NotesOnScourge.txt (4,118 bytes)   

Notes on scourge

AI 7 - Scourge 5 - Honeycomb

Notes:
-It takes 34 min for scourge to reach Mark II on first spawner
-Mark II happens when a builder has 1200 metal
-Builders aren't consumed when building a scourge spawner or armory
-The spawner and armory have CRAZY amount of health and shield. I mean CRAZY. The Mark II are 200k shield short of being as tough as Dire Core Guardposts, which are mark VII!!! They even have less weakness than Core Guard Posts!
-Mark III Spawner have 1m health and 3.5m shield. Almost nothing has bonus against them and they even have an anti-frigate/Lone weapon. They are insane.
-Once bottled up or limited to create more buildings, builders keep accumulating metal, which is why scourge can replace losses very fast since a high mark builder will have a huge income combined with the ability to upgrade to its mark +1.
-Crippling the Scourge economy requires a combination of killing high mark builders (to make them lose on metal income, stored metal and ability to upgrade and thus raise their income) and a combination of destroying their building (to cost them metal to rebuild)
-Scourge have a "science unit" of some kind that has no clear source. It'd be important to figure out, because more science units allow them to build more builders (they have a cap of how many they can have) and they can replace killed builders extremely fast (if with Mark 1 ones)
-The builders have 1000 speed, which allow them to flee many ships of the player. Builders flee if player has local superioty like AI forces. This is why subterfuge/generalist worked surprisingly well over classic (Ambush/concussion/Splash) loadout which is the default in the scenario. V-wing, spider and Warbird can immobile the builder. Though only spider has the range to hit them before they flee, and only Warbird can catch them before they run.
-At roughly 1h15 Scourge reach Mark III.
-Linear cost increase for scourge upgrades means higher marks are comparatively cheaper for them. Much like the player, as their income and source of income increase, they can rank up faster and faster. This isn't entirely unintended, but might not be adjusted for the new balance. To see.

*Scourge Economy:
-The spawner creates up to 3 builders and a stream of neophytes, all Mark II
-The builder accumulates xp and metal every seconds, more with every mark. It spend its metal for building a Mark I spawner or armory, and for upgrading building with a linear metal cost increase for upgrading (2500 metal to build, 1200perMark to upgrade - Base builder income is 2.5 metal/second)
-Higher marks creates get more metal, thus spawner creates neoyphytes faster that morph into warriors. Armories raise the Mark of warriors, and from Mark III onward can transform warriors into exotic forms. Builders get upgraded by armories and can upgrade building up their mark +1 (Mark 1 builder upgrade Mark 1 spawner to 2).
-As costs are linear with marks and buildings/builders get more income per marks, this causes the growth of scourge to be exponential even if they didn't have more buildings, which they do.
-How do Scourge gain "science unit" is unknow. However when they do, they can create Mark II then Mark III neophytes and builders right off the spawners. "Science unit" also allows Scourge to have more builders.

*Suggestion:
-Heavily reduce the Health and shield, possibly also modify some stats, so that Spawner and Armories can be killed much more easily. This is the most important balancing factor right now, because those buildings have way way way too much health and shield.
-Reduce builder speed from 1000 to a slower 600 or 400. It'd be much easier to catch them after you find them, without constant transport-micro or turbo-raiders, and it'd allow more techs to be viable. (Currently you want light hull and fast units wayyy too much here, or/and subterfuge).
-Potentially raise the metal cost per mark of upgrading buildings for builders, which would let the scourge spread fast, but slow their mark escalation to account for the overall longer playtime needed to win in the beta.