Implicit conversion from float -10.602409638554207 to int loses precision