What Is the Ratio Test?
The ratio test decides whether a series Σa(n) converges by asking how each term compares to the one before it. Compute the limit of consecutive-term ratios:
L = limn→∞ |a(n+1) / a(n)|
- L < 1: the series converges absolutely, eventually every term is a fixed fraction of its predecessor, so the tail is dominated by a shrinking geometric series.
- L > 1: the series diverges, terms eventually grow so they can't even approach zero.
- L = 1: the test is silent, convergence and divergence both remain possible, and another test must decide.
The comparison with geometric series is the entire proof idea: if the ratios eventually stay below some r < 1, the terms are trapped underneath a convergent geometric envelope. The ratio test is the geometric series' convergence criterion, exported to every series that behaves geometrically in the long run, which, conveniently, includes most series built from exponentials and factorials.
How to Use This Calculator
Enter the general term as an expression in n. The calculator samples |a(n+1)/a(n)| at n = 10, 100, 1,000, 10,000, 100,000, reads off the limiting value, and applies the three-way verdict, with the sampled ratios plotted against the critical line at height 1, so you can watch L settle above, below, or onto the boundary. Factorial-flavored terms are the test's home turf; note that this calculator's parser handles products and powers (write n! effects via ratios like the examples), and terms mixing exponentials against polynomials showcase the test at its clearest.
Worked Example
Test Σn²/2ⁿ. The ratio is:
|a(n+1)/a(n)| = [(n+1)²/2ⁿ⁺¹] · [2ⁿ/n²] = ½ · (1 + 1/n)²
As n → ∞ the (1 + 1/n)² factor melts to 1, leaving L = ½ < 1: converges absolutely. The polynomial n² inflates each ratio slightly, but the exponential's steady halving wins, the general moral being that exponentials beat polynomials, which the sampled ratios (0.605 at n = 10, 0.505 at n = 100, …) confirm numerically as they descend toward ½.
When to Reach for the Ratio Test
| Term shape | Ratio test outcome |
|---|---|
| Exponentials: rⁿ, and factorials n! | Decisive, ratios collapse cleanly |
| Mixed: polynomial × exponential | Decisive, exponential part sets L |
| Pure powers of n: 1/nᵖ | Always L = 1, use the p-test instead |
| nth powers: (stuff)ⁿ | Works, but the root test is often cleaner |
Common Mistakes to Avoid
- Treating L = 1 as "converges." L = 1 means no information: 1/n² (convergent) and 1/n (divergent) both give L = 1. The inconclusive chip demonstrates it live.
- Forgetting the absolute value. The test examines |a(n+1)/a(n)|; for alternating series the signs are stripped first, and L < 1 then delivers the stronger conclusion of absolute convergence.
- Dividing in the wrong order. The next term goes on top: a(n+1)/a(n). Inverting yields 1/L and flips every verdict.
- Concluding divergence from a few large early ratios. Only the limiting behavior counts, early terms may bounce before the trend locks in, which is why the calculator samples far out and plots the trajectory.
Real-World Applications
The ratio test's biggest client is the power series: applying it to Σcₙxⁿ produces the radius of convergence, the |x| range where a Taylor expansion is trustworthy, making this test the quality-control instrument behind every series-based computation in physics and engineering. (The Radius of Convergence Calculator runs exactly that computation.) The exponential series Σxⁿ/n! passes with L = 0 for every x, which is why eˣ, sin, and cos are computable everywhere, a fact your calculator's firmware relies on at every keypress.
Beyond pure math, consecutive-term ratios diagnose growth processes: population models, compound returns, and branching processes (each generation a ratio times the last) are healthy, stable, or explosive according to whether their effective L sits below, at, or above 1, epidemiology's reproduction number R₀ is culturally the same object, with the same critical threshold at 1. And in numerical computing, the observed ratio of successive correction terms tells an algorithm when its series expansion has converged enough to stop, a live, running ratio test inside production code.
Frequently Asked Questions
Why does L = 1 leave the test inconclusive?
Because at ratio exactly 1 the geometric-comparison proof loses its grip: the terms shrink slower than any geometric series, and their fate depends on finer structure the ratio can't see. Both π²/6-convergent 1/n² and divergent 1/n live in this blind spot, the p-test exists precisely to adjudicate it.
How does the test behave with factorials?
Beautifully, factorials are its signature win. For Σxⁿ/n! the ratio is |x|/(n+1) → 0, so L = 0 < 1 for every x: the series converges everywhere, faster than any geometric series eventually. Any n! in a term is a strong hint to try ratios first.
What's the difference between the ratio test and the root test?
Both compare against geometric decay; the root test reads the decay rate from |a(n)|^(1/n) instead of consecutive ratios. When the ratio limit exists, both give the same L; the root test is strictly stronger in edge cases and cleaner for terms that are explicit nth powers.
Does L < 1 tell me what the series converges to?
No, the test certifies convergence without locating the sum. Finding the value takes other machinery: closed forms, telescoping, or numerical partial sums via the Partial Sum Calculator.
Can the sampled ratios mislead the numerical verdict?
Rarely, but slowly-converging ratio sequences (like ones approaching 1 from below at rate 1/n) can look settled prematurely. The calculator samples out to n = 100,000 and declares "inconclusive" within a tolerance band around 1 rather than forcing a call, mirroring the honest analytic answer.
Does the ratio test handle alternating series?
Through absolute values, yes: it tests Σ|a(n)|, so L < 1 certifies absolute convergence, stronger than plain convergence. When L = 1 for an alternating series (like the alternating harmonic series), the ratio test is silent and the alternating series test takes over, often rescuing convergence the ratios can't see.
What happens when the ratio test limit equals exactly 1?
The test is inconclusive and gives no information. Series like the harmonic series and the p-series with p equal to 2 both produce a limit of 1 yet behave oppositely, so you must switch to a comparison, integral, or p-series test instead.