Thursday, October 30, 2008

Does Length Of Clue Affect Its Complexity?

Sponsored Links

I looked at the New Indian Express (NIE) crossword after a gap of a few months. My usual mode of solving the NIE puzzle is without grid; I go to the Orkut Crossword community, glance through the clues on the clue-sheet and give the answers that I can.

This is pretty easy to do with the NIE, it is not a taxing puzzle (though an enjoyable one). This time, I also noticed that the clues are short, much shorter than those of The Hindu. I recalled thinking that Neyartha’s puzzles in The Hindu were quite wordy, and wondered if clue length influences the difficulty of the clue in some way. After all, the higher the number of words in a clue, the more the permutations by which the solution can be derived.

Compare these two anagram clues:
NIE Oct 30 2008: Saves an agitated girl (7)
The Hindu 9362: Beat around the bush when asked to quote advice by mistake (10)

Both anagrams, but the second is more difficult to identify than the first. Among other reasons for this, one surely is that the first has single-word definition and anagram indicator, the second has multi-word ones.

A regexp check shows average clue length (number of words per clue) of the NIE as below:

Puzzle Date Average Clue Length
30-Oct-08   6.29
29-Oct-08 5.44
15-Jul-08* 5.64

*The last puzzle attempted before the hiatus.

The average in NIE based on the above data is 5.79 words per clue.

A similar analysis of the Hindu crosswords published between 26-Sep-08 to 26-Oct-08 threw up some interesting results.

The average in The Hindu is 7.50 words per clue.

There were 4 compilers at work in this duration - Nitaa Jaggi, Gridman, M.Manna and Neyartha. Of these, Gridman and M.Manna tend to write shorter clues, Nitaa Jaggi and Neyartha longer ones. The individual averages are:

Puzzle Nos. Compiler    No. Of Puzzles Average Clue Length
9338 - 9347    Nitaa Jaggi 10 7.86
9348 - 9353    Gridman 6 7.11
9354  -9360    M. Manna 7 7.20
9361 - 9362 Neyartha    2 7.90*

*of which 17% of the clues excluded definition

If assigning an absolute complexity rating, Nitaa Jaggi's puzzles would be termed Easy, Gridman and M.Manna Medium and Neyartha's Complex. So, there does not seem to be a direct correlation between clue length and complexity.

Another thought is that very long clues either obfuscate or overexpose, so sway towards the extremes - Complex or Easy, but seldom middling.

Anyway, here's a graph showing average clue length per crossword for a whole month, for you to mull over and draw your own conclusions!  

Average Clue Length

Related Post:
Thoughts On Crossword Complexity

If you wish to keep track of further articles on Crossword Unclued, you can subscribe to it in a reader via RSS Feed. You can also subscribe by email and have articles delivered to your inbox.


Chaturvasi said...

I know NIE puzzles from the Seventies. Clues of Roger Squires - that's the name of the compiler of this UK syndicated crossword - have always been short and precise. They may be easy but they have high entertainment value.
THC has only now started giving bylines. Your observations made by scrutinising only puzzles over a month may by and large be correct - but I would urge you to do this exercise for at least three months in succession so we may have reliable results.

Shuchi said...

I enjoy the NIE puzzles very much too, appreciate the economy of words and the 'punch' factor that its clues have.

For The Hindu, you're right we need more data to go upon! Will try to validate the initial observations again after a few months, as you suggest.

Chaturvasi said...

I may add that Roger Squires' puzzles appear in The Guardian under the pseudonym RUFUS and in The Herald under RFS and EAP (a pseudonym that the then Editor at a pinch bestowed on him and which is supposed to stand for Edgar Allan Poe).

Shuchi said...

Oh I didn't know that Rufus of The Guardian was the same as the NIE compiler! Aren't the clues there more layered than in the NIE? The Guardian Rufus puzzles are tougher for me than the NIE. Do you know if compilers clue differently for different publications?

Chaturvasi said...

Of course, they do.

If you take The Times/The Financial Times, The Guardian, The Independent and The Daily Telegraph, there are some compilers who supply puzzles to two or three of these publications. One of them may also be supplyiing crosswords to a church magazine. These compilers set crosswords keeping in mind the specifications laid down by their editors.
The difficulty levels too might vary.
In any case, compilers always keep a 13x crossword (which NIE's is) less easy than a 15x. The two 15x's that NIE carries in a fortnight are also on the easy side.

Chaturvasi said...

In the above "In any case, compilers always keep a 13x crossword (which NIE's is) less easy than a 15x" should have been "...less hard..."