Magpie & AA & Air

Alright, it's time for that yearly balance thread.

I'll get straight to the point.

Magpie is fundamentally badly designed unit

Let's look at it's strengths and weaknesses:
Strengths:
Excellent speed
Good maneuverability
Great HP/cost
Perfect accuracy
Homing projectile
Fast projectile speed
Good range
Weaknesses:
Absolutely terrible damage/cost

How did we end up with a unit that has only one weakness? And it's not even a weakness can be used to kill it.

Magpie is effective against any unit, as long as it's valuable, punishing potential interesting gameplay
Imagine this scenario:
You run a raid of 25 Glaives. You manage to evade the enemy's defenses. You enter the enemy's core territory: rich, full of eco and undefended.
Now - imagine a unit's value - it's based on the unit's state and position. A raider in the enemy base is worth much more that one in the factory. Magpie's role, due to it's bad damage/cost and perfect accuracy, is to fire at units that are the most valuable.
So your Glaives just die to Magpies, preventing any interesting gameplay.
With Raven you could have outran the projectile.
With Thunderbird you can spread out or wait for the disarm to run out.
With Phoenix you could at least have spread out your units. (I do think this problem affects Phoenix as well, it's another unit that's faster then raiders and counters them)
With Magpie you just die. The missiles home towards your units and there is nothing you can do. There is way to limit your losses or even kill it without AA.
This can happen with any game move that increases the value of your units, unless you have a swarm of AA to protect you.

How to counter Magpie?
Don't even bother with flex AA, Magpie both outranges it and outmaneuvers it.
There is only one answer - AA. And not a reasonable amount of AA. If you want to deny Magpie from attacking your forces, you need a stupid massive amount of AA. The issue lies here: Lanchester's square law

Zero-k has many game mechanics to reduce the effect of Lanchester's square law on the game, but none of them apply to AA and Magpie. Both just deal damage. There are no AA units that deal status damage and there are only few that do AoE. Ettin kinda works, but good players will spread out their Magpies. Only Artemis does enough AoE to truly stop Magpie. I believe Magpie was originally intended to be a counter to Lance (or other arty), but it fails to be one since Lance is usually co-located with massive amounts of AA and other targets are not.

But in most cases it's just two damage streams competing. And as modeled by Lanchester's square law, the bigger force will do overwhelming damage on the smaller force. And Magpie is faster then AA, so it can choose where to engage.

Because of this,a Magpie swarm might just be the best counter to mobile AA. You would think that ground units would be the counter to mobile AA, but how would they do that? They can't catch them. Crasher is faster then Scorcher. But a Magpie swarm can actually catch them. A large enough Magpie swarm can counter reasonably-sized AA forces. (I fully believe that in practice most mobile AA units die from player inattentiveness.)

Also, the AA units built by the arms race that Magpie forces counter every other plane/gunship as well, and it counters them better then Magpie. So building Magpie discourages building any air unit that isn't Magpie. There is no way to counter Magpie specifically.

Conclusion
I hope the arguments presented here make logical sense to people. Fixing this, though, will not be easy.
I hope the game developers carefully reexamine units with the unholy combination of high range and fast speed.
Magpie, sadly I don't think is salvageable in it's current state. Nerfing will not fix it's fundamental problems. I believe it would be ideal if it was fundamentally reworked.

I am not sure how you reach the assessment magpie has terrible damage per cost because it's the second most efficient bomber in the factory:

damage per metal:
magpie: 1.71
raven: 2.67
likho: 1
phoenix 1.38 +0.34 aftertburn

Likho is terrible on paper because it does exactly 2000 damage for 2000 metal, but it does it in one punch which often makes the difference between piercing a shield and being blocked by it. Magpie will be reliably blocked by sufficient shields, while likho will pierce half depleted shields and that is a very important distinction. A stack of aegis/aspis that has been previously pressured will still efficiently block a volley of magpies, while a likho will just punch right through.

Phoenix is pretty awful. The damage listed above is purely theorical because it assumes all firebombs hit the target when really, when do they ever? The idea is that they saturate fire an area and the weak targets will burn. That's its job.

