Implicit conversion from float -9.061032863849775 to int loses precision