In short

Not every series is a pure AP, GP, AGS, or a clean telescope. This article collects the techniques for handling the rest: series with dependent indices (double sums where the inner limit depends on the outer index), partial fractions applied to series with rational general terms, split-and-regroup strategies that reduce an unfamiliar sum to known pieces, and a selection of JEE-level problems that combine multiple techniques. The unifying principle: every closed-form summation works by reducing the problem to a sum you already know how to evaluate.

Here is a sum that looks innocent: \displaystyle\sum_{i=1}^{n}\sum_{j=1}^{i} j. The inner sum runs from 1 to i — not to a fixed n, but to a variable upper limit that changes with each pass of the outer sum. If you try to swap the order of summation without thinking, you get a wrong answer. If you evaluate the inner sum first, you get \frac{i(i+1)}{2}, and then the outer sum becomes \sum_{i=1}^{n} \frac{i(i+1)}{2} = \frac{1}{2}\sum_{i=1}^{n} i(i+1), which the method of differences handles cleanly: \frac{n(n+1)(n+2)}{6}.

That double sum was not hard — but it required you to recognise the dependent index, evaluate from the inside out, and then apply a technique from a different chapter. Most "miscellaneous" series problems work this way: they combine two or three simple ideas in a non-obvious order. This article is about building fluency with those combinations.

Series with dependent indices

Double sums: inside out

When the inner index depends on the outer one — \sum_{i=1}^{n}\sum_{j=1}^{i} f(i,j) — you always evaluate the inner sum first, treating i as a constant. The result is a single-variable function of i, which you then sum over i.

Take \displaystyle\sum_{i=1}^{n}\sum_{j=1}^{i} (i + j).

Inner sum (fix i, sum over j):

\sum_{j=1}^{i} (i+j) = \sum_{j=1}^{i} i + \sum_{j=1}^{i} j = i \cdot i + \frac{i(i+1)}{2} = i^2 + \frac{i(i+1)}{2} = \frac{2i^2 + i^2 + i}{2} = \frac{3i^2 + i}{2}

Outer sum:

\sum_{i=1}^{n} \frac{3i^2 + i}{2} = \frac{3}{2}\sum_{i=1}^{n} i^2 + \frac{1}{2}\sum_{i=1}^{n} i = \frac{3}{2} \cdot \frac{n(n+1)(2n+1)}{6} + \frac{1}{2} \cdot \frac{n(n+1)}{2}
= \frac{n(n+1)(2n+1)}{4} + \frac{n(n+1)}{4} = \frac{n(n+1)(2n+2)}{4} = \frac{n(n+1) \cdot 2(n+1)}{4} = \frac{n(n+1)^2}{2}
Triangular grid of a double sum with dependent indicesA triangular array of cells. Row i=1 has one cell (j=1). Row i=2 has two cells (j=1,2). Row i=3 has three cells (j=1,2,3). Row i=4 has four cells. Each cell contains the value i+j. The outer sum runs over rows, the inner sum over columns up to the row number. j → 1 2 3 4 i↓ 1 2 3 4 2 3 4 4 5 6 5 6 7 8 row sum = 2 row sum = 7 row sum = 15 row sum = 26 total = 50
The double sum $\sum_{i=1}^{4}\sum_{j=1}^{i}(i+j)$ fills a triangular grid. Each row has $i$ cells, and the cell at position $(i,j)$ holds the value $i+j$. Summing row by row (inside out) gives $2, 7, 15, 26$, and the total is $50$. The formula $\frac{n(n+1)^2}{2}$ at $n=4$ gives $\frac{4 \cdot 25}{2} = 50$.

Swapping the order of summation

When the indices are dependent, swapping the order of summation requires care. In \sum_{i=1}^{n}\sum_{j=1}^{i}, the summation region is the triangle \{(i,j) : 1 \leq j \leq i \leq n\}. If you sum over j first (for fixed i), j runs from 1 to i. If you sum over i first (for fixed j), i runs from j to n. So:

\sum_{i=1}^{n}\sum_{j=1}^{i} f(i,j) = \sum_{j=1}^{n}\sum_{i=j}^{n} f(i,j)

The two forms give the same answer — they scan the same triangle, just in different directions (row-by-row vs. column-by-column). Choosing the right order can simplify the inner sum dramatically.

Special techniques

Regrouping by splitting the general term

Sometimes a general term can be rewritten as a sum of terms, each of which belongs to a known family.

