Can a Human Solve this Extreme Sudoku?

Matthias asks us if this extreme sudoku can be solved by a human

[terryleddra]

yes I have solved it without bifurcation. It involved some logic chains to reduce potential candidates but it can be done (not quickly).

[jeffbrunton]

I was so stuck that I just ran this through a solver to see if I could follow the logical steps.


it used multiple splits, so yes, I think its beyond a human solution other than just by mutiple branches of trial and error


it may also be that it lacks uniqueness?

[rabidsamfan]

This is such a good example of the difference between force testing and actual bifurcation. By the time you'd looked at both versions of the two in row 3, you *knew* that the 9 in row 1 had to be. So instead of working the 1s down below, you could have placed the 9 in ink right away (whether or not you knew it was restricted the other way.) I think that's a function of how you mark the grid. By using a single digit on the borderline to represent two possibilities you can go fast, but it's harder to see when two choices both lead to the same end.


I've dug out a blank grid and now I'm going to see how long it takes me to solve this one. Looks like fun!

ETA: 25 minutes. Got nowhere bifurcating with the twos (even after watching you do it) so I had to start over from a different point. Nice. But yes, I definitely needed to bifurcate.

[joshuanorthey]

Hmm maybe I should try harder ones, I always considered any time you start brute forcing it (or trial and error) “cheating”, though obviously indirect proof is a form of deduction as well. Anyway, first time I have ever watched someone else solve and it was interesting. Definitely have a different method that is faster in some things and slower in others. I may watch more...