A number series is a sequence of numbers arranged according to a hidden rule. Your job is either to find the missing term, identify the next term, or — in the harder variant — spot the one wrong number that breaks the pattern.
Think of it like a locked box with a combination. Once you figure out the combination (the rule), every term falls into place instantly. The rule is always one of a handful of patterns, and after enough exposure, your brain starts recognising them the way a chess player recognises board positions — not by calculation, but by shape.
Here is the useful analogy: imagine each term in the series is a step on a staircase. In a simple arithmetic series, all steps are the same height. In a geometric series, each step is a fixed multiple of the previous one. In trickier series, the step heights themselves follow a pattern — they might be growing by a fixed amount, or doubling, or following squares. Your first job is always to look at the staircase, not just the steps.
The key mental shift for SBI Clerk: stop trying to find the formula first. Find the differences first. If the differences are constant, you're done. If not, look at second-order differences (differences of differences). If those are constant, it is a quadratic pattern. If the differences are multiplied by a fixed ratio, it is geometric growth in the differences. If the terms themselves multiply by an increasing integer (×2, ×3, ×4...), think factorial relatives. If the terms are clean numbers like 1, 2, 6, 24, 120, think factorials directly.
The reason number series problems are considered high-value in SBI Clerk prelims is that once you train the pattern-recognition reflex, these are among the fastest questions in the paper — often solved in under 30 seconds with the right approach.
Write out the terms. Immediately compute the differences between consecutive terms.
Example: 3, 8, 18, 38, 78, ?
Differences: 5, 10, 20, 40 — each doubles. Next difference = 80. Answer = 78 + 80 = 158.
Example: 6, 20, 42, 72, 110, ?
Differences: 14, 22, 30, 38 — each increases by 8. Next difference = 46. Answer = 110 + 46 = 156.
These are the most common traps. The rule looks like: next = previous × a ± b.
When you see this, do not look for differences first — they will be irregular and mislead you. Instead, check: can you get the second term by multiplying the first by 2 (or 3) and then adjusting by a small constant?
Example: 5, 11, 23, 47, 95, ?
Check: 5 × 2 = 10, not 11. Try 5 × 2 + 1 = 11. Then 11 × 2 + 1 = 23. Then 23 × 2 + 1 = 47. Pattern confirmed. Answer = 95 × 2 + 1 = 191.
These series have a signature look: the terms grow fast, and consecutive ratios are 2, 3, 4, 5, 6... or some shifted version.
Example: 7, 14, 42, 168, 840, ?
Ratios: 14/7 = 2, 42/14 = 3, 168/42 = 4, 840/168 = 5. Next multiply by 6: 840 × 6 = 5040.
This is closely related to factorials. Note that 7 × 2 × 3 × 4 × 5 × 6 = 7 × 6! / 1! = 7 × 720 = 5040. Useful as a check.
When you see 1, 2, 6, 24, 120, 720 — or any subset of those exact numbers — you are looking at n! (n factorial).
1! = 1, 2! = 2, 3! = 6, 4! = 24, 5! = 120, 6! = 720
Variants include n! + 1, n! - 1, or starting partway through the sequence.
Example: 11, 13, 17, 25, 41, ?
Differences: 2, 4, 8, 16 — powers of 2. Next difference = 32. Answer = 41 + 32 = 73.
This is different from a geometric series in the terms themselves — the differences are geometric.
Some series combine multiplication with an addend that itself increases.
Example: 4, 9, 20, 43, 90, ?
Check multiply-by-2 plus something: 4×2=8, 8+1=9. Then 9×2=18, 18+2=20. Then 20×2=40, 40+3=43. Then 43×2=86, 86+4=90. Pattern: ×2 + incrementing integer (1, 2, 3, 4, ...). Next: 90×2 + 5 = 185.
Apply the same pattern-finding approach on all terms except the suspected wrong one. Once the pattern is clear from the correct terms, plug in to find what the wrong term should have been. The wrong number is usually ±1 or ±2 off from the correct value — examiners make subtle traps.
