Implicit conversion from float -88.06841046277665 to int loses precision