Implicit conversion from float -10.441696113074201 to int loses precision