The ordinary Binomial Theorem written with nCr=r!(n−r)!n! only makes sense when n is a non-negative integer, because n! is undefined otherwise. To expand something like (1+x)−4 or (1−x)1/3, the theorem must be restated in a form that never needs n! of n itself: for ∣x∣<1, (1+x)n=1+nx+2!n(n−1)x2+3!n(n−1)(n−2)x3+⋯+r!n(n−1)⋯(n−r+1)xr+⋯, where n can now be any real number — negative, or a fraction, or (as a special case) a positive integer, which is when the series happens to terminate and reduces back to the ordinary finite Binomial Theorem. Two structural differences from the positive-integer case matter a great deal in practice. First, when n is not a positive integer the series never terminates — it has infinitely many terms — and the condition ∣x∣<1 is not optional decoration but the exact requirement that makes that infinite sum converge to a finite value; an exam answer is expected to quote 'first three/four terms', not the whole series. Second, to expand a general (a+b)n (rather than (1+x)n) with ∣b∣<∣a∣, the standard move is to fa …