Implicit conversion from float -69.06693711967546 to int loses precision