Implicit conversion from float -11.00200400801603 to int loses precision