A sequence is arithmetic if subtracting any term from the one before it gives the same difference every time, and geometric if dividing any term by the one before it gives the same ratio every time. The test only works on consecutive terms, and it fails immediately if a division step hits zero. Solve one example of each before trusting a label.
The Two Definitions in Plain Terms
An arithmetic sequence adds a fixed amount with each step. That amount is the common difference, usually written d, and it can be positive, negative, or zero. The sequence 4, 7, 10, 13 has d = 3 because 7 β 4, 10 β 7, and 13 β 10 all equal 3. A geometric sequence multiplies by a fixed amount instead. That amount is the common ratio r, and it can be any constant except that a ratio of zero behaves strangely. The sequence 3, 12, 48, 192 has r = 4 because each term is four times the last.
The distinction matters because the two families behave nothing alike over long spans. Arithmetic growth is linear: the 100th term of 4, 7, 10, 13 is 4 + 99 Γ 3 = 301. Geometric growth is exponential, and the 15th term of 3, 12, 48, 192 is 3 Γ 4^14 = 805,306,368. That gap is why mistaking one for the other wrecks forecasts built on real data, from loan schedules to population counts.
Two formulas anchor everything downstream. The nth term of an arithmetic sequence is a + (n β 1)d, where a is the first term. The nth term of a geometric sequence is a Γ r^(n β 1). Every later method in this article is a way of deciding which formula structurally cannot be wrong for the data in front of you.
One naming trap catches almost everyone. In geometry class, "geometric" refers to shapes and angles. In sequence theory, it refers only to repeated multiplication. A geometric sequence has nothing to do with triangles, and a shape with sides in arithmetic progression has nothing to do with adding a constant. Keep the two meanings separate in your head and the word stops confusing you.
Step 1: Write the Terms and Subtract Consecutive Pairs
Start by laying the terms out with enough space to write between them. Given 6, 14, 22, 30, you write 14 β 6 = 8, 22 β 14 = 8, and 30 β 22 = 8 in the gaps. One number per gap, no shortcuts. This is the fastest arithmetic test available, and it costs about ten seconds for four terms.
If every gap shows the same value, the sequence is arithmetic and you already know d. Record it. If the differences vary, write them down anyway, because a set of differences that changes steadily is itself a clue you'll use later. The differences 2, 4, 6, 8 mean the differences are going up by 2 each time, which is a second-level pattern.
Watch for near-misses. A sequence like 5, 9, 13, 18 looks arithmetic until the final gap shows 5. One mismatched difference disqualifies arithmetic entirely unless you've copied a term wrong. Recheck the original list before moving on, because transcription errors are the most common cause of a failed test.
Use a simple table when the sequence runs past five terms:
| Position | Term | Subtract previous | Running result |
|---|---|---|---|
| 1 | 6 | β | β |
| 2 | 14 | 14 β 6 | 8 |
| 3 | 22 | 22 β 14 | 8 |
| 4 | 30 | 30 β 22 | 8 |
Four rows, constant column, arithmetic confirmed.
Step 2: Divide Each Term by the One Before It
Run the division test on the same grouping. For 5, 15, 45, 135, compute 15 Γ· 5 = 3, 45 Γ· 15 = 3, and 135 Γ· 45 = 3. Each answer is the common ratio r. If every quotient matches, the sequence is geometric and you have your multiplier.
If the division results scatter β say 2, 1.5, 2.25 β the sequence is not geometric. Non-integer ratios are perfectly legal: 8, 4, 2, 1 has r = 0.5, and 100, 130, 169 has r = 1.3. Decimal answers deserve extra care because rounding can disguise a mismatch, so carry at least three decimal places before you compare.
Check negative and fractional ratios too. The sequence 2, β6, 18, β54 alternates sign because r = β3. The sequence 16, 4, 1, 0.25 divides cleanly by 4 each time. If the signs flip back and forth on every term, that alternation is itself evidence of a negative ratio rather than an arithmetic pattern.
Order matters in division. Compute later term divided by earlier term, never the other way around, or your ratio gets inverted. A consistent quotient of 3 and a consistent quotient of one-third are both geometric, but only one matches the sequence if the multiplier truly is 3.
Step 3: Run Both Tests Before You Decide
Many learners try subtraction, then stop when it fails. That's half a test. A sequence could pass both checks in trivial cases and pass neither in others, so run the subtraction column and the division column side by side before declaring anything.
Sample a few seeds yourself. The first has a difference of 10 and a ratio of about 3; the second has a difference of 3 and no ratio; the third has a difference of 1 and no ratio. Interleave the tests on paper and you'll notice a pattern: repeated differences of 10 with dramatically growing terms cannot coexist beyond the first term or two, which tells you arithmetic immediately.
Here is a comparison view you can reuse as a scratch template:
| Test | Operation | What you compare | What a match proves |
|---|---|---|---|
| Arithmetic | Later minus earlier | Every difference equals d | Constant addition |
| Geometric | Later divided by earlier | Every quotient equals r | Repeated multiplication |
Neither test works on non-consecutive terms. Comparing the 2nd term to the 1st and the 4th to the 3rd is correct; comparing the 3rd to the 1st is a common error that produces meaningless ratios.
