\nIf you plot the z-velocity for the ball, you would get something like this:
\n
![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4206802%2F64c915a32f9a9f301d32881af94d3243%2FScreenshot%20from%202022-10-04%2012-14-05.png?generation=1664865926888950&alt=media\")
Does Rocket League have gravity?
\nIf so, is it equal to earth gravity ~= 9.8 m/s2?
\nIn case there is gravity in the z-axis (vertical) it can be used to create new features based on projectile motion equations.
\nThanks in advance.
\n", "messages": [ { "id": "1970212", "postDate": "10/03/2022 23:38:37", "content": "Does Rocket League have gravity?
\nIf so, is it equal to earth gravity ~= 9.8 m/s2?
\nIn case there is gravity in the z-axis (vertical) it can be used to create new features based on projectile motion equations.
\nThanks in advance.
\n", "rawMarkdown": "Does Rocket League have gravity?\n\nIf so, is it equal to earth gravity ~= 9.8 m/s2? \n\nIn case there is gravity in the z-axis (vertical) it can be used to create new features based on projectile motion equations.\n\nThanks in advance. \n\n@dster @ashleychow", "votes": null }, { "id": "1970215", "postDate": "10/03/2022 23:57:11", "content": "Hey moth, good question. Yes there is gravity, but unfortunately I don't know the gravitational constant in this dataset -- I imagine one could infer it from the data. I agree it could be useful to estimate the trajectory forward in time, keeping in mind that there are walls and a ceiling, not to mention the players themselves, which could disturb that trajectory.
", "rawMarkdown": "Hey moth, good question. Yes there is gravity, but unfortunately I don't know the gravitational constant in this dataset -- I imagine one could infer it from the data. I agree it could be useful to estimate the trajectory forward in time, keeping in mind that there are walls and a ceiling, not to mention the players themselves, which could disturb that trajectory.", "votes": null }, { "id": "1970255", "postDate": "10/04/2022 01:12:44", "content": "Thanks for your answer @dster I guess I can use motion equations with Earth's g and compare them to actual positions on small time lapses.
", "rawMarkdown": "Thanks for your answer @dster I guess I can use motion equations with Earth's *g* and compare them to actual positions on small time lapses.", "votes": null }, { "id": "1970581", "postDate": "10/04/2022 06:39:34", "content": "Pretty good rundown on RL physics
\nhttps://www.youtube.com/watch?v=9uh8-nBlufM
\n650 uu/s^2 (2/3 earth gravity)
", "rawMarkdown": "Pretty good rundown on RL physics\n\nhttps://www.youtube.com/watch?v=9uh8-nBlufM\n\n650 uu/s^2 (2/3 earth gravity)", "votes": null }, { "id": "1970632", "postDate": "10/04/2022 07:16:50", "content": "Yes, there is.
\nIf you plot the z-velocity for the ball, you would get something like this:
\n
This is just like a ball bouncing under gravity.
\nYou could better approximate it if you factored in event_time, but for simplicity and since there's no event_time in test data, this could be a simple approximation, leaving out any collisions:
\n$$ v_{z}^{t+1} = v_{z}^{t} - 1.3$$
\nAnother interesting thing is, there's also friction.
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4206802%2F811512f47eb945151b8ab6219361fcfb%2FScreenshot%20from%202022-10-04%2012-39-32.png?generation=1664867391064081&alt=media\")
The velocities seem to be decreasing by 0.3% every step, except for complete change due to some collision. Something like:
\n$$ v_{x|y}^{t+1} = v_{x|y}^t \\times 0.997 $$
", "rawMarkdown": "Yes, there is. \nIf you plot the z-velocity for the ball, you would get something like this:\n\n\nThis is just like a ball bouncing under gravity.\n\nYou could better approximate it if you factored in event_time, but for simplicity and since there's no event_time in test data, this could be a simple approximation, leaving out any collisions:\n\n$$ v_{z}^{t+1} = v_{z}^{t} - 1.3$$\n\nAnother interesting thing is, there's also friction.\n\n\nThe velocities seem to be decreasing by 0.3% every step, except for complete change due to some collision. Something like:\n\n$$ v_{x|y}^{t+1} = v_{x|y}^t \\times 0.997 $$", "votes": null }, { "id": "1971134", "postDate": "10/04/2022 13:26:18", "content": "Thanks @kaggleqrdl for the reference. I hadn't realized the complexity of the game's physics. It also seems that instead of having air resistance they have a half-life speed decay. Interesting
", "rawMarkdown": "Thanks @kaggleqrdl for the reference. I hadn't realized the complexity of the game's physics. It also seems that instead of having air resistance they have a half-life speed decay. Interesting", "votes": null }, { "id": "1978526", "postDate": "10/08/2022 18:47:42", "content": "How did you guys get in-game data?
