SUGGESTIONS: Gunship Balance

HARPY seems like it is in an acceptable place. It fills the role of an early-game "bully" that picks on stragglers and assists allies in scuffles. It's not sturdy enough to survive midgame AA but it can be reused as a a defensive support unit. It has some strange relationships with other units, like being effective in A2A dogfights against Swifts and even occasionally Raptors. It is on a tighter clock to get value than most air units generally are, as VERY modest AA is quite threatening to it.

LOCUST is somewhat weak/situational, and it should remain that way so that Gunship plops do not become overly oppressive on certain maps. That said, I think players who can devote attention to it while raiding and scouting with it should be better rewarded for doing so than they are now. In team games, the biggest risk with making a Locust as a Scout/Raider as early as possible is its vulnerability to Swifts, especially deep in enemy territory. Its ideal targets seem to be Wind Generators and Mexes.

NIMBUS is also somewhat weak/situational, especially in the sense that it compares poorly to Revenant. Revenant is also an inaccurate high damage unit that prefers a lot of the same targets, but in exchange for lower range, Revenant has much better speed, health, burst damage, and AoE potential that make it more reliable at pushing through AA and finding value than Nimbus. Like Locust, it's probably for the best that Nimbus, in its current role, remains situational and easily countered for the health of the game. It struggles against mobile AA more than Revenant because it's unlikely to ever hit mobile units, but just like Revenant, "Mass Nimbus" can punish some poorly-supported static AA configurations.


---Suggestions ---

Harpy: 1200 Health -> 1250 Health
Harpy: Adjust its Attack-Move behavior, such that it "follows" units more effectively, and turns from its targets less.

Harpy is a very juicy target for the recently-buffed Tarantula, as it's now killed in 4 shots instead of 5. Several Harpies can spend limited time inside of AA if it helps to accomplish a task, but that risk isn't worth taking against Tarantula. A small health buff may help it in this regard without breaking Quant's Rule, as it's not particularly defined by "fragility." It's worth noting this is also a large buff vs Angler.

Harpy ALSO has an issue with following its targets, where it turns away before getting into firing range, repeatedly. Even when Harpy (on A-Move) is retreating from ITS targets, it has a strange propensity to turn away and then back toward the target, moving inefficiently. According to CharmedMage, it may have something to do with its "SkirmLeeway" value or some similar value, but I'm not sure.


Locust: Firing Arc 150 -> 180
Locust: Vision 500 -> 560

These are suggestions to slightly improve Locust's tactical autonomy as a scout and a raider. A greater firing arc would allow it to deal slightly more damage when repeatedly engaging and disengaging vulnerable targets like Wind Generators and Mexes. More vision range would allow it to less often "run face-first" into unexpected threats.


Nimbus: Vision 660 -> 760 (or even 800?)

This is a significant buff... Many potential adjustments to Nimbus threaten to bend Quant's Rule, as many of its flaws seem to be necessary forms of awkwardness expected in "Artillery" units, which Nimbus essentially functions as. A vision buff, however, may mesh well with the unit's identity and give it a rare form of supportive value. This is a relatively large buff to creating 1 or 2 Nimbuses, without giving much extra value to "Mass Nimbus." This buff would allow a Nimbus to fire at great ranges via its own vision, especially when it's on a high elevation; of course, the greater the range, the greater Nimbus's inaccuracy and the worse the impact of its slow projectile speed.

Harpy can be extremely oppressive early on because of the slow damage. Eating 4 tarantula shots is actually quite a lot when you factor in that (talking about team-games) you rarely have the ressources to build that much aa in the first few minutes of the game.
I am not going to argue about the attack-move behaviour. The thing is that I expect harpy to be very strong even without more hp, so I am not that sure about raising it's health.

quote: In team games, the biggest risk with making a Locust as a Scout/Raider as early as possible is its vulnerability to Swifts, especially deep in enemy territory. Its ideal targets seem to be Wind Generators and Mexes.

And that one is just raider in general, although I agree that locust is underwhelming.


Overall, Gunships atthis point are an absolute staple in teams, mainly because of revenant and krow. Even tho I rarely get to play it,
I think GS is one of the coolest facs in this game. I think harpy, nimbus and locust are not so popular mainly because they are just overshadowed by rev and krow.
So, I am not arguing about the proposed changes, what I say is that the factory demands some attention beyond what is layed down here.
Mainly because too powerful units/factories can make the game a frustrating experience.

As said, I am not opposed to the changes. Just worried that they will push the factory a bit too far overall.

I like these Ideas, well reasoned and could add a lot to the factories game feel. Might be interesting to look into some other minor buffs for other gs that don't overwhelm breakpoints while improving feel and utility.
I have been considering very minor tweaks for gnat and locust primarily relating to breaking and acceleration speeds, for reasons of balance im primarily looking at increases in breaking power.
Additionally, thoughts on maybe something like a 30 hp increase on Gnat HP, mainly to keep its break point for ogre being able to maintain its one burst capability while helping make up for chip damage.

