Implicit conversion from float -174.37177280550776 to int loses precision