulk explosion" effect
- The Reprocessor's effect could just increase the efficiency of the eventual salvage


Thoughts?

[tadrinth]

I like that the ark/flagship/controller difference is just speed instead of efficiency.  Might contribute to micro where you move your ark to your wave-soaking planets between waves, to vaccuum faster.

I think the previous salvage system would have benefitted from regular waves giving less salvage and exos and reprisal waves giving more salvage.  I almost think the AI should have a particular unit that is somewhat scary, but drops a very large amount of debris.  Then regular waves never include those, while reprisal and exo waves tend to have a lot of them. 

Salvage also becomes a LOT more impactful at high difficulties under the old system since your econ doesn't scale with difficulty but the enemy ship counts and salvage scale up pretty sharply.  If the salvage from the AI is mostly concentrated in special units, then their numbers can be relatively static across difficulties. That seems like it might help a lot with tuning.  Under the new system, this might not be an issue; you won't be able to vaccuum stuff up any faster at higher difficulties, so at some point your chokes will just get saturated with debris, and you'll want to either have more chokes (to spread the debris across more Controllers) or to bring up the Ark or a Flagship to help vaccuum.

Oh, and you should definitely have Dire Guardians wandering around AI space, and they should drop *ludicrous* amounts of debris.  So you can go whale-hunting. 

[kasnavada]

I like this better than the old salvage, but I've got huge balance concerns about the snowballing effect.

One remark in any case, metal amounts shouldn't scale to the difficulty level. Capturing one planet should enable you to rebuild "X" cap of something no matter whether it's low or high diff. If so it would be much easier to balance. I think that at higher diff, you should need to recycle many more units to regain the same quantities of stuff - also to make it so this mechanic matters at every diff level.

[Draco18s]

Generally looking at this and nodding going, "Mhm, mhm."

The reclaimation mechanic made me perk up and go "huh, that'd be different/interesting/unique."

[keith.lamothe]

Salvage also becomes a LOT more impactful at high difficulties under the old system since your econ doesn't scale with difficulty but the enemy ship counts and salvage scale up pretty sharply.

One remark in any case, metal amounts shouldn't scale to the difficulty level.

Good point, I don't know why the size of that disparity didn't occur to me before.

I'm not fond of having salvage-from-the-AI be artificially concentrated into certain units, as I like the idea of this being an intuitive thing (exploding ships leave floating metal) rather than obvious "this is what the game designer wants me to do" signs.

But the idea of simply scaling player salvage down at the higher difficulties makes sense.

Under the new system, this might not be an issue; you won't be able to vaccuum stuff up any faster at higher difficulties, so at some point your chokes will just get saturated with debris, and you'll want to either have more chokes (to spread the debris across more Controllers) or to bring up the Ark or a Flagship to help vaccuum.

I don't figure the vacuum rates would be slow enough to not have enough time between, say, normal waves to clean it all up.

But if there's one constant in AIW, it's the player ability to surprise me with the sheer size of the glowing glass craters you create.

Oh, and you should definitely have Dire Guardians wandering around AI space, and they should drop *ludicrous* amounts of debris.  So you can go whale-hunting.

That does sound fun.

I'm sure someone would transform the Special Forces into a Metal Delivery Service.

I like this better than the old salvage, but I've got huge balance concerns about the snowballing effect.

In theory I agree; in practice I don't think it would turn out that way.

There are four general outcomes from an attack on an individual planet:

1) Player on offense, player wins.
2) Player on offense, AI wins.
3) AI on offense, player wins,
4) AI on offense, AI wins.

#1 seems unlikely to be problematic, becuase the AI has oceans of defense-in-depth, and the player can't store infinite metal. A player could get on an unusually long "run" if they planned ahead which objectives to follow through with and was aggressive about using the revolving resources, but I don't see that as a bad thing. Eventually the player will need to regroup.

#2 and #3 aren't really snowball situations; they're a momentum switches.

#4 is the potential problem, where the AI taking a few of your planets gives it enough extra momentum to turn a small territorial gain into something that cripples or kills you. I confess that it takes me some effort to see the problem with that, but I'll take your word for it

There are a number of ways we can keep AI-snowballs under control. One is to delay the entry of any "collected from non-AI planets" salvage into the AI econ by X minutes.

