Implicit conversion from float -110.49504950495049 to int loses precision