Implicit conversion from float -167.75656324582337 to int loses precision