Skip to content

Six Reasoning Mistakes That Show Up in Puzzles

Valid but untrue, one rule assumed unique, a conditional read backwards. Six errors that wreck puzzle answers, each with a checkable example and a fix.

By Published Updated 6 min read

Reign Creative LLC makes IQ Test & Brain Training Games, which this article links to, so we have a commercial interest in it. How we write and check articles.

Key takeaways

  • Validity is about the shape of an argument, not the truth of its conclusion. A perfectly valid argument delivers a false conclusion whenever one of its premises is false.
  • Three terms never determine a fourth. The run 2, 4, 8 continues as 16 under a doubling rule and as 14 under a growing-difference rule, and neither reading is wrong on the evidence given.
  • The Stanford Encyclopedia of Philosophy's working definition of a fallacy - an argument that seems to be better than it really is - is the most useful sentence in the subject, because every error below feels correct from the inside.
  • Reversing a conditional is the most common formal error in everyday reasoning. If a number is divisible by 6 it is divisible by 3, but 9 shows the reverse fails.
  • In a puzzle with an elimination structure, an answer consistent with the clues you have read is not a solution. It is a solution only when it is consistent with all of them and nothing else is.

Six errors account for most wrong answers on reasoning puzzles, and all six share a property that makes them dangerous: they feel correct while you are making them. The Stanford Encyclopedia of Philosophy's entry on fallacies offers a working definition worth memorising — a fallacy is "an argument that seems to be better than it really is." Each section below names one error, gives an example you can check yourself, and states the specific habit that catches it.

One: treating a valid argument as a true one

Validity is a property of shape. It says that if the premises are true, the conclusion must be. It says nothing at all about whether the premises are true.

Try this one:

All prime numbers are odd. 2 is a prime number. Therefore 2 is odd.

The form is impeccable. The conclusion is false. The first premise is false, and validity does not protect you from that. Run the same structure with true premises — all multiples of 4 are even, 12 is a multiple of 4, therefore 12 is even — and you get a true conclusion, because the inputs were sound rather than because the shape improved.

The reverse trap is just as common: a true conclusion does not certify the argument that produced it. You can reach a correct answer through a broken chain, which is why a puzzle you got right by accident teaches you nothing.

The habit: check the premises separately from the inference. They are two different jobs and doing them together is how one gets skipped.

Two: over-generalising from a short run

Three terms never determine a fourth. This is not a pedantic point, it is the central fact about sequence puzzles.

Consider 2, 4, 8. Doubling gives 16. But the differences so far are 2 and 4, and a rule where each difference is two larger than the last gives a next difference of 6 and therefore 14. Both rules fit every term you were shown. On the evidence presented, neither answer is wrong.

A more striking version comes from geometry. Mark points around a circle, join every pair with a straight chord, arrange the points so that no three chords cross at the same interior spot, and count the regions the disc is cut into. For one point through five you get 1, 2, 4, 8, 16 — and for six points the count is 31, not 32. You can verify it with a pencil, which is the honest way to meet a counterexample.

The habit: before committing, name a second rule that also fits. If you cannot find one, your confidence is earned. If you can, you have learned that the puzzle depends on convention.

Three: assuming the rule must be unique

The previous error is about not looking. This one is about believing there is nothing to look for.

Puzzle sequences are answerable because a generator fixed a rule in advance and the convention is that the intended answer is the simplest rule consistent with the terms shown. That convention is doing real work, and it is a convention rather than a mathematical fact. Forgetting this produces two failures: arguing that your answer "must" be right because it fits, and rejecting a published answer because yours also fits.

Both are avoidable by asking the right question — not "does my rule work?" but "is my rule the simplest one that works?" Our guide to number sequence puzzles sets out the standard families and a procedure for working through them in order of simplicity, which is the practical form of this discipline.

The habit: rank your candidate rules by how much machinery they need, and prefer the plainest one that survives every term.

Four: matching the words instead of the relationship

Analogy items punish surface association. Take:

Thermometer is to temperature as odometer is to ______

The relationship is instrument to the quantity it measures, so the answer is distance. The tempting wrong answer is "car", because odometers and cars go together in the way that feels like an association. Strength of association is not the relation being tested.

The same failure has a documented cousin in conditional reasoning. In the Wason selection task, people select the cards named in the rule rather than the ones that could falsify it; the PLOS ONE study by Rossi and colleagues reports that 78% of their sample "relied on total or partial matching". Matching what is salient is a fast, cheap heuristic that produces confident wrong answers.

The habit: say the relationship out loud as a full sentence before you look at any options. "A thermometer measures temperature" makes the correct completion nearly automatic.

Five: reversing a conditional

From "if P then Q" plus Q, nothing follows about P. This is the formal error the Stanford Encyclopedia lists as affirming the consequent, and its partner — denying the antecedent — fails for the same reason.

