Implicit conversion from float -21.518324607329845 to int loses precision