Implicit conversion from float -16.402877697841717 to int loses precision