[Draco18s]

But if there's one constant in AIW, it's the player ability to surprise me with the sheer size of the glowing glass craters you create.

Have you seen the 100% real physics space combat game on steam?
Look at the discussions, there's a thread on absurd player creations, one guy made a 844 gigaton nuke. Which, as he puts it, kills every fleet there is.

Another guy made 1 gigawatt laser with a 21m radius aperture that kills capital ships instantly at a range of 250km (and dies to a stray, small caliber round).

Another guy bugged the simulation to make a rocket launcher that had a reload speed of 180 picoseconds. Of course, it wasn't usable at those speeds, and eventually got it to slow down to 160ms, housed it in a drone with 4x100 missiles, housed those in a carrier (with more launchers), grand total of something like 4700 1kT nuclear missiles.

Edit:
Now that I'm not mobile:
http://store.steampowered.com/app/476530/
and
http://steamcommunity.com/app/476530/discussions/0/343787920116425693/

[Lord Of Nothing]

I like this quite a lot. It seems to fix quite a few of the problems with the original salvage mechanic that helps push it closer to it's intended role and away from stuff like the wave-recieving homeworlds. I agree on reduction with difficulty to keep it at about the same or only slowly increasing amount of rebuild-power, this seems very sensible.

I'm not so sure about faster collection rates for the ark and flagships, personally, though, but let's see how feels.

I guess a question this brings up is how expensive buildpower will be, since this is rather closely linked in with the player's ability to snowball past even significant defence in depth- if you can keep up with the income, storage caps become irrelevant.

The AI snowball issue, on the other hand, is interesting. I guess at some point the AI might have to start diverting it to either some sort of rainy day fund, or to something nasty that will try to kill you in a different way that will be just as hard to counter as the straight reprisal, maybe even harder, but will be slow enough that you have some time to stabilise and rebuild yourself first.

[Tridus]

Further Opportunities

I think we could bring back the Parasite/Reclamation mechanic by building on top of this, specifically:
- have "parasite" ships tag their targets with "reclamation damage", as before
- but when the ship dies with enough reclamation damage to reclaim it, it doesn't respawn on your side as in classic; instead it becomes debris like normal but with a "can be reclaimed by the owner" flag on it
- and if you salvage that debris, instead of getting the metal it respawns the unit there for you to use. If you don't have the available ship cap for them you just get the metal like normal.

This is easier to balance because there's a limit on how _fast_ you can actually spawn those reclaimed ships, you have to have a collector present, and there's a tradeoff (you don't get the metal).

Other reclamation-like mechanics from classic could also piggy-back on this:
- The Botnet's Zombie-reclamation
- The Saboteur's "hulk explosion" effect
- The Reprocessor's effect could just increase the efficiency of the eventual salvage

I like this, but if this is the case, Reclamation needs to be prioritized over just salvage collection. I don't want to fail to reclaim bombers because my shipyard is off building them while I salvage metal instead. Or something to keep me from losing out because I didn't shut off my shipyard kind of situation, because getting an entire ship is better than getting a portion of it's cost back. We don't want to unintentionally introduce annoying micro.

[keith.lamothe]

Have you seen the 100% real physics space combat game on steam?

Simply beautiful.

I guess a question this brings up is how expensive buildpower will be, since this is rather closely linked in with the player's ability to snowball past even significant defence in depth- if you can keep up with the income, storage caps become irrelevant.

I'm not sure about the cost of buildpower yet. My general thought is that your Ark and Flagships will be able to crank out ships at a modest pace, while your stationary dedicated building facilities will be able to crank them out at a fairly insane speed if you have the Metal for it.

On the one hand, I don't want the lack of engineers to mean a lack of ability to speedbuild. On the other hand, I'm not sure "speedbuild by default" is a good idea. Probably best to go with the suggestion of a UI element (slider, perhaps) for controlling the build rate of a dock. So its default speed is respectable, and you can specifically set it to "insane" if desired, but you'll know why it's suddenly vaporizing your economy.

