Implicit conversion from float -13.140495867768593 to int loses precision