Implicit conversion from float -109.63031423290204 to int loses precision