Implicit conversion from float -13.231939163498097 to int loses precision