Implicit conversion from float -16.03053435114505 to int loses precision