As much as we’d like to think otherwise, Google Deepmind isn’t spending millions of dollars on R&D just for the good of the world E-Sports community. Yes, it does kind of suck for Starcraft 2 if an AI is essentially able to find and then execute unbeatable, game-breaking strategies. It would force AlphaStar to depend more on long-term strategy and high-level thinking and potentially produceīut I think a certain amount of this is grounded in ego, a desire to say “this isn’t fair! We’d beat it if it were subject to the same limitations as us meatbags!” Which is kind of missing the point. On comment boards and reddit, I saw a lot of “it won because it can click faster!” “We should limit it to human APM !” “Make it use a keyboard and mouse!” “Nerf Protoss!” (Ok fine nobody said that last one Zerg is still OP). And that’s what got MaNa (the human pro) in the end, even though he was often in a better strategic position - he kept bringing units into fights that would be winnable against human opponents and then getting ripped to shreds by the AI. You can make up for plenty of dumb at the strategic level if you’re running around with a group of essentially un-killable units. It’s ability to exert flawless command over many units simultaneously meant that, time and again, AlphaStar would get into fights that a human would have lost, and then pull out a victory. If you watch all of the matches, incredible micro-management (or micro, as we call it in Starcraft circles) was clearly AlphaStar’s core advantage. No human could micro-manage so many units so flawlessly and so quickly. This Stalker rotation is something that really high-level human players do, but only with a few Stalkers at a time.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |