View Issue Details
ID | Project | Category | Date Submitted | Last Update | |
---|---|---|---|---|---|
0008969 | AI War 1 / Classic | Suggestion - AI Type Ideas | Jul 18, 2012 12:14 am | Aug 19, 2012 12:37 am | |
Reporter | Hearteater | Assigned To | |||
Status | considering | Resolution | open | ||
Summary | 0008969: Emulator AI | ||||
Description | This AI type selects a random other AI type to emulate. Each time the player destroys one of its Command Stations, it has a chance to change which AI type it is emulating. While emulating a target AI it: 1) Gets the Wave and Reinforcement multiplier of the target 2) Slowly builds defining structures (Fortresses, Force Shields, AI Eyes, BHG) of the target AI, starting at alerted planets and expand outward from there. 3) Temporarily has all special ship unlocks of the target AI 4) Has any other special abilities (such as Stealth, Camouflage, Backdoor Gate) The Technologist, Everything and Core AIs are never emulated. | ||||
Tags | No tags attached. | ||||
Internal Weight | Feature Suggestion | ||||
|
How about a simpler implementation with three kinds of emulator -- offensive, defensive, and both? The offensive emulator would send each wave as though it were a different AI type, the defensive one would just start out with a different AI-type-specific structure on each planet, and the both would do both at once. |
|
I'd be up for this if it were just a bit more interesting somehow, if this AI had some personality. Maybe instead of "emulator," it's got multiple personality disorder? Or maybe it's "fickle," in that it eagerly pursues lines of attack that it quickly bores of and leaves its half-constructed toys strewn across the cosmos. - Quickly shifts technologies, approaches, etc., but has weaker ships overall, and many of its stationary defenses have less than complete health. - But, when it attacks, it does so often with mixed waves, half-interestedly flinging its aborted attempts at innovation at the humans, as though distracted by the latest new thing it's "really working on." edit: Don't mind me, I'm just hijacking your entire idea, is all. |
|
@Doctorfrog, that sounds really fun, actually. I would love to play against that. |
|
I like both the ideas here, although I worry that if the AI is in "defensive" mode too much/too long it'll have all the benefits of being a defensive AI for when it switches back to offensive, and it wouldn't really "lose" anything. I did, however, just have another idea... 0009301 |
|
True, it does get benefits for being in a defensive-style AI, but generally it isn't getting strong waves then. And you can always try and force a personality change by claiming its systems. Of course, that could always make things worse. |
|
What I mean is, it gets to build up defenses, then swap AI types to an aggressive one, retaining almost all of the benefits of the defensive type. Aggressive AIs shouldn't be benefiting from Hard Targets, such as force fields, as often, yet the Emulator AI (either one) would be allowed to keep them. If you look at my Progressive AI, you'll note that each transition is clear, and there's no back-tracking (or shouldn't be, though I did dedicate a paragraph to the possibility). It's total number of structures may end up being higher, but it doesn't get them all at once. Pretty much once it transitions to defensive mode, it's lost most, if not all, of it's offensive structures, but it's also lost the ability to retaliate. Unlike an AI that can transition from defense to offense, retaining hard-to-hurt targets, while also bringing down the giant hammer: getting the best of both worlds. |
|
Yes, but during the time it is making defensive structures, it isn't attacking you with hard waves. That's the trade-off. And since the player has control of the circumstances of it swapping, they can somewhat decide whether they want to leave it with a permanent benefit or make it switch quickly. If that's too much of an issue, permanent structures can be limited to alerted worlds, so you are never in danger of excessive build-up across the entire map. |
|
[i]> permanent structures can be limited to alerted worlds[/i] That works, although I will maintain that taking a command station is risk/risk with no reward appearing in the equation. That is, it's only a bad idea to take another command station: either the AI swaps to offense and hits you hard, or it stays defense and builds up more. You've got no way to strategically balance the choice of destroying the command station, unlike say a co-processsor or a superterminal. |
Date Modified | Username | Field | Change |
---|---|---|---|
Jul 18, 2012 12:14 am | Hearteater | New Issue | |
Jul 18, 2012 11:21 am | tigersfan | Internal Weight | => Feature Suggestion |
Jul 18, 2012 11:21 am | tigersfan | Status | new => considering |
Jul 27, 2012 4:25 pm | martyn_van_buren | Note Added: 0027165 | |
Jul 27, 2012 10:45 pm | doctorfrog | Note Added: 0027167 | |
Jul 27, 2012 10:46 pm | doctorfrog | Note Edited: 0027167 | |
Jul 28, 2012 2:56 pm | martyn_van_buren | Note Added: 0027170 | |
Aug 18, 2012 10:56 pm | Draco18s | Note Added: 0027798 | |
Aug 18, 2012 10:56 pm | Draco18s | Note Edited: 0027798 | |
Aug 18, 2012 11:38 pm | Draco18s | Note Edited: 0027798 | |
Aug 19, 2012 12:13 am | Hearteater | Note Added: 0027808 | |
Aug 19, 2012 12:23 am | Draco18s | Note Added: 0027811 | |
Aug 19, 2012 12:29 am | Hearteater | Note Added: 0027813 | |
Aug 19, 2012 12:37 am | Draco18s | Note Added: 0027814 |