Implicit conversion from float -110.75757575757575 to int loses precision