Telescoping Finite Sums by Cancellation

Why Consecutive Differences Simplify

A finite sum can contain many terms and still have a simple value. This happens when the terms are differences between consecutive values of another sequence: a positive occurrence of an intermediate value can cancel a negative occurrence of that same value in the next difference.

For instance, in \((a_1-a_2)+(a_2-a_3)+(a_3-a_4)\), the terms \(-a_2\) and \(a_2\) cancel, as do \(-a_3\) and \(a_3\). Only the first and last values remain. The key is to expand the indexed sum far enough to see the pattern and keep track of the endpoint indices.

The Basic Cancellation Formula

Let \(a_m,a_{m+1},\ldots,a_{n+1}\) be a finite list of numbers, where \(m\) and \(n\) are integers with \(m\le n\). The sum of consecutive differences \(a_k-a_{k+1}\), with \(k\) running from \(m\) through \(n\), is

The sum includes \(n-m+1\) differences. Notice that the final difference uses \(a_{n+1}\), not \(a_n\). The index \(n\) identifies the last value of \(k\); because that term is \(a_n-a_{n+1}\), the sequence must be defined through \(n+1\).

To see what happens, write the first and last terms and include the intermediate differences:

Every intermediate value appears twice: once with a minus sign and once with a plus sign. For each index \(j\) strictly between \(m\) and \(n+1\), the term \(-a_j\) comes from the difference with \(k=j-1\), while \(+a_j\) comes from the difference with \(k=j\). They sum to zero. The first value \(a_m\) has no preceding difference in the sum, and the last value \(a_{n+1}\) has no following difference. Those are precisely the values that remain.

The reversed formula follows because \(a_{k+1}-a_k=-(a_k-a_{k+1})\) for every \(k\). Thus the reversed sum is the negative of \(a_m-a_{n+1}\), which is \(a_{n+1}-a_m\). The subtraction order determines the sign of the answer.

A Complete Derivation by Expansion

The cancellation formula can be verified directly for any finite index range. For clarity, suppose first that \(m<n\). The first few and last few terms of the sum are

On adding the parentheses, \(a_m\) occurs once with a positive sign, and \(a_{n+1}\) occurs once with a negative sign. Each value \(a_j\) for \(m+1\le j\le n\) occurs in the two adjacent differences as \(+a_j-a_j=0\). Therefore the whole sum is \(a_m-a_{n+1}\).

The same formula also holds when \(m=n\): there is one difference, and the two sides are both \(a_m-a_{m+1}\). This checks the shortest allowed sum as well as sums with several terms.

This argument explains why the endpoint indices are \(m\) and \(n+1\): each term in the sum compares a value at \(k\) with the next value at \(k+1\). The cancellation removes the values shared by neighboring differences, not the two values at the ends.

Worked Examples

How to Use the Formula Reliably

The cancellation is easy to see once the summand is in the form of a difference between consecutive sequence values. A useful procedure is to keep the index range visible while expanding, rather than trying to guess the answer from a few terms.

This procedure also helps distinguish the number of terms from the final sequence index. For example, when \(k=2,3,4,5\), the sum has four differences, but those differences involve sequence values through \(a_6\).

A Common Indexing Mistake

A frequent error is to use \(a_n\) as the final endpoint merely because \(n\) is the upper limit on \(k\). But the last summand is \(a_n-a_{n+1}\), so the final endpoint is \(a_{n+1}\). Consider again the given values \(a_2=7\), \(a_3=-2\), \(a_4=5\), \(a_5=1\), and \(a_6=9\):

This calculation has stopped at \(a_5\), leaving out the last difference \(a_5-a_6\). Expanding all four terms shows the correct endpoint:

The correction is not to change the number of differences; it is to use the sequence value one index beyond the upper limit of summation. A separate sign check is also useful: reversing every difference changes the endpoint order and therefore changes the sign.

Why Finite Telescoping Is Useful

Finite telescoping turns a sum with many terms into a calculation involving only two sequence values. It is especially useful when the summand is already written as a difference of consecutive quantities. Instead of adding every term separately, one can identify what cancels and evaluate the two remaining endpoints.

The method does not say that every finite sum simplifies this way. The cancellation requires neighboring differences to contain the same intermediate values with opposite signs. If the terms do not have that structure, this endpoint formula does not apply as written. For sums that do telescope, expanding a few terms is a reliable way to confirm both the cancellation pattern and the sign.