##### ###### ##### ### # # ### # # ###### ## ## ## ## ## ## ## # # # # # ## ##### #### ##### # # # # # # # #### ## # ## ## ## ## # # # # # ## ## # ###### ## ### # ### # ######
##### ###### ##### ### # # ### # # ###### ## ## ## ## ## ## ## # # # # # ## ##### #### ##### # # # # # # # #### ## # ## ## ## ## # # # # # ## ## # ###### ## ### # ### # ######
Expected value variant
Instead of probability, find the expected number of moves before the knight leaves the board. Let = expected moves from . If on board: where are valid destinations. Off-board destinations contribute 0 (no future moves).
This creates a system of equations. Solve iteratively: initialize all , then update until convergence. Or use Gaussian elimination for exact solution.
