Implicit conversion from float -10.594795539033441 to int loses precision