Reciprocal Products That Telescope

Why reciprocal products telescope

A product in a denominator can hide a difference that is much easier to sum. For example, the terms \(1/[n(n+1)]\) do not look like a familiar sequence of cancellations. But each term can be rewritten using two reciprocals. When those rewritten terms are added, most of them cancel, leaving only a few terms at the beginning and end.

This matters for both finite sums and infinite series. A finite sum can often be evaluated exactly by tracking the terms that remain. An infinite series can then be evaluated by asking what happens to its partial sums as the number of terms grows.

The basic decomposition

To find the decomposition, compare the two reciprocals \(1/n\) and \(1/(n+k)\). Their difference has denominator \(n(n+k)\):

Dividing this identity by \(k\) gives the form that matches the original term.

The factor \(1/k\) is essential: without it, the difference of reciprocals is \(k\) times the desired term. The simplest case is \(k=1\), which gives the especially direct identity \(1/[n(n+1)]=1/n-1/(n+1)\).

Finite partial sums and cancellation

Once the terms have been decomposed, add them in order. A reciprocal with a minus sign in one term appears with a plus sign in the next term when \(k=1\). For larger \(k\), the same cancellation occurs after a gap: the negative term \(-1/(n+k)\) cancels a positive reciprocal from the term whose index is \(n+k\), provided that index is included in the sum.

It is useful to define the \(m\)th harmonic number by \(H_m=\sum_{j=1}^{m}1/j\) for each positive integer \(m\). This notation lets us record the finite sum compactly. It does not assume that an infinite harmonic series converges; here it is only a name for a finite sum.

To see the terms behind the formula, the positive reciprocals contribute \(1+1/2+\cdots+1/N=H_N\). The negative reciprocals contribute \(1/(k+1)+1/(k+2)+\cdots+1/(N+k)\), which equals \(H_{N+k}-H_k\). Subtracting the latter from the former gives \(H_N-(H_{N+k}-H_k)\), and the decomposition supplies the overall factor \(1/k\).

Why the finite-sum formula works

The harmonic-number expression describes the same cancellation without requiring the terms to be written out individually. The following derivation also shows precisely which terms remain.

In equations, the complete argument is

This derivation is valid for every positive \(N\) and \(k\). If \(N\) is at least \(k\), the cancellation can also be seen directly: the terms with indices \(k+1\) through \(N\) cancel, leaving the first \(k\) positive reciprocals and the final \(k\) negative reciprocals. The harmonic-number formula handles the index ranges correctly even when \(N<k\).

From partial sums to the infinite series

An infinite series is defined through the limit of its partial sums. In this case, the finite-sum formula makes that limit accessible. For fixed positive integer \(k\), rewrite the formula as

There are exactly \(k\) terms in the remaining finite sum. Each is positive and at most \(1/(N+1)\), so

Since \(k\) is fixed, \(k/(N+1)\) tends to zero as \(N\) tends to infinity. Therefore the remaining sum tends to zero, and the partial sums tend to \(H_k/k\). By the definition of an infinite series as the limit of its partial sums, the series converges to that value.

A common coefficient error

A frequent mistake is to omit the factor \(1/k\). For \(k=2\), writing \(1/[n(n+2)]=1/n-1/(n+2)\) is incorrect. At \(n=1\), the left-hand side is \(1/(1\cdot3)=1/3\), while the proposed right-hand side is \(1-1/3=2/3\). They are not equal.

The correction follows directly from the common-denominator calculation: \(1/n-1/(n+2)=2/[n(n+2)]\), so the difference must be divided by \(2\). More generally, the numerator of \(1/n-1/(n+k)\) is \(k\), which is why the correct coefficient is \(1/k\). A second useful check is to test any proposed decomposition at a small value such as \(n=1\) before summing it.

When this method is useful

Look for this method when a term has two factors in its denominator, one of which is the other plus a fixed positive integer. The decomposition replaces the product with a difference whose terms can cancel across successive indices. After that, keep the finite partial sum visible: it shows both what cancels and what remains at the endpoints.

For an infinite series, do not infer the sum just from seeing cancellation in a few terms. Use the exact finite partial-sum formula and take its limit. Here the uncancelled part consists of a fixed number \(k\) of terms, each tending to zero; the bound \(k/(N+1)\) verifies that their total tends to zero. That final limit step is what turns a pattern of finite cancellation into a value for the infinite series.