I've been pondering the "expansion" dilemma some, around which there has been so much discussion. While the energy changes have increased the Haagenti-tested-and-approved minimum territory size, it seems like the territory vs. AI progress trade-offs might still become too easy after that size has been reached. Making expansion desirable while also Bad For Your Health is quite a tough nut to crack but strikes me as one of the really unique core mechanics of AI War so I think that no time spent on that dilemma can be considered wasted. If I had to describe the challenge as it stands currently, I'd say that it is keeping the "start in and conquer a well-defended cluster with sufficient reactor-room, get a sustainable task force behind enemy lines, and make sure the AI never gets any angrier so that it will remain easy prey" strategy from being optimal in all cases.
I like the drive towards additional structures for the player to capture, and some of the ideas I've seen here are pretty nifty. I think that if they remain more "bonus" items, though, that they won't really affect the dynamic as much as could be desired. I would suggest not hesitating to make at least one or two on the order of advanced factories/research stations (i.e., "must-have" rather than "nice-to-have") to really force wider and possibly non-contiguous expansion. For example, certain "enabler" structures could be required for the production of the mkII/III units, or at least the production times on the higher-level units could be penalized unless the enabler is held. That way, the "bonus" capturables don't have to shoulder the entire burden of making expansion necessary and can instead serve to make the decisions on placing/shaping captured territorial areas more fun and interesting.
Some other ideas on motivating expansion:
- Logarithmically increase the cost of mkII units and mkIII units, and make the output of harvesters multiplicative rather than additive. This would increase the importance of defending resourcing operations and the stragetic attractiveness of resource-rich systems, and perhaps make for interesting tradeoffs in terms of AI progress from conquering systems vs. the AI progress increase over the extra time required to afford advanced ships with lower incomes. However, it would make it a lot harder to recover from adverse events, and to be honest I'm really not fond of the timed progress increases, so I'm not so sure what I think of this one. Food for thought, anyway.
- Keep the ratios the same between the caps for different types of ships, but scale the base constant by which they are multiplied with the size of a player's territory: more worlds means bigger fleets, and low-resource systems would have some value still. Improved command stations could increase the contribution of a system to ship caps, and there could be capturables that would do likewise. This change could also lead to vicious cycles that would make recovering from a strong AI assault impossible, so it would probably make sense not to reduce caps from the maximum achieved by the player (except for lost capturable structures) as the AI progress cost will have been paid. Multiplayer values should probably be based on total human territory rather than being done per-player, as well.
It seems to me that currently planets are significant primarily in economic terms; they are less so tactically, and far less so logistically. Making the location of conquests in relationship to objectives (and to each other) more relevant could make for more meaningful decisions and motivate players to step out of their choke-point comfort zones. Some possibilities:
- Introduce a "coverage area" for reactors. If available energy were computed per star system, with a reactor's effects diminishing with distance, it would provide a much greater motivation for players to capture and secure forward bases. Alternately, command stations or specially-purposed structures could act as power relays, making all available reactor power applicable within a certain area (perhaps with special relays which extend power to a larger area as bonus capturable structures.) Mobile reactors or relays could keep the deep-strike task force viable but make it more risky/challenging. This would definitely be a major change and could be too computation-intensive, annoying, or complex (are ships unable to enter areas with insufficient power, or are they just forced into low-power mode on arrival, what happens when an area loses energy coverage, etc) but might be interesting, and add to the significance of galactic "geography."
- Supply lines: degrade the effectiveness of ships as the distance to the nearest friendly command station increases - perhaps a reload time penalty or something similar. Supply ships (something like the mobile reactor/relay mentioned above) could allow task forces to operate, but at increased risk, and the mechanic could be an interesting basis for new ship attributes or types. For example, starships, scouts, and potential deep-strikers of some sort could be "self-supplying," certain assault ships could be very powerful but also carry the "adjacent supplier only" attribute, and others could be designed (or repurposed - vampires/cutlasses maybe?) as mobile but still defense-only via a "fixed supplier only" attribute, etc.
- Give the AI the ability to consider player fleets as weak points to be targeted (like it will target resourcing, etc) as I've seen suggested here, or otherwise make it more adept at destroying forces that have broken through its front lines. Perhaps cross-planet raids could consider large forces outside human territory as valid objectives, or something similar, so that a fleet without nearby fortifications to withdraw to would be at greater risk of being tracked down and annihilated.
- Only allow colony ships to be built in systems with command stations, or even only allow them to be built at a player's home station. Actually, I think this ought to be done no matter what, as colony ships are more or less irrelevant when they can be built by mobile constructors (and where are the colonists coming from, in that case?)
Anyway, I know that Wall of Text hits harder than a super fortress so I beg forgiveness for being so prolix, and hope that there might be something of use (or at least passing interest) buried somewhere in here.