Implicit conversion from float -14.380165289256198 to int loses precision