Consider, if you will, the following groups of letters. AI; AOEIUMBH; AEOIUYHBCK; AEOIUYSBF; SEAIOUYH; EAIOUSY. Any idea what they represent?

Well some propellorhead called Nick Berry has worked out these are the letters and sequences to maximise your chances of winning a game of Hangman when the word you’re seeking is 1,2,3,4,5, or 6 letters long. He actually worked it out up to 20 letters, where your best guess would be IE.

If you play Hangman, you pretty rapidly come to the conclusion that the frequency of letter use should give you a good start. Most people are aware that E is the most common letter in English. Some will be aware that more words start with a T than any other letter. But it’s not quite that simple. This is English, after all, the bastard offspring of about a zillion different languages. Bastard in all senses of the word.

The letter frequency changes with the length of the word. That’s a bit counterintuitive, but Mr Berry’s computer algorithm shows this to be true. Hence the sequences detailed above. What’s even odder is the way that the best choices reduce in number as the word gets longer than five letters, hence the simple IE for 20 letter words. This flies in the face of reason, but since this bloke has taken the time and effort to do the research I have to trust him at his word. He’s a data scientist for Facebook, and like it or love it, Zuckerberg’s acolytes over at Menlo Park are good at number crunching. Hence the ability to increase your chances of guessing the first letter in any given length of word. Cunning.

Cunning but somewhat pointless. It gets worse. He’s also done a computer analysis to determine the last word anybody would guess using his methodology. Care to take a guess?

It’s ‘jazz.’

So now you know how to beat the next five year old who challenges you to a game when it’s the school holidays and tipping down with rain.

Advertisements