Or just ........... linearly extrapolate based off their networked position and time for shots to reach them or use most enemies really simple movement patterns inside the autoaim?
Using deep learning as seen in this article:
My idea is training an AI model by recording gameplay and labeling datasets with data recorded from the game's under-the-hood API. Logging everything from player/enemy position to environment, walls, pets, etc would provide data that could be paired with the recorded gameplay to train a ML model.
Using it would look like:
The program takes a video stream of the Realm window, OBS does this. It reads the stream, calculates autoaim based off object classification and the algorithm used in flash, then sends the mouse position tracked with the nearest enemy.
I could see a downside being not able to use the mouse while playing. Perhaps a toggle, perhaps the AI could be trained to automate everything like autoloot and portal enter. Or maybe the mouse position could be sent directly to the server?
This was just an idea I had when I came back here from the break I've taken since exalt release. I hope to see some inspiration in these forums soon
Or just ........... linearly extrapolate based off their networked position and time for shots to reach them or use most enemies really simple movement patterns inside the autoaim?
Rotmg ai exists, it has read one article about ai on its journey to self awareness, and it started this thread