My thoughts on conundrums, and setting them. Some of you may have heard these from me in person, perhaps several times, but this seems a suitable thread in which to write it all down so I can stop going on about it at events.
tl;dr: Don't choose obscure words for conundrums as a lazy way of making them difficult!
When setting a pile of conundrums for a co-event, there is a temptation to pick obscure words for the "difficult" conundrums, the ones intended to be given out on crucials between high-rated players, or in the final. Choosing conundrums whose difficulty arises from the low probability of either player even knowing the word is seen as an easy way of manufacturing lists of difficult conundrums.
There are two fundamental problems with this approach. First, it misunderstands the point of puzzle setting. Second, it doesn't even work.
Puzzle setter and former Only Connect question editor Alan Connor describes the setter's role in
The Joy of Quiz: "Novice setters sometimes fancy that their job is to beat the contestant; it is not. Their job is to lose following a struggle."
Alan Connor was talking about quiz questions, but the same applies to conundrums or any other kind of puzzle. No conundrum setter should be patting themselves on the back because their super-hard final conundrum "defeated" the finalists.
Of course, it's not at all easy to judge whether the players will get it, and sometimes they won't. But if when you reveal the answer, it's a virtually unknown word, the players and most of the spectators will utter a collective "...?", with the sense that it was a bit of a wasted 30 seconds. On the other hand, if they recognise the word, the collectively-uttered punctuation will be more like "!", a much better ending to the tournament.
Consider the spectacle for the audience, as well, especially for the final conundrum. "I'd have got that" or "I could have got that" is always better than "I'd never have got that". Setters are - rightly - concerned about making it too easy in order to avoid the crapshoot of an instant buzzer race, but so often they tend to overcompensate the other way and it goes unsolved. Conundrum solves around the 15-25 second region are rarest, but when they happen, it's because the setter pulled off their job perfectly - to lose, after a struggle.
In the COLON 2018 final, the conundrum, WETHARDON, lasted over 20 seconds being stared at by the two finalists before one of them got it. And even if they hadn't got it, when
the answer was revealed there would have been no anticlimactic "...?" Great conundrum.
It's worth noting that for all the conundrums at COLON - not just the final - Jen specifically set out to make a range of conundrum difficulties without resorting to obscure words for difficulty. There were probably more conundrums written for that event (well over 200) than for any other co-event in history, so if you're thinking of hosting an event there's no reason you can't do that too.
Another approach, if you don't mind deviating from Countdown tradition, is to do what Milton Keynes did this year, and make the conundrums ten letters long. In the past, MK's grand final conundrums were known for being so hard that the answers didn't even look like words (
ACHOLURIC, anyone?), but with ten-letter conundrums it was much easier to find words that the finalists hadn't practised to death on apterous but which were still normal, everyday words.
Let's have a look at the second fundamental problem - the misconception that ramping up the obscurity provides more of a challenge to the higher-ranked players. It doesn't. The best players have played all the conundrums on apterous, and what makes a conundrum difficult to them doesn't have a great deal to do with how often the word is used. As an example, consider the following two conundrum scrambles.
ISAWKHAZI
INERTSTAR
The answer to the top one is an obscure word which most people outside of word games don't know. The answer to the bottom one is a very common word which you'll all know. However, I guarantee you that most expert players will spot the top one instantly (maybe you have already?), and take some time over the bottom one (still haven't got it?). The difficulty is nothing to do with the obscurity of the words, it's about the distribution of the letters.
For the top conundrum, a seasoned player will instantly spot the W, the Z, the K, the As and Is, and buzz in immediately with
the answer. Most of the top players have seen this word before, and it's so easy to spot that they'll get it straight away. Furthermore, in terms of catering for a range of skill levels, this one is the worst of all worlds; a top player will get it instantly with no effort, but a novice player has almost no chance of getting it no matter how much they try.
Now let's look at the bottom conundrum. It's got lots of common letters. It's also got lots of building blocks of words, which conundrum solvers often use as short-cuts, and which in this case are nearly all red herrings. There's -IEST, there's -ATE, there's INTER-, -ISE, -ISER, -ITE... none of them are correct. Even if you didn't get
the answer* in 30 seconds, the reaction upon seeing it is "ah!" rather than "what's that?" Even better, although this is a difficult conundrum, there's no reason a novice player couldn't solve it, and it certainly wouldn't be unfair for this conundrum to crop up in their game. They might even get it before a significantly higher-rated opponent, which would have been an impossibility with the other conundrum.
* This conundrum was used for a game at Co:Leam in 2013, and is what I usually use as an example of a "good" conundrum. While it was a valid conundrum at the time, nowadays it has an alternative, more obscure, solution.
Here's all of the above, distilled down into a graph. Imagine every conundrum is somewhere on this graph. The x-axis represents word obscurity, and the y-axis represents the difficulty of unscrambling the word assuming it's in your vocabulary. I'm not saying don't ever set obscure words (players shouldn't be entitled to assume the word won't be obscure), but I'd recommend using them only rarely. Most co-event conundrums ought to be on the left-hand half of the graph. To control conundrum difficulty, vary the position on the y-axis rather than the x-axis!
Code: Select all
Difficulty in unscrambling
^
| The "rewarded effort" corner, The "WTF is that" corner,
| e.g. RESTRAINT. e.g. ACHOLURIC,
| HANDWROTE MRIDANGAM
|
|
| The "buzzer race" corner, The "worst of all worlds" corner,
| e.g. EXPLODING e.g. WAKIZASHI
|
+----------------------------------------------------------> Obscurity of word
