Wanted: A "Dumb" Cryptogram-Maker (1 of 2)
Hunting patterns...and hunting a way to hunt for patterns.
In my quest to uncover wordplay and word trivia at scale, I’m happy to use whatever modern tools I can find. There are a couple of respects, though, in which those tools let me down.
I’ve already gone into the issues with IPA transcription tools. IPA Now has since given me good enough results that I’ve subscribed to it it—it retains an irritating tendency to drop its R’s, but it’s close enough to what I need to scan thousands of words and phrases for sound-alikes.
But I can’t find any good resource when it comes to scanning thousands of words and phrases for their letter patterns.
I’ve mentioned shared letter-patterns before, but here’s a refresher: one notable example is ALFALFA and ENTENTE, the only common English words that have the pattern ABCABCA. A while back, I proposed using that common pattern as the basis for a cryptic clue:
Coded
entente with Little Rascal (7)—ALFALFA
I think such patterns have a lot of potential in puzzles. Right now, though, most of that pattern-matching is done in scaled-up cryptogram puzzles like this sample quote from Arthur Conan Doyle: RB RW Y QVCYB BPRJQ BS WBYVB IRUC TRBP Y WZYII JNZKCV SU VCYIIE QSSG KSSHW TPRAP YVC ESNV OCVE STJ.
If we start at the beginning—RB RW Y—we note two two-letter words that share an opening letter, followed by a one-letter word that share none of the two-letter words’ letters. There are only so many common phrases in English that match that pattern, so this is likely to be code for IT IS A. (Had the sentence ended with a question mark, it could’ve opened with IS IT A.)
From this, we can conclude that the seventh word, STA_T, is probably START. And from there, each conclusion leads you to another until you have the answer: IT IS A GREAT THING TO START LIFE WITH A SMALL NUMBER OF BOOKS WHICH ARE YOUR VERY OWN.
I think shorter patterns have a lot of other potential uses, and they’re fun to play with. They’re relevant to my list of celebrities from the other day: EMMA STONE matches GOOD THING and TOOK HEART, and if we remove spaces, SEEDLINGS.
BARRY KEOGHAN (star of Saltburn) has a number of matches without spaces—with spaces, he matches ex-reality star HOLLY MADISON, though they aren’t “fully encrypted” since “y” and “n” in one name correspond to “y” and “n” in the other.
Only the toughest cryptogram puzzles remove spaces and punctuation, and almost none leave any letters unchanged. But both those options would be available in non-puzzle wordplay collections or puzzles that provided other hints.
I found these matches using Adam Aaronson’s excellent Wordlisted (the “Cryptogram” feature) which lets you scan your own wordlist as well as a basic dictionary. Other cryptogram-solving websites will at least supply any words sharing the pattern found in their own extensive dictionaries.
So what’s the problem? If I can find stuff like that, why the complaint?
Simply put, no solution scales. I can’t take a list of roughly 400,000 words and phrases, or one of 10,000 movie stars, and study their patterns without taking the time to look at each one individually. And no one has that kind of time. I can scan my movie-stars list and find six anagram pairs in under twenty minutes. But I can’t tell you anything about the whole list’s letter patterns that isn’t an educated guess.
I think there can be a better approach—but I’d need help to get there. More details tomorrow!