Why Simplify Before Testing?
A term in a series can combine factorials, powers, and polynomials in a way that hides its most important features. Applying a convergence test immediately may create long expressions and make cancellation easy to miss. A short algebraic rewrite often exposes the limit with much less work.
The aim is not to change the series or guess its behavior from appearance. The aim is to express its terms exactly in a form suited to the test. In particular, factorial quotients can often be reduced to a few factors, while powers with fixed bases can be grouped into a single exponential factor.
The Tests and the Information They Use
For a series \(\sum a_n\), the ratio test examines the limit of \(\left|a_{n+1}/a_n\right|\), when the ratio is defined for all sufficiently large \(n\). If the limit is less than \(1\), the series converges absolutely; if it is greater than \(1\), the series diverges. A limit equal to \(1\) gives no conclusion.
The root test examines \(\limsup_{n\to\infty}|a_n|^{1/n}\). A value less than \(1\) implies absolute convergence, and a value greater than \(1\) implies divergence. A value equal to \(1\) is inconclusive. When the ordinary limit of the nth roots exists, it equals this limit superior.
Both tests focus on exponential-scale behavior. A fixed exponential factor, such as \(c^n\), remains visible under a ratio or nth root. A fixed power of \(n\), such as \(n^k\), contributes a factor tending to \(1\) in either calculation. This distinction is a useful guide to what to simplify and what to expect from the limit.
A Justification for Separating the Factors
Suppose a positive term can be written as \(a_n=c^n n^k\), where \(c>0\) and \(k\) is fixed. The ratio separates into an exponential contribution and a polynomial contribution:
Since \(\left(1+1/n\right)^k\to 1\), the ratio limit is \(c\). For the root test,
Taking logarithms of the polynomial factor gives \(\ln(n^{k/n})=k\ln(n)/n\), which tends to \(0\). For example, l'Hôpital's rule gives \(\ln(n)/n\to0\), since the derivative quotient is \((1/n)/1=1/n\to0\). Therefore \(n^{k/n}\to e^0=1\), and the nth-root limit is also \(c\).
This calculation explains why separating factors helps: for a fixed power of the index, the test limit is determined by the exponential factor. The polynomial factor still matters in many other convergence questions, but it does not change these particular limits when they are strictly below or above \(1\).
Worked Examples
Worked Example: Canceling Factorials Before the Ratio Test
Consider the positive terms
First expand the factorial in the denominator. Since \((n+2)!=(n+2)(n+1)n!\), canceling \(n!\) gives
Now form the ratio from this simpler expression:
Dividing the numerator and denominator of the final fraction by \(n\) shows that
Because \(3/5<1\), the ratio test proves that \(\sum a_n\) converges. Check the rewrite by multiplying the simplified expression by the canceled denominator: \(\left(3/5\right)^n/((n+1)(n+2))=3^n/[5^n(n+1)(n+2)]\), which is exactly the original term after canceling \(n!\).
Worked Example: Separating Exponential and Polynomial Factors
Let
Cancel the common power \(n^2\), and combine the fixed-base exponentials:
For the root test, take the nth root of each factor:
The factor \(n^{1/n}\) tends to \(1\), so the root limit is \(2/5\). Since \(2/5<1\), the root test proves that \(\sum a_n\) converges. The terms are positive, so this is also absolute convergence.
Check the rewrite by multiplying \(n(2/5)^n\) out: it equals \(n2^n/5^n\), which is exactly \(n^3 2^n/(n^2 5^n)\) after canceling \(n^2\). The polynomial factor contributes \(1\) to the root limit; the exponential factor supplies the decisive value \(2/5\).
Worked Example: Reducing a Factorial Quotient in a Ratio
Consider
Group the fixed-base powers as \((6/5)^n\), so \(a_n=(n!)^2(6/5)^n/(2n)!\). For the ratio, write the next term and divide by the current one:
Use \((n+1)!=(n+1)n!\) and \((2n+2)!=(2n+2)(2n+1)(2n)!\). The factorial and exponential factors then cancel to give
Divide the last fraction's numerator and denominator by \(n\):
Since \(3/10<1\), the ratio test proves convergence. As a check, the uncanceled ratio has the factorial quotient \(((n+1)!)^2/(n!)^2=(n+1)^2\) and the denominator quotient \((2n+2)!/(2n)!=(2n+2)(2n+1)\); these are precisely the factors used in the simplified ratio.
A Common Cancellation Mistake
In the first example, it would be incorrect to cancel \(n!\) against \((n+2)!\) and write \(n!/(n+2)!=1/(n+2)\). The factorial expansion shows why this loses a factor:
For instance, at \(n=1\), the exact quotient is \(1!/3!=1/6\), while the incorrect expression \(1/(n+2)\) gives \(1/3\). The missing factor is \(n+1=2\). The corrected term is \(\left(3/5\right)^n/((n+1)(n+2))\), not \(\left(3/5\right)^n/(n+2)\). Even when a flawed simplification happens to lead to the same broad test conclusion, it is not an equivalent expression and cannot be trusted.
A Reliable Simplification Routine
Before applying either test, use a brief sequence of algebraic checks. The point is to reach a form where each factor's role is visible, not to perform extra algebra for its own sake.
Distinguish factorials, fixed-base exponentials, and powers or other polynomial factors in \(n\).
Expand factorials only as far as needed to expose common factors. For example, \((n+r)!/n!\) becomes a product of \(r\) consecutive factors when \(r\) is a positive integer.
Rewrite \(b^n/c^n\) as \((b/c)^n\), and combine fixed powers of \(n\) using the ordinary exponent rules.
For the ratio test, compare the simplified term at \(n+1\) with the term at \(n\). For the root test, take the nth root factor by factor.
Confirm that the rewrite equals the original term, then compare the test limit with \(1\). A limit of \(1\) is inconclusive for either test.
What Simplification Can and Cannot Tell You
Simplifying reveals which components survive the test calculation. Fixed-base exponentials are often decisive: \((b/c)^n\) contributes \(b/c\) to both the ratio and nth-root limits. Fixed powers of \(n\) generally contribute a factor tending to \(1\). Factorial quotients may reduce to products of consecutive integers, whose behavior must still be included in the ratio rather than discarded.
These observations do not mean that polynomial factors are unimportant in every convergence question. In particular, if a ratio or root limit equals \(1\), the test does not decide convergence, and the remaining structure may matter. Simplification helps identify that boundary case; it does not turn an inconclusive test into a conclusion.
For signed terms, apply the ratio and root tests to \(|a_n|\). This makes the relevant quantities nonnegative and tests absolute convergence. Before calculating, ensure that any ratio used is defined for all sufficiently large indices; if terms are eventually zero or the ratio is not defined, the usual ratio-limit calculation needs separate attention.
Check Your Understanding
For each question, simplify the term before deciding which limit to calculate.
- Rewrite \(\frac{(n+3)!}{n!}\) as a product of consecutive factors.
- Simplify \(\frac{n^4 3^n}{n^2 7^n}\) by separating its polynomial and exponential factors.
- Why does the fixed factor \(n^k\), for fixed \(k\), contribute \(1\) to an nth-root limit?
- For \(a_n=\frac{(n!)^2}{(2n)!}\), write \(a_{n+1}/a_n\) using factorial recurrences before taking a limit.
- What does a ratio-test or root-test limit equal to \(1\) establish about convergence?
