Implicit conversion from float -11.559633027522949 to int loses precision