Is There a Faster Way to Spot the Odd One Out in CAT?

VARC

Odd one out is the question type candidates most often solve by feeling. Five sentences, four belong together, one does not, and the usual method is to read them all and notice which one seems off. When that works it is fast. When it fails, there is no fallback, because a feeling cannot be checked.

There is a faster and more reliable method, and it does not involve reading more carefully. It involves changing what you are looking for: not the sentence that feels out of place, but the sentence that cannot be linked into the chain the others form. That is a structural test with a definite answer, and this piece gives you the four links to check, the order to check them in, and what to do when two candidates survive.

You Are Looking for the Wrong Thing

The instinctive approach asks which sentence does not belong. That framing sends you hunting for difference, and difference is everywhere in five sentences written by the same author.

The reliable question is the opposite one: which four sentences form a connected sequence? Once four of them link into a chain, the fifth is the answer whether or not it felt odd, and you have a reason rather than an impression.

This matters because the trap in this question type is precisely a sentence that feels related. Setters build option sets where the intruder shares the topic, the tone and the vocabulary of the other four, so anyone reading for feel is being tested on the thing the question is designed to defeat.

Four Links, in Order of Decisiveness

Start With Reference Chains

If you check one thing, check this, because it is both the fastest and the hardest to argue with.

A pronoun or a demonstrative needs something to point at, and that something must have been introduced earlier in the sequence. So a sentence containing one is necessarily linked to a specific other sentence, and finding two or three such links builds most of the chain in seconds.

The intruder frequently has no reference into the others and nothing referring to it. That absence is a definite fact about the sentence rather than an impression, which is what makes this method checkable under pressure.

Build Pairs Rather Than the Whole Sequence

The method error that slows candidates down is trying to assemble the full chain in one pass, holding several arrangements in mind at once.

Work in pairs. Find two sentences where a reference, an article shift or a connective forces one to follow the other, and lock that pair. Then find another. Two locked pairs usually leave only one sentence unattached, and that sentence is the answer.

Pairs are cheaper to establish and far more reliable than whole-sequence impressions, and they give you something that stays put while you work on the rest.

The Standard Trap, Named

It is worth seeing the trap explicitly, because recognising it is most of the defence.

A well-built option set contains one sentence that is thematically perfect and structurally isolated, and one that looks slightly tangential but is load-bearing in the chain. A candidate reading for feel picks the tangential one and leaves the isolated one in place.

That is why the structural test beats the thematic one every time it is available. The setter can make a sentence feel like it belongs. They cannot give it a reference chain into sentences it does not connect to without making it actually belong.

Signal	Reliability	What it tells you
Pronoun or demonstrative reference	Highest	A forced link to a specific sentence
Indefinite to definite article shift	High	Direction between two sentences
Connective word	High	What kind of sentence must precede
Continuity of the specific subject	Moderate	Which object the chain develops
General theme or topic	Lowest	Usually the trap, not the answer

When Two Candidates Survive

Sometimes the links are genuinely thin and two sentences both look detachable. This happens and it is worth having a procedure rather than a coin flip.

First, test whether removing each candidate leaves a sequence that still reads as continuous. The correct intruder leaves four sentences that connect; the wrong one leaves a gap where its links were. That test is quick and it resolves most ties.

Second, check which of the two is doing argumentative work. A sentence that introduces a term the others use, or that states the claim the others support, is load-bearing even if it reads as general. The intruder is usually the one that could be deleted without anything else changing.

Let the Format Decide the Ending

If the tie does not resolve, the scoring decides what to do, and this is where candidates most often get it wrong.

Where the question is Type In The Answer, there is no penalty for an incorrect response: a correct answer is plus three and a wrong one is zero rather than negative. So your better candidate should always be entered, because a blank guarantees zero while a considered guess costs nothing.

Where it is a multiple choice question, an incorrect answer is minus one. A coin flip between two survivors has close to no expected value, and the attempt is worth making only once you have eliminated on structural grounds rather than on preference.

Cap the Time

This question type has a specific failure pattern: it feels nearly solved for a long time, which makes it unusually good at absorbing minutes.

Set a short ceiling before the exam. If a scan for references and one pass at building pairs has not produced a clear intruder, the links are thin for you on this question and the remaining time is better spent elsewhere in a 40 minute section. The ceiling matters more here than on most types precisely because the sense of being close is so persistent.

How to Practise So the Links Become Visible

This is a noticing skill, built by a specific review habit rather than by volume alone.

When you get one wrong, do not just read the correct answer and accept it. Find the specific link you failed to use: the pronoun, the article, the connective. Write it down. After twenty questions reviewed that way you will have a short, repeating list of the evidence types you habitually miss, and that list is worth more than a hundred questions attempted and moved past.

Do the same when you get one right. If you got it right on feel rather than on a link, that was luck, and luck does not transfer to a harder set under a running clock.

The Same Reading Works on the Neighbours

The habit built here pays across the verbal ability portion rather than in one question type, which changes how much the practice is worth.

Para jumbles are the same evidence used to order rather than to exclude. Para completion asks which sentence finishes a paragraph, and the correct closer is the one that resolves what was opened and matches the reference chain already running, rather than the one that merely sounds like a good final line.

Since reading comprehension is consistently the majority of VARC, the verbal ability types have to earn their share of a smaller practice budget. A skill that covers three of them at once is where that budget goes furthest.

The Summary

Stop asking which sentence does not belong and start asking which four form a connected chain. The first question sends you hunting for difference, which is everywhere. The second has a definite answer you can check.

Check reference chains first, because a pronoun needs an antecedent and that forces a link to a specific sentence. Then article shifts, then connectives, then continuity of the specific subject. Theme is the least reliable signal and it is usually the trap, since setters build intruders that share the topic and the tone of the four that belong.

Work in locked pairs rather than assembling the whole sequence at once. When two candidates survive, test which removal leaves a continuous sequence and which sentence is doing argumentative work. Then let the format decide: always enter your better candidate on a Type In The Answer question, and on an MCQ attempt only once you have eliminated on structure. And cap the time, because this type feels close for far longer than it deserves.

If these are being decided by feel, you are being tested on exactly the signal the question is built to exploit. A CAT preparation strategy review will show which link types you keep missing, and a personalised CAT preparation plan puts the reviewed practice into your week.