I found this due to the latest patch mentioning this.

As a lob that just recently "learned" to play a bit better and in no small part thanks to GS, I think the fac as a whole needs some extra consideration. For example I love the idea of Locusts as, well locusts, but I find that they lose to many matchups. They suffer against AA and many other units while not being exactly cheap. Maybe they could be made slightly frailer but more spammable or receive some buff so not everything rekts them? Honestly IDK enough of balance to say what the solution is. But locusts are rarely used in the games I have been (at least team ones).

In fact, I think that probably the factory with less used units is GS? While certainly there are units that see little action I think that GS in general only uses like 3 units often: Harpy early on, Nimbus and Revenant. Yes there is the Krow, Gnats and Blasties, but the first is more of an strider and the later are a very rare sight...

I dont think transports see any use besides the odd hercules to kidnap units or rescue a comm (and ofc there is ship herding when ppl want the funsies).

This are just my thoughts on this from an usage and fun perspective. As I said I am more on the lob side and cant really speak with solid knowledge in terms of balance but the usage rate of many GSs likely says something.

I apreciate the extra links here about modifying the game as it piques my interest on checking stuff. I think I will start to try and check these too soon, mostly to get a better feeling for stuff. Thanks!

I feel that locusts share the same characteristics as hermits and a couple of other units where in small numbers or as part of a mixed force they are not very impressive, but as a monospam army of many thousands of metal they can be very scary.

Since they are somewhat weak in small numbers it's unintuitive that they would be great in large numbers, which is why only blues and purples use armies of hermits and only Blxssom uses armies of locusts.

B2498478 6 on Scaryland v1.02

Do we agree that potato's com in the isolated alcove would have died to any raider in that metal value? Two players' worth of raiders (bandits + locust) vs one engicom. After that ninse sent his engicom too far ahead alone and died to one player's worth of raiders while trying to back out and the locusts didn't even participate in that fight. I am not sure that match proves anything about locust, other than maybe ground raiders might have blown up with the enemy engicom while locusts did not?

Losing the first com was a pretty hard blow to both income and expansion potential.

Map presence was a thing though. I'll concede that ground raiders might have taken more time to reach coolcat's engicom pushing on the opposite side alcove at the end of the match, but I think there was already a significant lead for south at that point so anything in that metal difference would have done the trick.



B2498497 8 on Fairyland 1.31

First encounter is glaives vs locust deep into south territory (glaive aggression) and the trade is nowhere near even as the locust eat dirt (520 lost in glaives to 660 lost in locust) while the locust ball value is nearly double and even has some bandit support at the end.

At 4:19 locusts kill an engicom and lose 900 value for the privilege. Again, considering how far exposed the engicom was, same value of any raider could have done exactly the same.

At 5:45, 1900 worth of locust kill 300ish worth of glaives after running from them in the first place and letting the glaives destroy a factory, 2 mexes and a light amount of other stuff. It's literally on top of an ally factory, so any unit could have responded to the glaive threat.

At 8:00 the locust engage a shieldball backed by gremlins protecting mid map. The locust eat dirt and kill nothing. South dies. I don't see what this match shows about what locust did. I do wonder though if that locust ball from 5:00 to 8:00 would have spent raiding lightly protected expansion from the edge of the map on the east edge, how would north have responded? At that point the pretty much only aa north had was 8 or so gremlins that happened to be where the locust were all the time. What if locust had used its flight to hit mexes instead and clean right edge of the map? I also wonder what if north's plane had gotten 2 or so raptors... just about the only use case where raptors are quite good. The 3 early swifts did nothing to the locust other than lightly tickle and then endanger themselves.


B2498511 8 on LimeQuat Republic v1.24

Locust spend the first 3 minutes flying around the enemy base and being a nusance. They don't actually kill anything other than a half built mex, but to the enemy team they probably feel like a significant threat. The locust retreat without losses and a 1.05 attrition damage ratio. Psychological damage?

Ending at 4:30 there was a locust raid that took out a few mexes but made the attrition damage&kill fall from 1 to 0.7 while the spectators comment that run looks scary. Mexes were sniped, but nothing else died and a copious amount of reclaim was left behind.

At 5:15 coolcat with a couple of pyros and blxssom with locust push on and kill a guardian com reinforced by 3 venoms. The com dies yet the push loses about equal value. It's the same part of the map the locust were at previously, so flight and speed don't really explain it. It's just two players pushing with greater value upon a lightly protected com.


At this point east is 2k metal produced behind, I think entirely explained by the fact that coolcat (with locust support) is facing off vs gounson and is controlling a supermex while east lost the equivalent supermext up north, not because of locust but rather due to spiders I think? I missed it I was looking at the locust.

