Implicit conversion from float -220.34482758620692 to int loses precision