If neither test passes, say so explicitly. Plenty of sequences are neither, and forcing a label on 1, 4, 9, 16 produces a wrong answer. Name the pattern you do see instead of inventing one.
Step 4: Recognize Sequences That Pass Both
Constant sequences pass both tests. The list 7, 7, 7, 7 has d = 0 because every subtraction gives 0, and r = 1 because every division gives 1. Both labels are technically valid, and which one you use depends on the formula you intend to apply. For an nth-term question, using d = 0 gives 7 + (n β 1) Γ 0 = 7; using r = 1 gives 7 Γ 1^(n β 1) = 7. Same answer, two roads.
When a problem expects one answer, arithmetic is the safer label for a constant sequence because the zero difference is unambiguous and no division is involved. Division by a nonzero constant is fine, but constant sequences are where students first notice the overlap, so be ready to explain it. Textbooks typically accept either label and grade on the reasoning.
The overlap stops the moment a term changes. A sequence with a single repeated value near the start, like 5, 5, 10, fails both tests once the third term arrives β the differences become 0 then 5, and the ratios become 1 then 2.
Decide by the question you're answering, not by preference. If you need to project forward many steps, test whether exponential shape or linear shape fits the data; the constant case is the only one where the two shapes coincide.
Step 5: Watch for Zero, Missing Terms, and Sign Flips
Division fails when the term you're dividing by is zero. The sequence 0, 0, 0, 0 has infinitely many ratios and no defined one; it is still arithmetic with d = 0. For a sequence like 0, 5, 10, 15, the first ratio is undefined β 5 Γ· 0 has no value β while the rest are impossible, so geometric cannot apply. Treat a leading zero as an automatic disqualifier for geometric labeling unless you know the sequence's generating rule.
Missing terms reverse the workflow. Given 4, __, 16, 28, you're asked to fill the gaps, not classify forms. The subtraction route gives 4, 10, 16, 22 β differences of 6 β while treating the visible terms as an arithmetic skeleton with d = 12 gives 4, 16, 16, 28, which is not a sequence at all. Work from the visible terms outward and confirm each filled value passes the test.
Sign flips carry information. Arithmetic sequences with negative d decline steadily and never alternate; geometric sequences with negative r do alternate. If the sign flips with every step, suspect a negative ratio and test division immediately, because subtraction would give inconsistent differences and mislead you.
Interleaved sequences punish single-test thinking. The list 1, 100, 2, 200, 3, 300 alternates two arithmetic sequences, and its overall differences are 99, β98, 198, β197, 297 β clearly not constant. Split it into the odd-position terms 1, 2, 3 and the even-position terms 100, 200, 300, and both halves turn out arithmetic. Real-world data often hides patterns this way.
Step 6: Compute d and r Once the Test Passes
Confirm the constant and then write it in the standard form. For arithmetic, d is the subtraction answer and appears in the nth-term formula a + (n β 1)d. For geometric, r is the division answer and appears in a Γ r^(n β 1). If no constant emerges, there is nothing to compute.
Test the formula against a term you didn't use. Given 3, 7, 11, 15, the difference is 4 and the eighth term is 3 + 7 Γ 4 = 31. Then extend the list by hand β 19, 23, 27, 31 β and the formula matches. This cross-check catches sign errors before they spread into later work.
Geometric formulas need the same cross-check. Given 2, 10, 50, 250, the ratio is 5 and the sixth term is 2 Γ 5^5 = 6,250. Write out three more terms manually and compare. If the two disagree, you've either miscounted the exponent or the sequence wasn't truly geometric.
Recompute d and r from different pairs. Using 15 β 11 and 135 Γ· 45 gives the same constants as the first pairs did, and if it doesn't, the sequence fails the test after all. Never skip the second pair check on short sequences, where one coincidence can fake a pattern.
Step 7: Confirm the Label Holds for Terms You Haven't Seen
A label earned on four terms is a hypothesis until you stress it. Suppose a sequence passes the arithmetic test with d = 5; predicting that term 40 is 4 + 39 Γ 5 = 199 is a real claim about data you don't have. Whether the claim is true depends on the rule behind the data, not on the four terms alone. Sequence problems usually assume the simplest rule that fits the given terms, so say that assumption out loud when you present an answer.
Real measurements rarely obey either form exactly. Suppose a shop's daily sales are 120, 138, 156, 174. The differences are constant at 18, so subtracting treats this as arithmetic and predicts 354 on day 14 β a fit, but growth by fixed addition. The ratios run 1.15, 1.13, 1.12, all near but not equal, so dividing would tempt a geometric label the data doesn't earn. Fit real data with whichever test returns a constant, and if neither does, prefer a regression rather than a label.
