Implicit conversion from float -110.1980198019802 to int loses precision