Implicit conversion from float -178.76712328767127 to int loses precision