Loading...
  OR  Zero-K Name:    Password:   

The no units Co-op meta

3 posts, 350 views
Post comment
Filter:    Player:  
sort
Sometimes in team games you see newer players that build no units but tons of defense, and this is a strange thing: where could they have learned it? In other RTS games, low elos do weird stuff but zero unit pure tower is not one of them.

Well, mega porc is the meta in Co-op. This is the result of a number of factors:

1. Brutal AI is very aggressive, teams up in pushes and have few gaps of easily exploitable unit composition at lower force density: it will defeat normal player (not some top 20 micro-god) spread out to do map control while not having full econ due to teammates backline doing porc. Now a entire team pushing can do the job, but team coordination is never a given and porc always works.

2. Brutal AI generate a good density of defenses: raiding is not particularly effective when with fraction eco, especially when constrained to fairly poor micro. The raiding player runs out of units and can not mass for pushing normally.

3. Brutal AI is very stupid in dealing with defenses: it does not preserve units, avoid chokes, tech up properly with right composition, build arty/silo/shiny and instead feeds into heavy defenses (stringer/deso/cerb/luci) continuously. On a most simple level, brutal AI generate lance, Grizzly and phantom comps, but not building targeting impaler ball.

So, what does a new player to do after doing some of tutorials to learn to multiplayer? Join a co-op room! An inside he'd see co-op specialists players doing some chokepoint mega porc, building no units but constructors (or scorp, paladin to clean up after "impenetrable" porc is built). After failing other styles the new player would learn to porc hard to win at this, and may move into a normal team game and draw the ire of everyone else in the room.

Perhaps the AI should be changed to focus on anti-structure with things like sling and impaler, and actively try to counter terraform with constructor restore. This will make it weaker to players that fight it properly but will reward better habits as opposed to brittle cheese strategy.
+1 / -0
2 years ago
Erm ... I admit I am only playing on "hard" so far, but in my experience, the AI does use artillery to take out static fortifications. Having played balanced annihilation some years back, gaat guns with nanos were my bread and butter defense, and now learning Zero-K, I tried the same thing - but the AI has forced me to abandon this tactic!

A single lance one-shotting my stingers one after the other, impalers arriving faster than I can build nano turrets, plenty of light artillery for shorter ranged defenses. I even had to disable desolators to avoid their destruction by groups of impalers. It takes one second of vision, and my fortifications are targeted by more artillery than I can repair. Oh, and let's not forget the occasional firewalker setting my nanos ablaze ...

(I admit it is somewhat weak to long ranged "defenses" such as the cerberus, and does not prioritize these as much as it should, but it does use artillery quite well)

And so the AI taught me to use mobile units to threaten their artillery to keep my defenses alive (which I still build, but much lighter, only to prevent small raiding groups from overwhelming my micro)
+2 / -0

2 years ago
The only thing the AI can't handle ATM is stacked shields. It doesn't use silo, ever, and it doesn't seem to properly calculate if it has enough damage to pierce through. It also rarely, if ever, uses shields or cloaking of its own.

The only thing the AI can do vs stacked shields is escalate to super, but even if you let it take 3/4th+ of the map, the AI doesn't seem to want to build a super before the 40+ minute mark. That's enough time for just about any player to slowly creep with heavy porc until the AI has lost too much ground and eco to actually escalate to super.

Surprisingly, the AI will try to nuke at some point. It's even not half bad at picking targets, but it does it so late that you will always have an anti by then. The AI doesn't know how to disable antis the way players do.

The AI seems to have gotten better at repelling early raids, but it also seems to have gotten way worse at expanding at a reasonable pace. In 1v1, if you deflect the early raids and then expand aggressively, you will MASSIVELY outpace it in macro.


That being said, I think it's right to claim that players who coop learn to porc where they spawn. I just don't see it as having any effect on team games. The only remedy to being a lob is experience. If they didn't learn to porc at spawn in coop, they'd just fail in some other way.

The nature of RTS games is that you need to know what your opponent can have, when it can have it, what trades reasonably well with it, when it's safe to take up more ground, when a situation has gotten unsafe and retreat is favorable to losing units... There are so many skills and judgement calls to develop. New players just have so much to learn. Whatever reasonable strategy you tell them to employ, odds are the run of the mill 100+ games player already knows how to counter.
+2 / -0