Implicit conversion from float -11.340996168582379 to int loses precision