Currently the algorithm will continue until it runs out of moves entirely, then backup to the last critical point, however major time can be saved it there is a periodic check for unreachable moves that kicks the backup sequence off earlier. Alternatively, if there is a way to detect unreachable moves that doesn't significantly impact per-move performance, it would be most efficient to mark the move that creates an unreachable area as failed, then begin backup.

Being able to see the board state itself proved very handy, but it does impact performance astronomically. Being able to toggle these debug states would be very handy.

Ideas for possible flags:

• Maximum performance (disables all progress reporting)
• Refresh interval (sets the time between progress updates)
• Board debug (prints current board state, perhaps scaled to fit console size)
• Pass starting position and board size as flags instead of hard coded params (default to 1,1 8x8)