BTW has anyone solved Dominoes?

On first look, it seems to be hamiltonian-cycle detection, which is NP-complete.

I'm baffled why it's medium and not hard.

Assuming it is what I think it is, Diamond and Dominoes do not belong in the same difficulty class.

NP complete means it will take more than polynomial time to run. Maybe the algorithm is still simple.

By the way #mathematics exists now.

Agreed. naive backtracking is definitely a "medium". But an efficient solution is beyond "hard" 🙂 I was thinking of somehow mapping it to exact-cover and use Knuth's DLX. Is that a decent approach?

DLX = dancing links?