Implicit conversion from float -234.51923076923077 to int loses precision