Implicit conversion from float -113.14685314685315 to int loses precision