Implicit conversion from float -110.79365079365078 to int loses precision