Implicit conversion from float -13.701380175658713 to int loses precision