Implicit conversion from float -11.669449081803009 to int loses precision