Cornflakes_91 wrote: ↑
Fri Nov 03, 2017 4:54 pm
i definitely agree with your basic sentiment, and the tech stuff to mitigate or even improve with various design decisions sounds cool
(how to implement and make the AI and pcg aware of it is a completely different can of worms, tho..)
but my main problem with cutting stuff apart in a game thats not a voxel building game is that it can be a very large random "you are fucked now" factor.
from either being unlucky with the enemy's hits or maybe even with the PCG for the ship you are sitting in (for when the player doesnt build eir own industry but just buys ready-made ships)
then its just a game over from something the player has no real control over, which is frustrating and no fun in any capacity.
I think mitigating the "random death roll" factor shouldn't be too bad though. The way I see it, having your ship get cut apart is more of a prolonged death animation than anything else. You *might* be able to salvage the situation or escape, even with the loss of function, but it isn't guaranteed. Instead of reaching zero and suffering a critical existence failure, at some point you slowly lose functionality until completely disabled. Getting hit at the very beginning, or even a bit after the midpoint of battle, shouldn't be taking chunks out of your ship unless you are getting completely destroyed by an alpha strike, which would kill you normally anyway. This allows for cool comeback moments, where you get hit really hard and lose some engines, but manage to hit back and win. And if you don't manage to make a comeback, then at least you still get to watch a nifty destruction animation as your ship is cut to pieces. Having the tactical capabilities of combatants change as they are damaged is something that could add a lot to combat, as long as it wasn't overdone, as you point out. We aren't playing Limit Theory: Dead Space Edition, after all- it doesn't have to be excessive or happen all that often to add some interesting wrinkles to "shoot until it dies" combat.
As for your point about teaching the AI and PCG algorithms to appropriately use those "design altering technologies", it is a pretty tough problem and is probably a pipe dream. On the other hand, it isn't too farfetched to propose a machine learning system to figure it out, similar to Flat's proposal for self-learning pilot AI. I think the main issue would actually be generating the technologies themselves. Without a human to be like "oh, it has this novel property", I'm not sure how you would go about giving things new, unique rules. That might not be necessary though. Just having a few would give enough variety were the combinations would remain suitably unique over the course of play, especially if you don't run into them all that often.