Magpie has better homing than raven because it can reliably bomb raiders, but doing so is usually a poor use of it. The reload time makes it an expensive "riot" option. It's just convenient that it can catch raiders at all, providing an extra safety net should you need it as some raiders slip through a crack in the front.

The magpie slightly outranges the likho (550 to 500) but the likho has way better AOE (96 vs 24) and homing (164 vs 82).

The one thing magpie is exceptionally bad at, and used to be worse in the past, is rearm cost. It's 0.057s per metal, compared to the raven's 0.0267 and the likho's 0.013.

It also doesn't have exceptional speed by any means. Magpie moves at 252elmos/s while likho at 270 and raven at 246. Magpie barely, barely outruns raven but by so little that if I didn't tell you that and you just glanced, you'd think they move the exact same speed. I don't know why you think magpie moves quickly. It does not. Not exceptionally fast for a plane anyways. The swift moves at 390, the raptor at 228 (lol it can't even catch up to bombers despite being purely a response unit), thunderbird at 270, owl at 360 and sparrow at 210 without the detonate boost. What makes you say magpie is fast? I am genuinely curious.


Ravens cannot hit moving raiders at all. Likhos are too expensive to just open with. Magpies provide safety at first and you can group them in small squads to snipe early artil like emis and lances. I don't know why people keep making them beyond like 40, which is already a lot. I think that's too much and I can ususally think of better ways to spend metal than making even more magpies at that point. Usually I plate a tank fac and get emis or something. Or plate shieldbots and snitch shieldballs... whatever high risk low cost strategy that I think can be more beneficial than afking 20k's worth of magpies only to send them into artemis half way through the game. Just make 2 artmis. On most map that covers pretty much the entire base. That's 4.8k to stop a seemingly infinite number of magpies. Limit stockpile of each artemis to 10 so at most that will cost you an extra 800/per... a relatively small investment to 100% counter a monospammer and it will also shut down owls, which is priceless. The blind team tends to lose.

All AA counters magpies effectively. It's just that no aa outside artemis counters a sudden ball of 15k+metal's worth of magpies for a tenth of the cost, and why would it? Why would a ground aa unit counter thousands of times its own cost?

The question you need to ask yourself is how do games get to a point where a player can mass 15k+ of mostly idle magpies without their team falling behind and then suddently wake up and decide to do something with it? That's not game mechanics. That's just disfuctional TAW behaviors. TAW makes no sense because the vast majority of players in TAW never learned to do antying other than TAW stuff, and they explain it away saying that "teams is a different skill". I wish I could find it but around a week ago there was a lobby in which 2-3 high ranked players were facing 7-8 low ranks with equalcom set to 1. It went exactly as expected. The low ranked players didn't know how to handle raider phase, were reduced to 1/3 of the map instantly and then the high ranked players could essentially win any way they saw fit. It happened in every of the 5 or so games they played. That's what TAW teaches, and it's all TAW teaches.

I covered that. None of those costs are hidden.

Lanchester's Square Law doesn't apply to Magpies within a single engagement, because they shoot once and need to rearm. No amount of scaling up Magpies will let the same Magpie fire twice in the same bombing run. Contrast, say, Harpies, where Lanchester's Square Law does apply because the first ones to die grant the others more firing time before they too get shot down.

One practical application of this is that more spread engagements favor the AA over the Magpie, so a Magpie flock spreading out to avoid even the small AoE of a Tarantula or Vandal is receiving extra shots by doing so.

And if the Magpies are making repeated runs at the same player's AA squad, that player can choose to prioritize building more AA to keep the collapse at the end of Lanchester's Square Law from occurring.

petition to rename magpie to pigeon

I think there are three separate questions here which it is easy to muddle up.

(1) How, as a Zero-K player, should you play against Magpies to maximise your chances of victory?

