Implicit conversion from float -52.42636746143057 to int loses precision