What does the Hivemind thinks of X-Com Apocalypse?

[Darth Canoli]

If I were going to make an auto-resolve, what I'd do is make it so you click it, it looks over what you're up against and how you previously performed in similar situations, and tells you what your expected results are, and then you click okay or not depending on whether you think this assessment is accurate. The autoresolve will thus actually learn from you and model your results.

Funny but this exact idea comes up each time i play a tactical with heavy combat, it's so obvious and yet, nobody implemented this kind of system as far as i know.

 

[Norfleet]

It also would likely greatly improve the quality of AIs in many games, since one of the major failings of game AIs is how they'll take an action based on their built-in assessment of a situation ("Will I win this?"), and, of course, they're totally wrong about it and get their asses handed to them. But this never prompts them to reassess the situation differently, because there's no mechanism for them to learn this. So they'll do it again, and again, and again, until they've gotten all their doodz killed. If the AI learned from disconnects between what autoresolve (probably also the same mechanism used to predict relative strengths), and what actually happened, the quality of both systems would improve.