Implicit conversion from float -110.25773195876292 to int loses precision