Implicit conversion from float -80.63492063492063 to int loses precision