Implicit conversion from float -110.26075619295958 to int loses precision