Implicit conversion from float -9.167927382753419 to int loses precision