Implicit conversion from float -11.811023622047244 to int loses precision