I find I do very badly at word based odd-one-out-type puzzles. It generally seems to be the case (at least, in my experience) that almost any of the items can be chosen as the odd one out, and this choice justified in some way.

A word could be the odd one out because of its structure (number of syllables, where the emphasis falls, etc.), the type of word (adjective, verb, etc.), the relationship of the word to humans (what kind of thing it refers to), the relationship of the word to the world in general (what aspect of the world, field of science, etc. it refers to), how the word is spelt or pronounced (rhymes, starting letters, syllables in common, etc.), and various other things.

There seem to be so many different aspects of a given word that in a given list of four or five things (which is what these puzzles - or at least, the ones that I've encountered - usually offer) there are likely to be multiple candidates for odd-one-out.

Often, when I see the answer to one of these puzzles, it's either something I considered, but discarded in favour of something else, or something I never considered because I was busy trying to find which of the possibilities I'd already thought of it was. It's almost never something which is (to me) more obviously fitting than the possibilities I'd considered.

Now, if everybody else had the same experience, I'd be inclined to assume these questions are just a bit imprecise by nature. Evidently, though, since some people are very good at them, there must be a way to determine which answer is the intended one, and this is very likely to be a transferable skill (in the sense that it can be transferred to new puzzles).

Even if it isn't a learnable skill, I'd still be interested to know what it is. The answer may be that I'm just naturally not good at this sort of puzzle. That's fine. But why? What prior knowledge or mental skill is required to determine which of the possible answers is the one the asker wants?

  • $\begingroup$ Question "why" is not for puzzling, it's about cognitive sciences, I'd say. $\endgroup$
    – rus9384
    Commented Oct 12, 2017 at 8:33
  • $\begingroup$ Generally, I tend to look for the most obvious 'strong' link between all but one of them, and from there onwards point out which one doesn't fit. On virtually any question you can come up with a reason for any of the answers to be the one one out. If your reasoning seems farfetched, it usually is. $\endgroup$ Commented Oct 16, 2017 at 12:38

4 Answers 4


I can sympathize. This is from The Guardian, 2015:

odd ones out

From left to right, starting with image 2, we have:

  • no outline
  • not a square
  • green
  • small

leaving image 1 because it is the only image that is not an odd one out!

Which I think qualifies it as an odd one out.

So the answer given might not be unique, which makes some of the puzzles unsolvable.

  • 4
    $\begingroup$ That's horrifically meta. $\endgroup$ Commented Oct 13, 2017 at 22:02

The Situation that you are talking about is a lot more common than you know. And it's not only limited to odd-one-out puzzles. There are riddles which are basically "guess what I am thinking" puzzles. Odd-one-out puzzles usually land in this basket.

There is nothing wrong with your thinking or your reasoning. The fault, almost always, lies with the question. And that is primarily because odd-one-out puzzles are by default too-broad. It's really hard to come up with a set of words/items that differ in one and only one characteristic.

As for your second question ( What prior knowledge or mental skill is required to determine which of the possible answers is the one the asker wants?), the answer is none. You just have to be lucky enough to be thinking exactly what the asker was thinking. And there is no "trick" to that. Either you have to have learnt the art of mind-reading, or you just have to be lucky. Obviously, mind-reading is not something possible, the answer is, "You just have to be lucky".

  • $\begingroup$ Thank you for your answer, Sid. Do you think I'm mistaken in believing that some people are consistently good at these sorts of puzzles, then? I'll admit I don't know this for a fact, but if it is true then it suggests that there is, if nothing else, a difference between how they approach these problems and how I (and presumably you, based on your answer) do. $\endgroup$ Commented Oct 11, 2017 at 14:23
  • 2
    $\begingroup$ @TheTermiteSociety If you have data to show that people are consistently good at these sort of puzzles, then, please do show them to me. I(and many others here at Puzzling) firmly believe that these "puzzles" are simply based on luck. $\endgroup$
    – Sid
    Commented Oct 11, 2017 at 14:25
  • 2
    $\begingroup$ @TheTermiteSociety I can imagine that once you see the kind of puzzles a certain author has, one may learn to think the same way. So, the pattern you are picking up is not "the odd one out", but rather "which kind of oddities the author thinks about". $\endgroup$
    – Davidmh
    Commented Oct 11, 2017 at 19:52

My hint would be to rephrase the common question from "Which is the odd one out" to "Which item is not in the set to which all others clearly belong ?". This helps to reframe thinking and leads to better, easier answers. The common task becomes "find a set in which all but one of the items is actually in" rather than picking on a single excluding characteristic of one item.

The Guardian example of "JonMark Perry" above, as he shows, has multiple correct answers. Personally I would have gone with "no Outline" from the choices.

To improve odd-one-out thinking look at the question backwards and find a simple all encompassing set that excludes a single candidate answer.


You will almost always find multiple grouping rules: each of them will result in a different solution; this is also true for other kind of puzzles.

From my experience, most of these rules will be overcomplicated: when you guess the intended solution (or when someone else guesses it) you will be absolutely sure that that is the correct one because it is simple but still it holds for exactly $n-1$ items leaving an odd-one-out.

If you cannot find such simple rule, then according to me the puzzle is not well-formed.

The example proposed by JMP is a meta-puzzle thus it can be considered as an exception.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.