Implicit conversion from float -14.86555697823303 to int loses precision