Always attempt in this order:
When you write the series and immediately below it write the differences, you create a "difference ladder." Keep taking differences row by row until you hit a constant row. The level at which differences become constant tells you the series type: constant at row 1 = arithmetic; constant at row 2 = quadratic/second-order; constant (as ratio) at row 1 = geometric in differences.
Example: Series 6, 20, 42, 72, 110 → Row 1 differences: 14, 22, 30, 38 → Row 2: 8, 8, 8 (constant). So extend: next Row 1 difference = 38 + 8 = 46, next term = 110 + 46 = 156.
Standard approach (trial patterns one by one): ~45 seconds. Difference Ladder (systematic): ~15 seconds.
For series that grow fast (each term roughly 2× or 3× the previous), skip the difference ladder entirely. Divide term 2 by term 1, term 3 by term 2, and so on. If ratios are 2, 3, 4, 5 (increasing by 1), you have an incremental-multiplier series. If ratios are exactly equal, you have a geometric series.
Example: 7, 14, 42, 168, 840 → ratios 2, 3, 4, 5 → next ratio = 6 → answer 840 × 6 = 5040. This saves you from computing large differences (which would be 7, 28, 126, 672 — unhelpful).
Computing large differences and failing to spot pattern: ~50 seconds. Ratio check: ~10 seconds.
Memorise the factorial sequence cold: 1, 2, 6, 24, 120, 720, 5040. Any series containing two or more of these exact values is almost certainly factorial or factorial-based. You do not need to calculate — recognition alone gives you the answer.
Spotting 1, 2, 6, 24, 120 and realising immediately it is factorials takes under 3 seconds. Trying to find differences (1, 4, 18, 96 — which look random) and then ratios: ~40 seconds of confusion before seeing the pattern.
When differences are irregular and ratios are not clean integers, try: next = current × 2 ± small constant. Compute (term2 - term1×2). If that value is the same constant throughout, you found the rule.
Example: Series 5, 11, 23, 47, 95 → 11 - 5×2 = 1. Check: 23 - 11×2 = 1. Check: 47 - 23×2 = 1. Rule confirmed: ×2 + 1. Answer: 95×2 + 1 = 191.
This test takes 3 subtractions, ~8 seconds. Guessing the pattern through trial-and-error: ~40 seconds.
For series like 4, 9, 20, 43, 90, where ×2+constant does not quite work with a fixed constant, check if the addend is itself incrementing (1, 2, 3, 4...). Write out: term2 - term1×2, term3 - term2×2, etc. If those residuals are 1, 2, 3, 4 — you have the rule.
Example: 9-8=1, 20-18=2, 43-40=3, 90-86=4 → next: 90×2 + 5 = 185.
This adds only 4 extra subtractions to the Multiply-and-Adjust test. Total time from series to answer: ~12 seconds vs ~50 seconds of blind pattern-hunting.
In the exam hall, run this mental checklist in order — stop the moment a check succeeds:
Step 1 — Visual Scan (2 sec): Are any terms recognisable as factorials (1, 2, 6, 24, 120, 720)? If yes, verify and answer.
Step 2 — Fast Growth Check (3 sec): Do terms roughly double or triple each step? If yes, compute ratios. Constant ratio = geometric. Ratios 2, 3, 4, 5... = incremental multiplier. Both types solved in under 10 seconds.
Step 3 — Difference Ladder (8 sec): Compute first-order differences. Constant? Done. Not constant? Compute second-order differences. Constant? Quadratic pattern — extend and answer.
Step 4 — Differences Doubling (5 sec): If first-order differences are 5, 10, 20, 40... (doubling), next difference = previous × 2. Extend and answer.
Step 5 — Multiply-and-Adjust Test (8 sec): Compute term_n+1 - term_n × 2. If constant → rule is ×2 + constant. If incrementing (1, 2, 3...) → rule is ×2 + incrementing addend.
