Thinking flexibly is key to cryptic crossword solving

When it comes to thinking about cryptic crossword solvers, what kind of image springs to mind? Maybe Chief Inspector Morse, a vicar, or a bowler-hatted Civil Servant? But would you be right…? Kathryn Friedlander shares new research lifting the lid on the mind of the cryptic solver – and finds that they are an academically able group, tending towards science, with fantastically flexible problem-solving abilities.

We’ve known for some time, since our research in 2016, that cryptic crossword solvers appear to be a bunch of high academic achievers, and that – perhaps paradoxically – their particular strengths seem to lie in the areas of science and maths, rather than in word-based areas.

For example, in the large survey of 805 solvers I conducted with Philip Fine, we found that over 80% of them (regardless of expertise level) had been to university; and 12% had PhDs. We also found that over half (56%) worked in STEM fields, including medicine or finance, and this rose to 66% among our Super-experts (they’re the kind of people who set crosswords for the newspapers, who solve advanced – barred – cryptic crosswords or who speed-solve in competitions).

Of course, we’re not claiming that you need to have gone to university to do cryptics. In fact, most of our participants had already started to solve in their mid-teens. But our research did suggest that there is a minimum threshold of flexible problem-solving ability for tackling cryptic crosswords, which is being reflected indirectly in the very high levels of university participation.

Flexible thinking in crossword solving – our study

So when it came to testing 28 crossword solvers in our labs (18 Super-experts and 10 Ordinary solvers), we had a hunch that they might do quite well at the sort of test which probes flexible problem-solving skills, known as ‘fluid intelligence‘. After all, the cryptic crossword puzzle is in itself a kind of ‘code-cracking’ exercise, requiring solvers to suppress the surface meaning of the clue in order to detect the concealed mechanisms. The algebraic/ cryptographic nature of the cryptic clue means that wordplay components may then be flexibly recombined or anagrammed to form new units. Sometimes even the definition or the solution pathway isn’t clear, meaning that these puzzles act as highly challenging ‘insight puzzles‘, leading to that magical Aha! moment.

Our face-to-face trials at Buckingham

What we did in our trials at Buckingham was to set our participants a particularly tricky fluid intelligence test (the AH5), deliberately devised for those who pursued education beyond high-school. The 72 test questions require people to handle complex challenges involving the separation, recombining and sequencing of information, and the detection of the rules operating in each case. In other words – all the kind of tasks you’re faced with when cracking a cryptic crossword!

Even the highest performing groups in the test manual only solved an average of 45 of these 72 questions correctly in the 40m time limit – giving an indication of just how tricky this test is.

We also got our participants to solve a full cryptic crossword set by Phi (Paul Henderson), which later appeared as Independent crossword #7835. We expected our Super-experts to romp away at this, but for our Ordinary solvers to find it a challenge in the 45m time limit. And that’s just what we found: for example, one Super-expert solver raced home in 10m47s, with the other Super-expert finishers taking an average of 23m30s. But the Ordinary solvers didn’t finish, on the whole, and had between 1 and 13 clues remaining at the end of the time. Take a look at the bar chart in the side panel, and compare the huge difference in performance.

Cryptic crossword solvers excel at fluid intelligence

When it came to the results of the fluid intelligence test, we found that our 28 cryptic crossword solvers did exceptionally well, scoring 44 out of 72 as an overall group. This is almost as good as the very best comparison group in the manual (Oxford Science Scholarship students: 45 out of 72), despite our group being considerably older (the average age of our solvers was 53yrs, compared to the 20yr old students). Fluid intelligence tends to decline with age on a pretty stable trajectory, so this was a very impressive performance.

What’s more, the Super-experts did particularly well at the test, scoring 47 out of 72, compared to the Ordinary solvers at 39 out of 72. This makes the Super-experts the highest scoring cohort we are aware of, doing better than any of the groups in the manual itself. The Ordinary solvers still compared well with other high-ability student populations such as Oxford Zoology students (39 out of 72), Medical students (38 out of 72) and Cambridge Arts students (37 out of 72). You can see the Super-expert scores plotted in the graph below (orange), compared with the Ordinary solvers (blue), the Oxford Science Students (green) and a group of mature students from the test manual (grey).

Success on the crossword and Fluid Intelligence scores were related

Finally, we plotted the results of the fluid intelligence test with the time taken to solve the crossword, and found an inverse relationship. That’s to say, as the scores on the fluid intelligence test went up, the time taken to solve the cryptic crossword went down. To do this, we had to extrapolate the finishing times for those who still had clues remaining, by assuming that they would have continued to solve clues as a steady rate beyond the 45m time limit.

What does this all mean? Implications for expertise research

Well, the main message is that, as suggested by our survey, cryptic crossword solvers do appear to be a pretty bright bunch, with particular strengths in fluid intelligence – the kind of thinking that allows you to perform mental gymnastics, and to solve novel problems on the hoof. Both groups did very well at the test, beating most of the high-ability groups in the test manual; but the Super-experts did exceptionally well.This suggests that flexible thinking lies at the heart of successful cryptic crossword solving.

From a wider perspective, it also feeds into our understanding of how people become experts at anything. There’s been a popularly-held belief that all it takes is focused practice (10,000 hours of it, to be precise!) and that anyone could succeed in a field if they persevere for long enough, but that position is becoming more and more difficult to defend as our knowledge increases. Here, our results suggest that would-be solvers may struggle to get a foothold in the activity without a certain threshold level of flexible thinking. And it’s very difficult to persevere with a hobby like cryptic crosswords if you feel that you just don’t ‘get’ them.

Additionally, our two crossword groups had both been solving for over 3 decades on average, accumulating many thousands of hours’ practice along the way, but there was a large difference in outcome, with some becoming highly expert solvers, and others remaining at more everyday levels of expertise – as we demonstrated by the Phi puzzle challenge. Differences in flexible thinking may be part of this picture, although we would need to conduct a larger-scale trial to be certain of this.

Our full paper has just been published in a peer-reviewed journal. If you are interested in learning more details about expertise, our trials and the results we found, it’s free to access by clicking on the link above. Do leave a comment below if you have any queries about the study, and we’ll be happy to respond.

This study was published as Friedlander, K. J., & Fine, P. A. (2020). Fluid Intelligence is Key to Successful Cryptic Crossword Solving. Journal of Expertise, 3(2), 101–132.


Man in thought:;

Independent Crossword #7835 – author’s own publication proof.

Graphs – from Journal of Expertise paper

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s