Implicit conversion from float -22.877697841726615 to int loses precision