I think there are four main paths to success against Magpie. In loose order of preference:

(a) Go kill them early with Swifts. If you kill their first bunch of Magpies they will have minimal game impact for a very long time, and you can transition to a more efficient bomber like Thunderbird or Phoenix and have game-winning impact while they pick up the pieces. Unfortunately not viable if their team provides them with a bunch of anti-air cover from the start. A representative replay: B2456693 16 on IncultaV2

(b) A moderate number of Magpies can be beaten with the right kinds of mobile anti-air. Ettin is best in class, several others are serviceable, some are bad. However, if the Magpies get to make a direct attack run against your AA it may not go well for you. Try to hide your AA at least from radar so they don't know which radar dots are dangerous. Ideally use cloaks and shields. I don't have an immediate replay coming to mind for this one.

(c) If the Magpies reach critical mass, make and protect Artemises. Artemises are pretty annoying to protect, but if you do then this will solve your Magpie problem. Representative replay: B2459631 20 on Comet Catcher Remake 1.8

(d) Make an army that is fairly resistant to Magpies (lots of HP, shields, and/or cloak) and just go kill their land teammates. This is hard to make work against a truly overwhelming number of Magpies, but hopefully by that point you can revert to plan (c). Representative replay: B2444093 8 on Isidis crack 1.1 (sort of, this example is not especially clean)

(2) Is Magpie overpowered?

This is a more complicated question than it looks because it depends on context, but my answer is "mostly not". I think my personal winrate against them is fairly good, at any rate.

I think that a blue or purple rated player would struggle to maintain their rating if Magpie was a large part of their repertoire. More often than not, teams will build their army compositions around countering the opposing high-rated players, ESPECIALLY air players in my experience, and I don't think Magpie can cope with that.

Magpie's early game tempo is generally pretty bad. This matters much more the smaller the teams get and the more skillful the average player in the room is. I can certainly imagine that in a large low-rated team game played by a bunch of newbies Magpie could be super oppressive. I lack data to substantiate or reject that hypothesis. I don't recall Magpie ever looking at all viable in Palladium. I think that Magpie makes it harder to keep Zero-K fun across a wide range of team and map sizes.

Like a few other things (cough Bertha) Magpie is probably stronger on maps which I choose to dodge.

(3) Does Magpie make Zero-K a more fun game?

I suppose the answer to this is inherently personal. Personally I don't enjoy what the existence of Magpie does to Zero-K.

I think that fundamentally Magpie makes the "slippery slope" of getting ahead or falling behind unpleasantly slippery. If you are ahead, the losses you take from Magpie can often be absorbed by your greater metal income and you just go on winning without even bothering to directly contest against the Magpies. If you are behind, you can't keep Artemises alive and nothing else is efficient enough to fight against a metal deficit.

All sorts of clever things you used to be able to do by sneaking a few raiders into an undefended corner of the map are rendered outright infeasible by Magpie. Even the quick-response gunships like Harpy and Locust take more time to arrive on the scene and clean up.

Most other units in Zero-K are hyper-countered by some stuff, and in turn hyper-counter some stuff themselves. With the exception of Artemis (which is, again, hard to keep alive) Magpie doesn't really have much in the way of hyper-counter matchups in either direction. In that sense I think it is discordant with the rest of Zero-K's design.

Several things which a new player would think OUGHT to be effective against Magpies actually aren't, or at any rate they need to be quite carefully controlled. Swifts need to avoid head-on engagements against armed Magpies. Mobile anti-air needs to hide. Most of the static AA doesn't have adequate DPS and/or range to compensate for the fact that it cannot be hidden. As I have said above, Magpie is quite beatable - but you need to know how and be in a position to do it. Being put under pressure by the opposing purple, and then getting blasted off the map by some random bronze spamming Magpies, is a very unpleasant way to lose a game.

Besides playing against Magpies, I also don't enjoy having a Magpie player on MY team. It generally signals that I am about to suffer through an outnumbered frontline battle while the Magpie player sits at the back for ages building up. An air player opening with Phoenix, Thunderbird or even Raven will likely be much more helpful to me in the early game.

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

This all leads to the question

(4) What could be done with Magpie to make Zero-K a better game?

to which I personally have no better answers than "delete it". My feeling is that the fundamental concept of a cheap long-range bomber is unsalvageable, though I am willing to be proved wrong.

Hmmmm.

I think here's how I would balance it while giving it a unique spot.

I think currently there's not much air that works in hard AA. Everything kinda just dies to it.