Take T_k = \frac{k^2 + 1}{k!}. Split it:

\frac{k^2 + 1}{k!} = \frac{k^2}{k!} + \frac{1}{k!}

Now \frac{k^2}{k!} = \frac{k}{(k-1)!} = \frac{(k-1)+1}{(k-1)!} = \frac{1}{(k-2)!} + \frac{1}{(k-1)!} for k \geq 2.

So T_k = \frac{1}{(k-2)!} + \frac{1}{(k-1)!} + \frac{1}{k!} for k \geq 2, and each piece is a shifted version of \sum \frac{1}{m!}, which converges to e. The sum of the entire series can be expressed in terms of e.

The principle: if you cannot sum the general term directly, break it into pieces you can sum.

Splitting a complicated general term into summable piecesA box containing the general term (k squared plus 1) over k factorial, with arrows splitting into three pieces: 1 over (k-2) factorial, plus 1 over (k-1) factorial, plus 1 over k factorial. Each piece is labelled as a shifted version of the sum of 1 over m factorial. (k² + 1) / k! 1/(k−2)! 1/(k−1)! 1/k! + + Each piece is a shifted version of Σ 1/m! — all converge to e.
The term $\frac{k^2+1}{k!}$ splits into three pieces, each a shifted factorial sum. Summing each piece separately gives the total in terms of $e$.

Partial fractions in series

When the general term is a rational function of k, partial fraction decomposition is your main tool. Factor the denominator, decompose, and sum each piece — often each piece telescopes.

Take T_k = \frac{1}{k^2 - 1} for k \geq 2. Factor: k^2 - 1 = (k-1)(k+1). Decompose:

\frac{1}{(k-1)(k+1)} = \frac{1}{2}\left(\frac{1}{k-1} - \frac{1}{k+1}\right)

This is a telescoping sum, but with a "gap" of 2 — the \frac{1}{k-1} from term k cancels with the -\frac{1}{k+1} from term k+2, not term k+1. Write out the first few and last few terms:

\sum_{k=2}^{n} \frac{1}{2}\left(\frac{1}{k-1} - \frac{1}{k+1}\right) = \frac{1}{2}\left[\left(\frac{1}{1} - \frac{1}{3}\right) + \left(\frac{1}{2} - \frac{1}{4}\right) + \left(\frac{1}{3} - \frac{1}{5}\right) + \cdots\right]

The cancellations leave: \frac{1}{2}\left[\frac{1}{1} + \frac{1}{2} - \frac{1}{n} - \frac{1}{n+1}\right] = \frac{1}{2}\left[\frac{3}{2} - \frac{1}{n} - \frac{1}{n+1}\right].

The "gap-2 telescope" is a standard pattern in JEE. The surviving terms are always the first few and the last few — exactly how many depends on the gap.

Gap-2 telescoping in the sum of 1 over (k-1)(k+1)Terms written as differences with a gap of 2. The first two positive parts and the last two negative parts survive. All middle terms cancel in pairs separated by two positions. (1/1 − 1/3) + (1/2 − 1/4) + (1/3 − 1/5) + (1/4 − 1/6) + ··· + (1/(n−1) − 1/(n+1)) cancel (gap 2) cancel survivors: 1/1 + 1/2 − 1/n − 1/(n+1) A gap-2 telescope: 2 terms survive at each end.
In a gap-2 telescope, each positive part cancels not with the next term's negative part, but with the term two positions later. The first two positive terms and the last two negative terms survive. For a gap-$g$ telescope, $g$ terms survive at each end.

The trick of adding and subtracting

When the general term is close to something telescopable but not quite, adding and subtracting a strategic piece can make it work.

Consider T_k = \frac{k}{(k+1)!}. Rewrite:

\frac{k}{(k+1)!} = \frac{(k+1) - 1}{(k+1)!} = \frac{1}{k!} - \frac{1}{(k+1)!}

The sum telescopes: \sum_{k=1}^{n} \frac{k}{(k+1)!} = \frac{1}{1!} - \frac{1}{(n+1)!} = 1 - \frac{1}{(n+1)!}.

The key move was writing k as (k+1) - 1 — adding and subtracting 1 in the numerator to make the fraction split into a telescoping difference. This "add the right thing to the numerator" trick is one of the most useful in series problems.

