Implicit conversion from float -9.098360655737707 to int loses precision