Maybe make some changes to allow it to have this feature, but weak its impact on the game generally.

- Make it cost substantially more more. 500-1000.
- Increase its armour.
- Increase potentially increase its speed or range it fires from.

Effectively low, the significantly increased cost should make it not effective for what it's being used for atm, which is to kinda just kill everything.

If those changes were made, you might have a unit good for sniping high value units with low HP, but not great at other things. Swarms might be effective, but only in long game.

I think it boils down mostly to being not interactive enough. It's not really a balance thing.
Theres just not too much you can do on the receiving end of a magpie strike compared to say a raven/phoenix/tb strike where riots can still shred a few planes.

Most of that discontent comes from the combination of survivability in the form of range and difficult to avoid damage.
Take away the range and you have raven again, which is why I'm interested in pushing for making magpie to be dodgeable/unguided.


In the Refumble mod I've implemented a variant of magpie that acts as a tossbomber that one can dodge, trading it's tracking for some aoe.
That profile makes that magpie variant bad at picking off the starting leaks (Swifts job), worse at sniping single heavy targets (Ravens job), but decently good at grinding down dense/squishy armies.

Ofc that needs more testing too, but I think a plane focusing on sortie througput and airpad logistics, rather than the single strikes all other bombers focus on is interesting.

Ground being better at AA than air itself makes a whole lot of sense to me.

If you make a flying unit that is so potent at AA it dwarfs any other kind of AA, all you're doing is making sure nobody ever builds any kind of flying unit ever again because, if they do, they are met with the uberflyingaa. Uberflyingaa is countered by none other than uberflyingaa, so the whole air game becomes make more uberflyingaa than your opponent or don't make any air at all. Boring.

We saw this years ago when scallops were super op and outranged and outdamaged all other forms of sub and sea. Water was just who made the most scallops. It's boring, it's bad and it should never be repeated ever again.

quote: The problem is because they are so tanky

Here is the heatlh per metal ratio of all the planes:

swift 2
magpie 3.52
phoenix 2.69
raven 3.33
odin 3.47
spawrrow 1.52
raptor 3.67
likho 1.18
thunderbird 2.91
owl 2.79

Of all the criticisms of the magpie so far this is the only one I agree with. Until I bothered to calculate these I didn't realize just how durable the magpie is for a spammable unit. It's closer to an assault in health ratio than... what is it supposed to be? A flying skirm (aka flying scalpel)? It already has range for safety. Maybe it doesn't need that much health? I always thought bombers were meant to be used as hit and run. You're supposed to send them where you can get kills without retaliation and get out of there before AA can be mobilized, and lose them if you make the wrong call and stumble upon undetected aa like unscouted artemis, cloaked ground units... If they weren't meant to be played like that they would do less damage per shot and didn't need to rearm, so they'd be used like gunships. Having to rearm planes is a very ZK/C&C thing. Most RTS do not do this. Bombers are just other regular units that linger in the battlefield firing continuously.

I'd be totally on board with magpies having a skirm's health ratio rather than an assault's. Heck, raven is an assault and it has less. Phoenix is slower than both magpie and raven, and it has even less health. Why? I'd also be for removing all homing from raven and doubling its health, but I'm probably going to be alone on that one.

You could also reduce magpie range, but then it just becomes a worse phoenix/raven so you might as well delete the unit at that point.

If going for major redesing, then I really like the laser plane using the sparrow model from future wars. I don't know if it's shaman's original idea, but the plane has a battery. It fires lasers continously but not instantly so it depletes its battery over time which means it is exposed to enemy fire for a prolonged period, then flies back to a pad to recharge the battery. This is less alpha-strike-y and it gives the enemy time to respond as well as increases the effectiveness of the response.

quote: Magpie's early game tempo is generally pretty bad. This matters much more the smaller the teams get and the more skillful the average player in the room is. I can certainly imagine that in a large low-rated team game played by a bunch of newbies Magpie could be super oppressive. I lack data to substantiate or reject that hypothesis.

From my anecdotal experience, this is spot on. Depending on team-size and competence, Magpies are either super oppressive or borderline useless. The only thing I'd like to add is that the oppressive state often extends to mid-level play, not just to newbie games.