Implicit conversion from float -10.162866449511398 to int loses precision