How to Solve a Cryptogram

Frequency is the strongest signal you have. In ordinary English, E accounts for about 12.7% of letters, then T at 9.1%, A at 8.2%, O at 7.5%, I at 7.0% and N at 6.7%. The old printer's mnemonic ETAOIN SHRDLU lists them roughly in order. In a short cryptogram the ranking will not be exact, but the most frequent symbol is nearly always one of E, T or A — so pencil in E and see whether the rest stays sensible.

A standalone symbol can only be A or I. That is a free letter from a single glance, and both are common enough to constrain everything around them. If the puzzle has two different one-letter words, you have both and can stop guessing.

Two-letter words are a small set: OF, TO, IN, IS, IT, BE, AS, AT, SO, ON, OR, AN, DO, UP, MY, BY, WE, HE, NO. Three-letter words are dominated by THE, AND, FOR, ARE, BUT, NOT, YOU, ALL, CAN, HAD, HER, HIS, ONE, OUT, WAS. If a three-letter word appears twice in a short text, THE is the first thing to try — and it hands you T, H and E at once.

Two identical symbols side by side is a strong tell. In English the common doubles are LL, EE, SS, OO, TT, FF, RR, NN, PP and CC — roughly in that order. Doubles almost never occur with A, I, U, H or Y, so a doubled symbol rules out a large part of the alphabet in one move.

An apostrophe is nearly a gift. The letter after one is almost always T, S, RE, LL, VE or D, and the letter before a lone 'T is almost always N (as in DON'T, CAN'T, ISN'T). One apostrophe often unlocks three letters.

Once a few letters are in place, English word shapes do the rest. Watch for -ING, -ED, -LY, -TION, -MENT and -NESS. A three-symbol ending whose middle symbol is already N is very often -ING, and that confirms G, which is rare enough elsewhere to be worth having.

At some point you will recognise a word from its shape rather than deduce it. That is fine — but propagate the guess through the rest of the puzzle before believing it. If it forces an impossible sequence somewhere else, the guess was wrong, and finding that out early costs less than unwinding ten letters later.

The twelve most common letters in ordinary English, as a share of all letters. Your most frequent symbol is very likely one of the first three.

Here is the Christmas puzzle from the Emopuz archive, exactly as it appears on its own page:

Start with the shape, not the symbols. There are 2 words and the first is 5 symbols long. Look for two identical symbols side by side — in English a double is most often LL, EE, SS, OO or TT, which rules out most of the alphabet in one move.

Now count the repeats. The most frequent symbol here appears 3 times. A symbol that both doubles and recurs elsewhere narrows the field fast: fix that one letter and the surrounding words start to resolve on their own.

The fastest way to get good at this is volume. Emopuz has a free emoji cryptogram for every special day of the year, and a solver that does the frequency counting for you while you concentrate on the deductions.