Expose cancellation before taking a limit

Try

Write S=1+r+r²+r³ and rS=r+r²+r³+r⁴ directly below it. Subtract the second line from the first. Nearly every term cancels. Predict the same pattern for N+1 terms before passing to any infinite expression.

Predict

For r=1/2, the uncanceled final power shrinks. For r=−1, it alternates without shrinking. For r=2, it grows. Predict why “r less than 1” is not the correct convergence condition when negative ratios are allowed.

Experiment

Compare finite endpoint terms in two types of cancellation.

Family	Finite sum	Remaining tail expression
Geometric, n=0 through N	(1−r^(N+1))/(1−r), r≠1	r^(N+1)/(1−r) when convergent
Telescoping, n=1 through N	Σ(1/n−1/(n+1))	S_N=1−1/(N+1)

For the telescoping example, write the first three terms and the last two terms before crossing anything out. The surviving first and last quantities show why the answer is not zero, despite most internal terms canceling.

Understand

Subtracting rS_N from S_N gives (1−r)S_N=1−r^(N+1). If r≠1, divide by 1−r. When |r|<1, the final power tends to zero and the infinite sum is 1/(1−r). Multiplying every term by a fixed first-term factor A gives A/(1−r). At r=1 the finite sum is N+1; at r=−1 it oscillates. Magnitudes greater than 1 also fail the necessary term condition.

Starting index changes the first term. The stream 3(1/4)ⁿ beginning at n=0 sums to 4. Beginning at n=1 instead removes the initial 3, so the sum is 1. Always identify the first actual term rather than blindly labeling a coefficient as the first term.

For 1/[n(n+1)], partial fractions gives 1/n−1/(n+1). The Nth partial sum is 1−1/(N+1), tending to 1. For 1/[n(n+2)], the decomposition is half of 1/n−1/(n+2). Two initial terms and two tail terms survive, giving the limit (1/2)(1+1/2)=3/4. Telescoping does not mean canceling symbols across an undefined infinite expression; it means deriving a valid finite identity and then taking its limit.

A synthetic discounted payment model makes indexing consequential. Payments of 100 at the end of each year with a fixed annual discount factor 4/5 have present values 80,64 and so on. The sum is 80/(1−4/5)=400. A payment made immediately would add a different initial term. This example assumes a perpetual stream and a fixed discount factor; it does not establish a real investment value or account for default, inflation uncertainty or changing payments.

The remainder of a geometric series has an exact formula, not merely a convergence classification. That makes geometric tails useful benchmarks for later tests: if a complicated nonnegative tail is bounded by a geometric tail, it inherits a finite bound. Some alternating geometric tails have signed remainders, so report whether an error claim concerns magnitude or direction.

Apply

For each practice sum, write its starting index, first term and finite partial sum. In the shifted telescoping problem, retain both boundary terms. State |r|<1 explicitly whenever you use an infinite geometric formula, and explain assumptions when interpreting a sum as a physical or economic model.

Continue

Many series have no simple finite cancellation. Next use necessary term behavior and positive integral comparisons to classify their tails and estimate the uncomputed remainder.