Implicit conversion from float -13.909465020576121 to int loses precision