ducdontfeelsowell/24020064_24020082_24020338_CaroAI — reverse-engineered prompt
Reverse engineered prompt
Build me a simple Python console game for playing Caro against the computer.
I want the game to ask for the board size, at least 9 by 9, let the player choose X or O, then choose whether the AI uses normal Minimax or Alpha Beta pruning, and choose search depth from 1 to 5. The player should enter moves as row and column numbers starting from 1, and the game should reject bad input like wrong format, out of range coordinates, or occupied cells.
The rules are simple Caro, whoever gets 4 in a row horizontally, vertically, or diagonally wins, no blocked ends rule. If the board fills up, it should be a draw.
After the AI moves, print useful info like the score, how many board states it evaluated, and how long it took to think. Keep it running in the terminal with a clear board display. Use Python 3 and numpy if needed.
Want more depth? Deep Reverse