Skip to main content
Game Development

Return to Question

fixed typo, moved linked image into question to prevent linkrot
Source Link
Pikalek
Pikalek
  • 12.5k
  • 5
  • 45
  • 52

When trying to prototype a (top-down, so far) boxing simulation I've run into a huge roadblock with trying to set the space in which the boxers will fight, because it's important for the fighters' reach to be a factor in setting the distance of the fight, which is easier in a continuous space. On the other hand, they should also be able to cut angles where the opponent is out of position when compared to themselves (like Tyson didTyson did below), which would probably be easier to represent in a discrete space, like a hex, but the rotation would be very rigid. The difference between stances would also be more difficult to represent in a discrete gamespace, since foot position is entirely different, as well as the reach.

Mike_Tyson_vs_Clifford_Etienne

I've been coding this in Godot, so it would not be that much more difficult to do a continuous space and use their signals and collision shapes to do it, my only concern being the computational cost of actions would be raised with the amount of positions available, so I'd rather have an abstraction.

Is there any obvious solution I'm missing? Some way both things can be represented (in a way an AI can easily interact with) without unnecesaryunnecessary granularity? Or a way to do it in a continuous space that wouldn't greatly increase computational overhead with the massive array of new positional possibilities?

When trying to prototype a (top-down, so far) boxing simulation I've run into a huge roadblock with trying to set the space in which the boxers will fight, because it's important for the fighters' reach to be a factor in setting the distance of the fight, which is easier in a continuous space. On the other hand, they should also be able to cut angles where the opponent is out of position when compared to themselves (like Tyson did), which would probably be easier to represent in a discrete space, like a hex, but the rotation would be very rigid. The difference between stances would also be more difficult to represent in a discrete gamespace, since foot position is entirely different, as well as the reach.

I've been coding this in Godot, so it would not be that much more difficult to do a continuous space and use their signals and collision shapes to do it, my only concern being the computational cost of actions would be raised with the amount of positions available, so I'd rather have an abstraction.

Is there any obvious solution I'm missing? Some way both things can be represented (in a way an AI can easily interact with) without unnecesary granularity? Or a way to do it in a continuous space that wouldn't greatly increase computational overhead with the massive array of new positional possibilities?

When trying to prototype a (top-down, so far) boxing simulation I've run into a huge roadblock with trying to set the space in which the boxers will fight, because it's important for the fighters' reach to be a factor in setting the distance of the fight, which is easier in a continuous space. On the other hand, they should also be able to cut angles where the opponent is out of position when compared to themselves (like Tyson did below), which would probably be easier to represent in a discrete space, like a hex, but the rotation would be very rigid. The difference between stances would also be more difficult to represent in a discrete gamespace, since foot position is entirely different, as well as the reach.

Mike_Tyson_vs_Clifford_Etienne

I've been coding this in Godot, so it would not be that much more difficult to do a continuous space and use their signals and collision shapes to do it, my only concern being the computational cost of actions would be raised with the amount of positions available, so I'd rather have an abstraction.

Is there any obvious solution I'm missing? Some way both things can be represented (in a way an AI can easily interact with) without unnecessary granularity? Or a way to do it in a continuous space that wouldn't greatly increase computational overhead with the massive array of new positional possibilities?

Source Link
Victor
Victor
  • 13
  • 2

How to best simulate the space in a boxing game to account for footwork and angles?

When trying to prototype a (top-down, so far) boxing simulation I've run into a huge roadblock with trying to set the space in which the boxers will fight, because it's important for the fighters' reach to be a factor in setting the distance of the fight, which is easier in a continuous space. On the other hand, they should also be able to cut angles where the opponent is out of position when compared to themselves (like Tyson did), which would probably be easier to represent in a discrete space, like a hex, but the rotation would be very rigid. The difference between stances would also be more difficult to represent in a discrete gamespace, since foot position is entirely different, as well as the reach.

I've been coding this in Godot, so it would not be that much more difficult to do a continuous space and use their signals and collision shapes to do it, my only concern being the computational cost of actions would be raised with the amount of positions available, so I'd rather have an abstraction.

Is there any obvious solution I'm missing? Some way both things can be represented (in a way an AI can easily interact with) without unnecesary granularity? Or a way to do it in a continuous space that wouldn't greatly increase computational overhead with the massive array of new positional possibilities?