Blog of coding, TCG's and whatever else I'm doing.
Hit detection is a key element of video games and needs to be optimised to ensure the efficient calculation of data in a low latency environment. This is a short investigation into how these work and various optimisation that are used. All code used here has not been tested and is mainly research or concept material.
Valorant hitboxes are represented as pills, a line with a radius spanning around it. For this one we will only look at bullets (represented as lines) and how they are calculated for a hit.
Based on this video there are about 17 hitboxes per agent.
Firstly, we need to establish how camera/bullet direction is calculated. Typically this is stored as yaw (horizontal) and pitch (vertical) for the camera angle. This helps establish where a person is looking. One example is CS2/GO.
Spatial Partitions are used to break the map into smaller more manageable areas.
The following code gets a line and calculates the next box a line will enter. To ensure you do not indefinitely calculate bounding boxes there is a world box limit. (This is an assumption based on the fact it would be impossible to stop a line if there was not which would be computationally expensive)
Example code
3D partition example
Bounding boxes are typically axis aligned and can help optimisations by providing simple calculations on whether a line can possibly intercept more complex hitboxes or polygons. So rather then calculate 10 hitboxes, we calculate 1 larger one first to ensure it is possible for the smaller ones to be reachable. There is a variety of ways of this being stored however a possible way is storing x,y,z coords and their lengths.
test