Implicit conversion from float -15.354330708661422 to int loses precision