", "rawMarkdown": "How did you guys get in-game data?", "votes": null } ], "comments": [ { "id": 1970215, "author_name": "dster", "author_url": "", "post_date": "10/03/2022 23:57:11", "content": "Hey moth, good question. Yes there is gravity, but unfortunately I don't know the gravitational constant in this dataset -- I imagine one could infer it from the data. I agree it could be useful to estimate the trajectory forward in time, keeping in mind that there are walls and a ceiling, not to mention the players themselves, which could disturb that trajectory.
", "votes": null, "replies": [ { "id": 1970255, "author_name": "alejopaullier", "author_url": "", "post_date": "10/04/2022 01:12:44", "content": "Thanks for your answer @dster I guess I can use motion equations with Earth's g and compare them to actual positions on small time lapses.
", "votes": null, "replies": [] } ] }, { "id": 1970581, "author_name": "kaggleqrdl", "author_url": "", "post_date": "10/04/2022 06:39:34", "content": "Pretty good rundown on RL physics
\nhttps://www.youtube.com/watch?v=9uh8-nBlufM
\n650 uu/s^2 (2/3 earth gravity)
", "votes": null, "replies": [ { "id": 1971134, "author_name": "alejopaullier", "author_url": "", "post_date": "10/04/2022 13:26:18", "content": "Thanks @kaggleqrdl for the reference. I hadn't realized the complexity of the game's physics. It also seems that instead of having air resistance they have a half-life speed decay. Interesting
", "votes": null, "replies": [] } ] }, { "id": 1970632, "author_name": "aatiffraz", "author_url": "", "post_date": "10/04/2022 07:16:50", "content": "Yes, there is.
\nIf you plot the z-velocity for the ball, you would get something like this:
\n
This is just like a ball bouncing under gravity.
\nYou could better approximate it if you factored in event_time, but for simplicity and since there's no event_time in test data, this could be a simple approximation, leaving out any collisions:
\n$$ v_{z}^{t+1} = v_{z}^{t} - 1.3$$
\nAnother interesting thing is, there's also friction.
\n
The velocities seem to be decreasing by 0.3% every step, except for complete change due to some collision. Something like:
\n$$ v_{x|y}^{t+1} = v_{x|y}^t \\times 0.997 $$
", "votes": null, "replies": [] }, { "id": 1978526, "author_name": "carneirofernando", "author_url": "", "post_date": "10/08/2022 18:47:42", "content": "How did you guys get in-game data?
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1970212": "Does Rocket League have gravity?\n\nIf so, is it equal to earth gravity ~= 9.8 m/s2? \n\nIn case there is gravity in the z-axis (vertical) it can be used to create new features based on projectile motion equations.\n\nThanks in advance. \n\n@dster @ashleychow", "1970215": "Hey moth, good question. Yes there is gravity, but unfortunately I don't know the gravitational constant in this dataset -- I imagine one could infer it from the data. I agree it could be useful to estimate the trajectory forward in time, keeping in mind that there are walls and a ceiling, not to mention the players themselves, which could disturb that trajectory.", "1970255": "Thanks for your answer @dster I guess I can use motion equations with Earth's *g* and compare them to actual positions on small time lapses.", "1970581": "Pretty good rundown on RL physics\n\nhttps://www.youtube.com/watch?v=9uh8-nBlufM\n\n650 uu/s^2 (2/3 earth gravity)", "1970632": "Yes, there is. \nIf you plot the z-velocity for the ball, you would get something like this:\n\n\nThis is just like a ball bouncing under gravity.\n\nYou could better approximate it if you factored in event_time, but for simplicity and since there's no event_time in test data, this could be a simple approximation, leaving out any collisions:\n\n$$ v_{z}^{t+1} = v_{z}^{t} - 1.3$$\n\nAnother interesting thing is, there's also friction.\n\n\nThe velocities seem to be decreasing by 0.3% every step, except for complete change due to some collision. Something like:\n\n$$ v_{x|y}^{t+1} = v_{x|y}^t \\times 0.997 $$", "1971134": "Thanks @kaggleqrdl for the reference. I hadn't realized the complexity of the game's physics. It also seems that instead of having air resistance they have a half-life speed decay. Interesting", "1978526": "How did you guys get in-game data?" }, "source": "meta" }