JEE-level problems — combining techniques

JEE problems on series often require chaining two or three techniques. Here are the most common combinations:

  1. Partial fractions + telescoping: Factor the denominator, decompose, and telescope. Gap-1 and gap-2 telescopes cover most cases.

  2. Splitting + standard formulas: Rewrite the general term as a sum of standard forms (k, k^2, k^3, k \cdot r^k) and sum each separately.

  3. V_n method + simplification: Express the general term as a product of consecutive integers, apply the V_n formula, simplify.

  4. Double sum + swapping order: Evaluate a double sum by choosing the easier order of summation.

  5. AGS + differences: When the general term has both polynomial and exponential parts, use the AGS multiply-subtract technique to reduce the polynomial degree, then telescope.

The next two examples illustrate these combinations.

Two worked examples

Example 1: Find $\displaystyle\sum_{k=1}^{n} \frac{k}{(k+1)!}$

Step 1. Rewrite the numerator.

\frac{k}{(k+1)!} = \frac{(k+1) - 1}{(k+1)!} = \frac{k+1}{(k+1)!} - \frac{1}{(k+1)!} = \frac{1}{k!} - \frac{1}{(k+1)!}

Why: writing k = (k+1) - 1 splits the fraction into two terms whose factorials differ by exactly one level — the signature of a telescoping pair.

Step 2. Recognise the telescope.

T_k = f(k) - f(k+1) where f(k) = \frac{1}{k!}.

Why: the sum of f(k) - f(k+1) from k = 1 to n collapses to f(1) - f(n+1).

Step 3. Evaluate.

\sum_{k=1}^{n} \frac{k}{(k+1)!} = \frac{1}{1!} - \frac{1}{(n+1)!} = 1 - \frac{1}{(n+1)!}

Why: all intermediate terms cancel, leaving only the first and last values of f.

Step 4. Verify for n = 3: \frac{1}{2!} + \frac{2}{3!} + \frac{3}{4!} = \frac{1}{2} + \frac{2}{6} + \frac{3}{24} = \frac{1}{2} + \frac{1}{3} + \frac{1}{8} = \frac{12 + 8 + 3}{24} = \frac{23}{24}.

Formula: 1 - \frac{1}{4!} = 1 - \frac{1}{24} = \frac{23}{24}. Confirmed.

Result: \displaystyle\sum_{k=1}^{n} \frac{k}{(k+1)!} = 1 - \frac{1}{(n+1)!}.

Partial sums of k over (k+1) factorial approaching 1A graph with n on the horizontal axis from 1 to 8 and partial sum on the vertical axis from 0 to 1.1. A dashed red line at 1 marks the limit. Dots climb from 0.5 and converge extremely quickly to 1. n Sₙ 0 0.5 1 S∞ = 1 1 2 3 4 5 6 7 1/2 5/6 23/24
The partial sums of $\sum \frac{k}{(k+1)!}$ converge to $1$ with extraordinary speed. By $n = 4$ the sum is $1 - \frac{1}{120} = 0.9917$, and by $n = 7$ the error is less than $10^{-5}$. Factorial decay is far faster than geometric or polynomial decay.

The convergence here is much faster than in any GP-based series. The factorial in the denominator shrinks each term so aggressively that a handful of terms essentially captures the entire sum.

Example 2: Evaluate $\displaystyle\sum_{i=1}^{n}\sum_{j=1}^{i} ij$

Step 1. Evaluate the inner sum.

For fixed i: \displaystyle\sum_{j=1}^{i} ij = i\sum_{j=1}^{i} j = i \cdot \frac{i(i+1)}{2} = \frac{i^2(i+1)}{2}

Why: i is a constant with respect to the inner summation over j, so it factors out.

Step 2. Set up the outer sum.

\sum_{i=1}^{n} \frac{i^2(i+1)}{2} = \frac{1}{2}\sum_{i=1}^{n} (i^3 + i^2)

Why: expanding i^2(i+1) = i^3 + i^2 splits the sum into two standard power sums.

Step 3. Apply standard formulas.

\frac{1}{2}\left[\sum_{i=1}^{n} i^3 + \sum_{i=1}^{n} i^2\right] = \frac{1}{2}\left[\frac{n^2(n+1)^2}{4} + \frac{n(n+1)(2n+1)}{6}\right]
= \frac{n^2(n+1)^2}{8} + \frac{n(n+1)(2n+1)}{12}
= \frac{n(n+1)}{24}\left[3n(n+1) + 2(2n+1)\right] = \frac{n(n+1)(3n^2 + 7n + 2)}{24}

Why: factoring out \frac{n(n+1)}{24} from both terms keeps the algebra clean.