In numbers it is unmistakable. If a whole number is divisible by 6, then it is divisible by 3. True. Reverse it: if a whole number is divisible by 3, then it is divisible by 6. False — 9 is divisible by 3 and not by 6.

There is exactly one safe reversal, the contrapositive: flip the two parts and negate both. If a whole number is not divisible by 3, then it is not divisible by 6. That is true, and it is true whenever the original is.

The habit: when you want to run an if-then backwards, write the contrapositive instead. If the contrapositive does not give you what you need, the inference you wanted was not available.

Six: stopping at the first consistent answer

Consistency with the clues you have read is not a solution. It is a candidate.

This is the error that ruins elimination puzzles. In a grid puzzle, an assignment that satisfies the first three clues can be flatly contradicted by the fifth, and the only defence is to use every constraint before declaring victory. Our worked logic grid walkthrough shows the disciplined version: record all the negatives, propagate every positive across its row and column, and count what remains.

The same failure appears in arithmetic as an answer that is plausible but never checked against magnitude, which is one reason estimation is worth treating as a checking tool rather than as a shortcut.

The habit: after you have an answer, spend one pass confirming that no remaining information contradicts it. It costs seconds and it catches this whole category.

Catching your own errors in practice

NRICH, the mathematics education project at the University of Cambridge, describes a progression in reasoning that ends with justifying — "a correct logical argument that has a complete chain of reasoning to it" — and proving, "a watertight argument that is mathematically sound". The useful implication is that reasoning quality is visible in the explanation, not in the answer. If you cannot say why, you do not yet know.

That has a direct consequence for how you practise. Timed formats reward the first plausible answer, which is exactly the condition five of these six errors need. IQ Test & Brain Training Games, which we build at Reign Creative, ships an untimed accuracy mode that gives feedback, an explanation and an optional hint on each question, and that format — not the sprints — is the one to use when you are trying to catch your own reasoning rather than your speed. Those drills record score, accuracy, streak and response time. The app's IQ-style test, memory test and personality questionnaire are separate, and its listing describes their results as estimates based on app activities, not clinical or professional assessments, so treat none of them as a verdict about you.

You can see how the app describes its games and tests on the IQ Test & Brain Training Games app page, find the rest of this material in the logic and pattern puzzles hub, and browse the wider education and brain section for related writing on study method and evidence.

Good to know

Frequently asked questions

Can a valid argument have a false conclusion?
Yes, and this is the single most useful thing to know about validity. Validity only guarantees that the conclusion follows if the premises are true. Take: all prime numbers are odd; 2 is prime; therefore 2 is odd. The form is impeccable and the conclusion is false, because the first premise is false. Checking form is never enough.
Why can a number sequence have more than one right answer?
Because finitely many terms never pin down a rule. Given 2, 4, 8, doubling gives 16 and a rule where the gaps grow by two each time gives 14. Both are consistent with everything you were shown. Puzzles are answerable only by convention: the intended answer is the simplest rule that fits, not the only possible one.
What is affirming the consequent?
Treating an if-then statement as though it works in both directions. From "if P then Q" and Q, nothing follows about P. In numbers: if a whole number is divisible by 6 it is divisible by 3, but 9 is divisible by 3 and not by 6. The only safe reversal is the contrapositive, which flips both parts and negates them.
How do I check my own reasoning on a puzzle?
State the rule you are using in a full sentence, then look for the case it would get wrong. Most errors survive because they are never articulated. Writing out "I am assuming the difference is constant" makes it obvious that you should test whether the ratio is constant too, and takes about three seconds.
Do timed drills make these errors more likely?
Speed pressure rewards the first plausible answer, which is precisely the condition several of these errors need. If you want to catch your own reasoning, an untimed format with feedback and an explanation is the better tool. IQ Test & Brain Training Games ships an untimed accuracy mode with hints and explanations for exactly that reason.

Sources

  1. Fallacies (opens in a new tab)Stanford Encyclopedia of Philosophy, Stanford University — accessed
  2. Helping Reasoners Succeed in the Wason Selection Task: When Executive Learning Discourages Heuristic Response but Does Not Necessarily Encourage Logic (opens in a new tab)PLOS ONE — accessed
  3. Reasoning: The Journey From Novice to Expert (opens in a new tab)NRICH, Millennium Mathematics Project, University of Cambridge — accessed

About the author

— The independent app studio that builds the apps covered on this site and publishes every article here. On Google Play it publishes as Reign Collective Apps.

We use AI tools to help research, draft and edit our articles. Each article lists the sources its factual claims rely on, and details about our apps come from their live Google Play listings. No article on this site has been reviewed by an independent expert unless the reviewer is named on it.

Spotted an error? Tell us — how corrections are handled is set out in our editorial policy.

Keep reading