Implicit conversion from float -80.75471698113208 to int loses precision