Step 4. Factor the quadratic 3n^2 + 7n + 2 = (3n+1)(n+2).

\sum_{i=1}^{n}\sum_{j=1}^{i} ij = \frac{n(n+1)(n+2)(3n+1)}{24}

Step 5. Verify for n = 2: inner sums are 1 \cdot 1 = 1 (for i=1) and 2 \cdot 1 + 2 \cdot 2 = 6 (for i=2). Total = 7. Formula: \frac{2 \cdot 3 \cdot 4 \cdot 7}{24} = \frac{168}{24} = 7. Confirmed.

Result: \displaystyle\sum_{i=1}^{n}\sum_{j=1}^{i} ij = \frac{n(n+1)(n+2)(3n+1)}{24}.

Triangular grid of the double sum ij for n=4A triangular grid where row i has i cells. Cell (i,j) contains the product i times j. Row 1: 1. Row 2: 2, 4. Row 3: 3, 6, 9. Row 4: 4, 8, 12, 16. Row sums and total are shown. j=1 j=2 j=3 j=4 i=1 1 → 1 i=2 2 4 → 6 i=3 3 6 9 → 18 i=4 4 8 12 16 → 40 total = 65 Formula: 4·5·6·13/24 = 1560/24 = 65 ✓
The double sum $\sum_{i=1}^{4}\sum_{j=1}^{i} ij$ fills a triangular grid. Cell $(i,j)$ holds $ij$. Row sums are $1, 6, 18, 40$, and the total is $65$. The formula $\frac{4 \cdot 5 \cdot 6 \cdot 13}{24} = 65$ matches.

The triangular grid makes the structure of the dependent-index sum visible. Each row i has i cells, and the row sum \frac{i^2(i+1)}{2} grows cubically. The total is a quartic expression in n divided by 24.

Common confusions

Going deeper

If you came here for the techniques — dependent-index sums, partial fractions in series, split-and-regroup, and the JEE problem patterns — you have the toolkit. The rest of this section connects these techniques to a broader mathematical landscape.

Generating functions

A generating function encodes a sequence (a_0, a_1, a_2, \dots) as the coefficients of a power series:

A(x) = \sum_{k=0}^{\infty} a_k\,x^k

When you know A(x) in closed form, you can extract properties of the sequence — sums, convolutions, recurrences — by manipulating the power series. For example, the generating function of the constant sequence 1, 1, 1, \dots is \frac{1}{1-x} (the geometric series). Differentiating gives \frac{1}{(1-x)^2} = \sum k\,x^{k-1}, the generating function of the sequence 1, 2, 3, \dots — and this is exactly the differentiation trick from arithmetico-geometric series.

Generating functions unify many of the techniques in this article: partial fractions decompose a generating function into simpler pieces, telescoping corresponds to "clearing" factors in the generating function, and double sums arise from multiplying generating functions.

Abel summation

The discrete analogue of integration by parts is Abel summation (also called summation by parts):

\sum_{k=1}^{n} a_k\,b_k = A_n\,b_n - \sum_{k=1}^{n-1} A_k\,(b_{k+1} - b_k)

where A_k = \sum_{j=1}^{k} a_j is the partial sum. This identity transforms a sum of products into a different sum that may be easier to evaluate. It is the discrete version of \int u\,dv = uv - \int v\,du.

Abel summation is particularly powerful when a_k has a known partial sum (like A_k = \frac{k(k+1)}{2} when a_k = k) and b_{k+1} - b_k is simpler than b_k (e.g., when b_k is a slowly varying function). This technique appears in olympiad problems and in the analysis of series convergence.

Interactive: partial sums of 1 over k squared minus 1A graph where a draggable point controls the upper limit n. The curve shows the partial sums of 1 over (k squared minus 1) for k from 2 to n. A dashed line at y equals 3/4 shows the infinite sum. n Sₙ 0.75 3/4 drag to change n
Drag the point to watch the partial sums of $\sum_{k=2}^{n} \frac{1}{k^2 - 1}$ converge to $\frac{3}{4}$. The formula $\frac{1}{2}\left(\frac{3}{2} - \frac{1}{n} - \frac{1}{n+1}\right)$ gives the exact partial sum. As $n \to \infty$, the sum approaches $\frac{3}{4}$.

Where this leads next