Implicit conversion from float -11.464435146443506 to int loses precision