Implicit conversion from float -10.98591549295773 to int loses precision