Step 6 — Elimination (last resort, 15 sec): Use the answer choices. Plug back to verify which option fits the pattern suggested by the last two known differences or ratios.
Do not spend more than 60 seconds on any single number series question in SBI Clerk prelims. If you are past Step 5 without a clear pattern, mark your best guess using Step 6 and move on.
Why this question: Tests whether you recognise the factorial pattern immediately or get misled by computing differences (1, 4, 18, 96 — which look random and waste time).
Solving path: Apply the Factorial Fingerprint trick. The terms 1, 2, 6, 24, 120 are 1!, 2!, 3!, 4!, 5! exactly. Next term = 6! = 720. No calculation needed beyond recognition.
Why this question: Tests the incremental-multiplier pattern where ratios are 3, 4, 5, 6, 7 — not the more obvious 2, 3, 4, 5 starting from 2. The first multiplier being 3 can throw you off.
Solving path: Use the Ratio Check trick. 12/4 = 3, 48/12 = 4, 240/48 = 5, 1440/240 = 6. Ratios are consecutive integers starting at 3. Next ratio = 7. Answer = 1440 × 7 = 10080.
Why this question: Classic ×2 + 1 series. The trap is that the differences (6, 12, 24, 48) look geometric — you might mistakenly apply a ratio rule to the differences instead of recognising the underlying multiply-and-adjust rule.
Solving path: Differences are 6, 12, 24, 48 — doubling. You could use Step 4 (next difference = 96, answer = 95 + 96 = 191). Or apply the Multiply-and-Adjust test: 11 - 5×2 = 1, 23 - 11×2 = 1. Rule is ×2 + 1. Either path gives 191.
Why this question: Represents second-order arithmetic — the differences of differences are constant. This is the cleanest test of the Difference Ladder technique.
Solving path: First-order differences: 14, 22, 30, 38. Second-order differences: 8, 8, 8. Pattern confirmed. Extend: next first-order difference = 38 + 8 = 46. Answer = 110 + 46 = 156.
Why this question: Differences are powers of 2 (2, 4, 8, 16). Examiners bank on aspirants computing the series as geometric in the terms themselves (wrong approach) rather than geometric in the differences (correct approach).
Solving path: Compute differences: 2, 4, 8, 16. These are 2^1, 2^2, 2^3, 2^4. Next difference = 2^5 = 32. Answer = 41 + 32 = 73.
Taking differences when the series multiplies. For fast-growing series (4, 12, 48, 240...), computing differences gives 8, 36, 192 — useless. Always check ratios first when terms are growing rapidly. The Ratio Check should be your first move on any series where each term is clearly several times larger than the previous.
Missing the incrementing addend. When the ×2 + constant test gives different constants (1, 2, 3, 4 instead of 1, 1, 1, 1), aspirants often conclude "no pattern" and waste 30 seconds re-checking. Train yourself to immediately look at whether those residuals are themselves in arithmetic progression.
Confusing a factorial series with a multiply-by-incrementing-integer series. They are related but not identical. 7, 14, 42, 168, 840 is not a factorial series — it is 7 × 2!, 7 × 3!, 7 × 4!... essentially, but it is faster to just track the ratios (×2, ×3, ×4...) rather than forcing a factorial frame.
Stopping at first-order differences. When first-order differences are not constant, many aspirants guess a wrong pattern instead of computing second-order differences. Always go one level deeper before switching strategy.
Spending too long on wrong number series. The wrong number will always be close in value to the correct number. Once you identify the correct pattern from the other terms, the "wrong" term will be off by a predictable amount based on the rule — check the options to see which one is 1 or 2 units off from what the pattern predicts.
Not cross-checking with answer options. Even if you identify the pattern, verify by plugging the answer back in. For multiply-and-adjust patterns, this takes 5 seconds and saves you from sign errors (forgetting whether it's +1 or -1, ×2 or ×3).