Implicit conversion from float -113.0035335689046 to int loses precision