It is, but for 18 move solutions and fewer, it is actually mathematically most likely that the solution is optimal, due to the rarity of the scrambles of length (solution length -1). Which explains why my 18 and Harry's 17 were both optimal. I guess Sebastiano's 18 was not, but we now have 3...
What is your current method for recognition?
None of the above ;) You have a physically twisted corner that cannot be solved with algorithms. Solve first two blocks, phyiscally twist all of the corners upwards so that their yellow stickers are on top, and then lookup the CMLL. This should...
ah, it's a 4x4. that would be irresolvable on a 3x3, but this is a special case on a 4x4 where you can run into that called PLL parity. just do:
2R2 U2 2R2 u2 2R2 u2 U2
Then, resume your method from how the cube looks afterwards, and you should be able to solve it
Like, solves CP and CO? You could put the F2L corner at UBL and do the non-conjugated Y perm that is used for Old Pochmann in 3BLD
R U' R' U' R U R' F' R U R' U' R' F R
which, again, is just a Y perm without the first and last F/F' moves