Implicit conversion from float -80.19230769230768 to int loses precision