Implicit conversion from float -11.596638655462186 to int loses precision