Show HN: Quake 1 movement physics reinforcement learning project
1–10 of 21 posts
Re: Show HN: Quake 1 movement physics reinforcement learning project
#2This is awesome -- I grew up playing Quake 1 and really fell in love with it more as we as a community learned to exploit the movement physics. Still my favourite FPS and still arguably the one with the steepest learning curve.
I didn't have a chance to watch the video yet, but out of curiosity does this actually use QuakeWorld or base Quake 1? As I remember the physics are different between the two, right?
Re: Show HN: Quake 1 movement physics reinforcement learning project
#3This is awesome -- I grew up playing Quake 1 and really fell in love with it more as we as a community learned to exploit the movement physics. Still my favourite FPS and still arguably the one with the steepest learning curve. I didn't have a chance to watch the video yet, but out of curiosity does this actually use QuakeWorld or base Quake 1? As I remember the physics are different between the two, right?
It's difficult to tell, but my guess would be quakeworld.
QW has slightly faster strafing, so there is a significant difference.
Re: Show HN: Quake 1 movement physics reinforcement learning project
#4How difficult would it be to encode the optimal running technique as a rule instead of learning it with RL?
Re: Show HN: Quake 1 movement physics reinforcement learning project
#5I just thought that's the geekiest thing I've seen in a while... good job! :)
Re: Show HN: Quake 1 movement physics reinforcement learning project
#6How difficult would it be to encode the optimal running technique as a rule instead of learning it with RL?
It wouldn't be difficult at all if the optimal running technique was known before hand; I think the goal of many of these RL exercises is to either i) find a better solution than what we may have imagined or ii) confirm that our knowledge was indeed the best possible solution!
Re: Show HN: Quake 1 movement physics reinforcement learning project
#7This is awesome -- I grew up playing Quake 1 and really fell in love with it more as we as a community learned to exploit the movement physics. Still my favourite FPS and still arguably the one with the steepest learning curve. I didn't have a chance to watch the video yet, but out of curiosity does this actually use QuakeWorld or base Quake 1? As I remember the physics are different between the two, right?
It's the Quake 1 physics, not QW. I haven't looked at the QW physics in detail, but I think the main difference is that there's a single frame of ground friction applied between every jump, which ends up making a difference to the optimal techniques.
Re: Show HN: Quake 1 movement physics reinforcement learning project
#8How difficult would it be to encode the optimal running technique as a rule instead of learning it with RL?
It wouldn't be difficult at all if the optimal running technique was known before hand; I think the goal of many of these RL exercises is to either i) find a better solution than what we may have imagined or ii) confirm that our knowledge was indeed the best possible solution!
Correct, it wouldn't be hard to take the best known strategy and implement it perfectly. Aside from the two points you mention, it also serves as a precursor to applying RL to real levels in which we may not be so sure about the optimal strategy.
Re: Show HN: Quake 1 movement physics reinforcement learning project
#9Wild, I was just playing QW a week ago for the first time in 20 years and someone was showing me how to do this running technique after destroying me in a one on one.
Re: Show HN: Quake 1 movement physics reinforcement learning project
#10How difficult would it be to encode the optimal running technique as a rule instead of learning it with RL?
It wouldn't be difficult at all if the optimal running technique was known before hand; I think the goal of many of these RL exercises is to either i) find a better solution than what we may have imagined or ii) confirm that our knowledge was indeed the best possible solution!
> It wouldn't be difficult at all if the optimal running technique was known before hand;
You're right, but this is a simulated environment. It means the physics in it are driven by a small set of known (and most likely deterministic) rules.
So I wonder if it would it be possible to analytically determine what is the best possible running strategy.