Implicit conversion from float -10.06711409395973 to int loses precision