Implicit conversion from float -30.052539404553414 to int loses precision