Implicit conversion from float -13.111739745403113 to int loses precision