Implicit conversion from float -13.556485355648533 to int loses precision