IDBI Assistant Manager common mistakes
IDBI Assistant Manager · Understand the traps and how to avoid them.
In puzzle and seating-arrangement sets, the candidate draws one arrangement, finds a contradiction partway through, and then erases and restarts from scratch two or three times instead of checking the fixed clues first.
Under a sectional time limit the urge to 'get going' beats the discipline of isolating the anchor clues. Restarting feels productive but burns 4-6 minutes on a set that should take 90 seconds, and the panic then spills into the next set.
What to do instead
Before drawing anything, spend 20 seconds classifying clues as fixed (absolute positions, definite left/right) versus variable (either/or, 'not adjacent'). Build the skeleton from fixed clues only, then test variable clues against it. If a contradiction appears, identify which single clue caused it and branch only that clue rather than redrawing the whole grid.
On syllogism and inequality questions the candidate answers from the meaning of the sentences rather than from the Venn diagram or the chain of symbols actually drawn.
Everyday language makes 'some A are B' feel like it implies 'some B are not A', and inequality chains like A>B, B≥C, D<C invite the reader to assume D<A without checking. The candidate trusts intuition and skips the mechanical step.
What to do instead
For syllogisms, draw the minimum-overlap diagram for the premises and then ask whether a counter-diagram exists that makes the premises true and the conclusion false; only a conclusion with no counter-diagram follows. For inequalities, rewrite the whole chain in one direction (all '>' pointing the same way) before judging any option, and treat 'either/or' conclusions as valid only when the two options are individually undecidable but jointly exhaustive.
In data interpretation the candidate reads the chart, then performs every calculation at full precision with long division, spending six or seven minutes on one five-question set.
School habit trains exact arithmetic, and DI sets are designed so that exact computation is unnecessary. The candidate also fails to notice that many questions ask for a ratio or percentage change where approximation to two significant figures gives the same option.
What to do instead
Convert every DI question into an estimate before computing: round populations and amounts to two significant figures, compute the percentage change approximately, and only do exact arithmetic if two options fall within the error band. Practise the standard shortcuts (percentage change ≈ difference divided by base, ratio comparison by cross-multiplication with rounded numbers) until they are automatic.
The candidate attempts all 35-40 quantitative questions in order, gets stuck on a time-consuming arithmetic word problem, and leaves 8-10 easy simplification or approximation questions unattempted.
Sequential answering is the default exam behaviour, and the paper mixes difficulty deliberately. Without a pre-planned triage rule, the candidate has no reason to skip and keeps sunk-cost investing in a single question.
What to do instead
Adopt a two-pass rule: on the first pass answer only questions solvable in under 45 seconds, marking everything else. Return to marked questions only after the easy ones are banked. In mocks, record how many questions you left blank versus got wrong, since unattempted easy questions are the cheapest marks lost.
In reading comprehension the candidate reads the passage once at normal speed, then answers each question by re-reading the entire passage to locate the evidence.
The candidate treats RC as a memory test rather than a location test, so each of five questions triggers a full re-read. Time per set balloons and accuracy drops because the final questions are answered under pressure.
What to do instead
Read the passage once for structure only, noting in the margin where the author's claim, the counter-claim, and any example sit. Then answer questions by returning to the relevant paragraph, not the whole passage. For 'according to the passage' questions, verify the exact wording of the option against the sentence rather than accepting a paraphrase that sounds right.
On error-spotting and sentence-correction questions the candidate picks the option that 'sounds wrong' without identifying the grammatical rule being tested.
Spoken English tolerates constructions that written English rejects, so ear-based judgement misfires on subject-verb agreement across intervening phrases, parallelism in lists, and misplaced modifiers. The candidate also misses errors of omission, where a preposition or article is simply absent.
What to do instead
Check each sentence against a fixed checklist in order: subject-verb agreement, tense consistency, pronoun reference, parallelism, articles and prepositions, modifier placement. Read the sentence with the suspect part removed to see whether the remainder still parses. For 'no error' options, only choose them after the full checklist has been run.
The candidate memorises static facts (capital cities, founding years) but cannot answer questions on current policy rates, recent scheme launches, or the functions of regulators.
Static material is easy to revise from a book, whereas current affairs requires regular updating. The candidate also assumes banking awareness means history of banking rather than the operating framework of RBI, SEBI, IRDAI and the schemes currently running.
What to do instead
Maintain a one-page living sheet for the last six months covering policy repo rate and stance, CRR and SLR, inflation prints, major schemes and their eligibility, and recent appointments to regulator and bank leadership. Revise it weekly rather than in the final days, because awareness questions reward recency and precision, not volume.
The candidate answers banking-awareness questions by recalling a headline without knowing which institution or instrument it refers to, confusing, for example, monetary-policy tools with fiscal instruments.
News reading is passive; the candidate registers that 'RBI did something with liquidity' without encoding the mechanism. Under exam pressure, superficially similar terms (repo versus reverse repo, CRR versus SLR, bank rate versus MSF) collapse into each other.
What to do instead
For each concept, write a one-line definition in your own words plus the direction of its effect: who sets it, what it does to liquidity or credit, and which side of the balance sheet it touches. Build paired flashcards for the confusable pairs and test yourself by explaining the difference aloud, not by recognising the term.
In accounting questions the candidate treats the balance sheet as a list of items to memorise rather than applying the accounting equation, so any adjustment question (depreciation, outstanding expenses, prepaid items) produces a guess.
Professional-knowledge sections are often prepared by reading definitions rather than working through journal entries. Without the debit-credit framework internalised, the candidate cannot predict which side an adjustment moves.
What to do instead
Drill the rule that every transaction has two equal effects and that the balance sheet must balance after each adjustment. For each adjustment type, work three or four numerical examples by hand: record the journal entry, post it, and confirm total assets still equal liabilities plus capital. Learn the classification of accounts (personal, real, nominal) and the golden rules as a decision procedure, not a slogan.
The candidate attempts the professional-knowledge section with the same time-per-question as reasoning, running out of time on the final 10 questions in every mock.
Sectional timing is rarely rehearsed explicitly; candidates practise content, not pacing. Professional knowledge and awareness questions are answered fast when known and should be abandoned fast when not, but the candidate lingers, hoping recall will arrive.
What to do instead
Set a hard per-question ceiling for the awareness and professional sections (roughly 30-40 seconds) and skip immediately when the answer is not present within that window. In mocks, log the time spent on questions you ultimately got wrong; if any single question exceeded 60 seconds, that is the pacing leak to fix before the next attempt.