I completed this kata but the randomHardTest is broken for C#. At first I was receiving an overflow during summation in the determinant calculation, so I refactored to work with 64 bit ints, which fixed this problem, but the result was failing the randomHardTest with unexpected results. On closer inspection, the expected values of the test are not the real determinant values, they are the same as the 64-bit results, truncated to 32-bit. An explicit cast of my final 64-bit result allowed me to pass all tests.