Implicit conversion from float -80.09925558312656 to int loses precision