I like this, but if this is the case, Reclamation needs to be prioritized over just salvage collection.

Yea; when the debris is created it already knows whether it's reclaimable, so it can be put in a separate list which collectors check before the non-parasited debris list.

[kasnavada]

There are four general outcomes from an attack on an individual planet:

1) Player on offense, player wins.
2) Player on offense, AI wins.
3) AI on offense, player wins,
4) AI on offense, AI wins.

#1 seems unlikely to be problematic, becuase the AI has oceans of defense-in-depth, and the player can't store infinite metal. A player could get on an unusually long "run" if they planned ahead which objectives to follow through with and was aggressive about using the revolving resources, but I don't see that as a bad thing. Eventually the player will need to regroup.

#2 and #3 aren't really snowball situations; they're a momentum switches.

#4 is the potential problem, where the AI taking a few of your planets gives it enough extra momentum to turn a small territorial gain into something that cripples or kills you. I confess that it takes me some effort to see the problem with that, but I'll take your word for it

I have a slightly different assessment...

1) Player on offense, player wins.
=> gets metal for winning, partial fleet lost easily rebuilt, refleeting to minimum, working as designed.

2) Player on offense, AI wins.
=> No refleeting, potential lost entire fleet situation. High possibility of entering into 4. Possibly highly hard on newbies - if they misjudged one enemy unit or defense and get fleet wipe, it going to be difficult to get back from that.

Basically it means that the AI will be the next one to attack, and this time it's situation 3 or 4. In case of a retreat with partial fleet, defense or partial fleet should be ok and giving "sane" amount of defenses, but in the case of full wipe, it's a very high chance of snowball starting - but even if not, it's netflix time until the AI attacks. Which I'd prefer not to have.

3) AI on offense, player wins,
=> Potential issue, depending on how the defenses are done comparing to AI War I. Because on AI War I the player will mostly not lose even against very large odds.
I think the game needs lesser defenses - my opinion only -  in order for this to actually not happen most of the time. There was a discussion on reducing the "power difference" between attack and defense which talks about this at depth.

4) AI on offense, AI wins.
Working as intended, I guess. It's complete snowball, but then again, the AI was too passive so basically I don't have much of an issue with that.


Basing on this, I think situation 2 needs command similar to a "recycle entire planet" command, which scrapes an entire planet in order for the player to be ready to repulse the next attack with a full fleet. Hopefully, leading into 3 without too much risk of going into 4. Either makes it into metal or into special "emergency" defense ships. Other ideas could be done, but a giant "panic button" with some kind of drawback sounds cool to me.

[WolfWhiteFire]

Well players can no longer set up chokepoint planets that all AI attacks have to pass through by maximizing all your defenses, since each planet now has a limited supply instead of there being an overall supply that can be distributed anywhere, so you can set up chokepoint planets but not fortify them as heavily. Also personally if balancing is needed I would rather AI waves be made stronger and bigger than defenses being weaker, that way we get to see more ships explode or explode us.

[keith.lamothe]

Well players can no longer set up chokepoint planets that all AI attacks have to pass through by maximizing all your defenses, since each planet now has a limited supply instead of there being an overall supply that can be distributed anywhere, so you can set up chokepoint planets but not fortify them as heavily. Also personally if balancing is needed I would rather AI waves be made stronger and bigger than defenses being weaker, that way we get to see more ships explode or explode us.

Bear in mind that finite power per planet doesn't mean there won't be major chokepoints. You'll just need to "work with the grain" of the map by using the planets with naturally high power, etc.

But yea, the idea is to avoid things getting quite so out of hand in the sense of "I can stop every conceivable attack within this 100 cubic meters of space around this one wormhole".

[Squashyhex]

I think we could bring back the Parasite/Reclamation mechanic by building on top of this, specifically:
- have "parasite" ships tag their targets with "reclamation damage", as before
- but when the ship dies with enough reclamation damage to reclaim it, it doesn't respawn on your side as in classic; instead it becomes debris like normal but with a "can be reclaimed by the owner" flag on it
- and if you salvage that debris, instead of getting the metal it respawns the unit there for you to use. If you don't have the available ship cap for them you just get the metal like normal.

