Implicit conversion from float -14.237288135593218 to int loses precision