Test the fringes of your data. If the sequence is long, check whether the constant holds in the middle where transcription errors cluster. A single anomalous term in an otherwise clean arithmetic run often points to a typo rather than a new pattern.
State the rule and the confidence separately. "Arithmetic with d = 4, consistent across eight terms" is an honest answer. "Definitely arithmetic" for a sequence built from three terms promises more than three terms can show.
Step 8: Apply the Label to nth-Term and Sum Word Problems
The label decides which formula solves the question, so classify first and compute second. For an arithmetic sequence, the nth term is a + (n β 1)d and the sum of n terms is (n/2)(a + last term). For a geometric sequence, the nth term is a Γ r^(n β 1) and the finite sum is a(1 β r^n)/(1 β r) when r β 1.
Work a ticket example. Suppose ticket prices start at $12 and rise by $4 each row of seats; that's arithmetic with d = 4, so the price in row 10 is 12 + 9 Γ 4 = $48, and the sum of rows 1 through 10 is (10/2)(12 + 48) = $300. Suppose instead a handout photocopies double in count each day; that's geometric with r = 2, so day 6 counts 1 Γ 2^5 = 32 and the total through six days is 1 Γ (2^6 β 1)/(2 β 1) = 63.
Know when a label is useless. Bouncing-ball heights, compartmental spread, and any situation where the multiplier is itself the object of interest are geometric through and through, and the arithmetic test will fail them by a mile. Similarly, fixed monthly payments with no interest are arithmetic; add any compounding and the geometric test is the only one that applies.
Check that the sum formula's assumptions fit. The geometric formula requires r β 1, which excludes constant sequences; use arithmetic for those. The arithmetic sum formula requires a constant difference throughout the range you're summing, so verify the constant one last time before plugging in.
Step 9: Avoid the Three Costliest Mistakes
Mismatched pairs lead the error list. Mixing consecutive and non-consecutive comparisons, like comparing term 4 with term 1, produces ratios and differences that no formula can use. Always compare adjacent terms unless the problem explicitly asks about a gap.
Reading too much into three points comes next. Any three terms can be fit to an arithmetic form if d happens to match, and many short lists fit both forms by accident. Find a fourth term and test against it before committing to a label on a real problem.
The third mistake is switching labels mid-problem. Compute one full example both ways β say 3, 6, 12, 24 arithmetic-first, then geometric-second β and watch which operations produce constants. Once a label earns a constant on consecutive pairs, keep that label for the whole problem; reclassifying halfway invalidates the arithmetic you've already done.
One more trap hides in calculator work. Double precision floating point handles roughly 15 to 17 significant decimal digits before rounding, per the IEEE 754 standard, and the geometric sums in finance and science routinely pass that boundary. A computed ratio can drift in the last digits and look nonconstant, so compare with the rounding you'd expect for the number of digits you entered, not for exact equality.
FAQ
Can a sequence be both arithmetic and geometric?
Yes, but only constant sequences qualify. The list 9, 9, 9, 9 has d = 0 and r = 1, so both formulas produce the same results. Any sequence with at least two distinct terms can pass at most one test.
What if the sequence is neither arithmetic nor geometric?
Name the other pattern instead. The squares 1, 4, 9, 16 have odd differences 3, 5, 7, and the Fibonacci form 1, 1, 2, 3, 5 adds adjacent terms. Neither family's test applies, and saying so is a correct answer.
How do I handle a zero in the sequence?
Test subtraction first. A zero term only breaks geometric labeling, so if the differences are constant the sequence is arithmetic and the zero causes no trouble. Division by zero disqualifies the geometric test entirely.
Is 2, 4, 8, 16 arithmetic or geometric?
It's geometric with r = 2. The differences 2, 4, 8 vary because the terms themselves are growing by doubling, so subtraction fails. Division gives 2 at every step, which is the definition of geometric.
Why does my computed ratio change in the last decimal place?
Floating-point rounding, not a broken pattern. Nightly compound interest of 7 percent annual, applied as 12 monthly steps, turns each term into the previous one multiplied by about 1.005654 β a full-decimal value that only survives to 16 or so significant digits. Compare with a sensible tolerance before calling the sequence non-geometric.
Can a geometric sequence have a ratio between 0 and 1?
Yes. The ratio 0.5 shrinks each term by half, giving 64, 32, 16, 8, and the ratio β0.5 flips signs while shrinking. Neither shape can be produced by fixed addition, so subtraction fails and the geometric label holds.
What's the quickest way to check an arithmetic sequence on a calculator?
Compute one gap, then add it to the last term and compare with the given next term. For 7, 19, 31, 43, compute 19 β 7 = 12, then 31 + 12 = 43. The final match confirms the arithmetic label, and one calculation beats four.
Does a sequence with a constant ratio have to be integers?
No. Geometric sequences with r = 1.25 and first term 16 produce 16, 20, 25, 31.25, which include a fractional term. Arithmetic sequences with integer d can also produce fractions if the first term is fractional, so check the ratio or difference rather than the term values.