Implicit conversion from float -132.69938650306747 to int loses precision