Implicit conversion from float -11.091405184174619 to int loses precision