Implicit conversion from float -13.09954751131221 to int loses precision