Implicit conversion from float -22.710280373831765 to int loses precision