Implicit conversion from float -162.38095238095238 to int loses precision