At 6:00 blxssom catches out of position bolas mid map and restores his attrition kills to 1.0.

Nothing at all happens with locust until 10:00. Potato punches a hole through the north front and Blxssom piggibacks where the resistance is 1 grizzly and 1 razor. 5k value in locuts vs 2.2. East was around 4k metal army deficit before the dive and multiple fronts are suddently contested at the same time with overwhelming force on each. Steelblue vs doctorche, coolcat vs gounson, potato vs chonkgoose and potato's recluse thinning out chonk's amphs pretty much the entire match. It's an all around disaster for east at this point. I don't think the locust did anything because they were locust here. Blxssom just picked encounters that were very favorable due to opportunity I think. I think Bucky has the right of it.


B2498525 6 on Desert Rumble 1.02

The first 4 minutes, the locust essentially spent time making a big blunder and then getting pickoffs at the center of the map to get back to 1.0 attrition kill. Blunder was flying over 3xgremlins+glaives, and then the regrouped locust killing back stray glaives.

At 5:00 the locust dive on the southwest corner. blxssom loses one locust and kills around 1k in value. Out of position and lower in value daggers dive in after mexes and llt are lost. It's likely flying had a good part to play in the success of the raid, although pulling off the same thing with land raiders was not impossible. Case in point, north loses their southwest corner, but takes northeast at the same time with glaive and venom.

At 6:08 locust fly into north's base. 3.3k value in locust (+1 pyro) vs 1.3k in various defenses that include 4 gremlins and 3 glaives. The sheer value difference makes the locust victorious, but considering a pyro made it in with the ball, there was opportunity for land units to do the same. After this successful dive the locust escape through a basically empty midfield. Blxssom is at 1.31 attrition kill and 0.82 attrition damage.

At 7:00, after having flown from mid to northeast, the locust pick off a few stray glaives from the northeast push. They got there probably faster than any other raider would, but also a player could have just split their ground forces a bit to do the same thing.

By 8:00, metal produced is basically equal. Army value and attrition difference is around 800. Worthy of note at this point however is that Blxssom is 1500 value killed ahead of the second highest. So far if there is one match to watch to see if the mobility of locust can deliver good resuts if used well, this would be the one. I still think land raiders could have done the same, but it's certainly easier to consistently find good opportunities if you can fly. The MO is very much still to hit where there is essentially no opposition at all.

At 8:30, 3k of locust dives into 1.3k's worth of a mix of gremlin and glaives. They lose equal value, meaning that the locust swarm survives with 1.9k value. It's a killing blow because there is nothing left to check the locust and there are 2 pyros on top of the enemy factory cutting off all production.

There are a couple of good chockes on desert rumble on which a snitch/imp could be planted and catch land raiders. Locust get to just fly in and out safely. The one thing I would say about locust in this match is that I think land raiders could have done the same, but locust are safe from the counters land raiders are subject to while AA of equal cost to those counters does not provide the same level of protection. I also still wonder if 2-3+ raptors would be enough to shut down locust. The caveat there being, as always, that specialized AA only counters air while locust are free to shoot at anything.


B2498527 6 on AlienDesert

At 3:10 locust dive on a very exposed engicom already pressured by rovers.
Nothing interesting happens. I'm not sure why this match is mentionned.

Raptors suck >_< >_< >_<. Every time I try I regret lol. Very very situational unit and unfortunately one of those cases is countering the genuinely okay anti air in swifts.

My vibe with locusts so far is that there probably needs to be more testing.

They are oppressive, but not necessarily game winning.

I am yet to see competently wielded Swift fail to shut Locusts out of a game.

I think Raptor is an okay unit, but I do not think it is where you want to be against Locusts unless you are VERY confident in your Raptor control.

I have not seen Trident but I expect it is too slow.

A lot of the power of Locust is that because it is so mobile it can always find the weakest link on a team. Whoever built the fewest things that can shoot up effectively. If your team does not have Swifts to rapidly reinforce the weakest link, then your weakest link needs to be quite strong.

Nice replays and detailed analysis Steel_Blue and Galamesh

I am not at the point of seeing that something needs to be done. It is neat that even Locust can be picked up and made to look good. That said, the first game made me do this: https://github.com/ZeroK-RTS/Zero-K/commit/0bb60a1de42d4b1da283c1c7f548c7cae9fdf95c

quote: Push Locust target priority below that of raiders.

I noticed that Locust struggled the most in the game against a planes plop. This shifts us into the much harder question of the necessity of air plops in small teams. I don't think we've reached a consensus on which situations require an air plop, but I would not be too put out if it turns out that most 4v4s on most maps want one. If it turns out that planes and gunships can take the air plop role about equally well, then that would be fine.