Implicit conversion from float -113.92749244712991 to int loses precision