Those formulas are all the same in mathematical terms. But we are talking floating point arithmatic here. Computing the factorials and then dividing is going to introduce errors into the computation for fairly small values of n here.