This is easier to balance because there's a limit on how _fast_ you can actually spawn those reclaimed ships, you have to have a collector present, and there's a tradeoff (you don't get the metal).

Other reclamation-like mechanics from classic could also piggy-back on this:
- The Botnet's Zombie-reclamation
- The Saboteur's "hulk explosion" effect
- The Reprocessor's effect could just increase the efficiency of the eventual salvage

I love this idea for reclamation. It doesn't instantly give you ships, as in AIWC, so its more like neutralising the ship without damaging it. Maybe would need another name then? I dunno. But the concept seems pretty sound

[Atepa]

Hrmm…
1) Would it need to leave debris? Is our ‘collector’ ship going from piece to piece to pick it up, because it sounds like they would just slowly vanish the pieces (starting with those closest to them). If it is just vanishing them, why make a visual cue at all? It could potentially just be a counter, although that is less pretty too.
2) For your goal, what about ties? I mean yes, absolutely there will be conflicts where there are a clear winner and loser in the outcome, but how would this type of mechanic work for fights where one side wins a pyrrhic victory. Just whomever got to the system first to collect the debris?
3) If this sort of change would let us keep things like parasite ships and reclamation, then I’m definitely all of it. There is nothing I enjoy more than reclaiming a bunch of AI ships and using those to hit the next system while the portion of my fleet that died is rebuilding.

2) Player on offense, AI wins.
=> No refleeting, potential lost entire fleet situation. High possibility of entering into 4. Possibly highly hard on newbies - if they misjudged one enemy unit or defense and get fleet wipe, it going to be difficult to get back from that.
Basically it means that the AI will be the next one to attack, and this time it's situation 3 or 4. In case of a retreat with partial fleet, defense or partial fleet should be ok and giving "sane" amount of defenses, but in the case of full wipe, it's a very high chance of snowball starting - but even if not, it's netflix time until the AI attacks. Which I'd prefer not to have.

I’m not seeing the problem with that… a newbie shouldn’t realistically WIN their first game, heck they shouldn’t win their first few games while they are learning the mechanics, unless they are playing on a difficulty so low that the AI literally just sits there the whole game and does nothing strategy wise. What lesson does a new person learn from the scenario you played out? 1) Start learning to recognize a failing attack and retreat. 2) More knowledge on the enemy unit or defense that destroyed them so that they can start formulating a way to counter it (assuming they survive)

Newbies stop being newbies by pushing them to get better / stronger. Yes you have to push lightly, but they still need to accept that they will not always win if they ever intend on playing the game at a decent difficulty level.

[MaxAstro]

Situation #2 is definitely the point of concern for me, because it's the time I find myself bored most often.

In situation 1, I usually don't need to refleet; in situation 3, I have plenty of salvage, and in situation 4 I'm far too panicked about how screwed I am to be bored.

The boredom comes when situation 2 hits and the AI does nothing to capitalize on that.  Even with the salvage mechanic in AI War I, there are many times that I lose my fleet, but I have enough defenses to hold against the reprisal wave - but maybe not fully, so the reprisal money goes into turrets - and then I'm sitting around refleeting.  Or the reprisal wave ends up being a ways out, and I'm sitting around waiting for it.

It would be nice if the AI was a lot more aggressive about identifying situation #2 and trying to kick you when you are down.  You just made a major tactical blunder: You lost your entire fleet on a failed attack.  The fact that you sent such a large fleet should worry the AI, and the fact that you lost it should tell it you are vulnerable.  It should be pouring resources into making you pay for that, or at least giving it a good try.

The quicker we get from #2 to either #3 or #4 the less boring downtime there is.  Sure, that's a bit rough.  But I agree with Atepa: The faster and more directly you punish someone for making a mistake, the quicker they will learn not to make that mistake.

My level of caring about this topic is high enough that for my "design an AI" thingy, I'm going to be suggesting an AI that is basically entirely focused on addressing punishing the player for overreaching/making mistakes, because that's the AI I want to fight.