Implicit conversion from float -16.41748942172073 to int loses precision