Implicit conversion from float -190.37383177570092 to int loses precision