by eleven » Mon Mar 18, 2019 8:37 pm
Agree.
Concerning the old post:
Using graph theory for sudoku problems was made popular by Gordon Royle, who has found most of the first 30000 known 17-clue puzzles with the help of sophisticated graph theory programs.
But forget it for manual solving.
And a warning to use the fever site: Because of the poor other solving methods some "diabolic" are simple, and some have multiple solutions, which are "solved" with the implemented graph algorithm.
E.g. i got this one: ..1...9...9.3.7.6.6...8...1....3.......6..............9...6...5.2.....3...8...7..