Implicit conversion from float -11.425661914460278 to int loses precision