Implicit conversion from float -113.0864197530864 to int loses precision