Solving Codeword Puzzles: Letter Frequency, Vowels and Cross-Checks

What makes a codeword different

In a codeword, every letter of a hidden quote is replaced by a number from 1 to 26. The same letter always gets the same number, and no two letters share one. There are no written clues: everything you need is in the pattern of numbers.

Compared with a cryptogram, where letters stand for other letters, numbers have one big advantage. Nothing on the board looks like a word until you place it, so you are never misled by a cipher that happens to spell something. Every square is either solved or unsolved, and you can see your progress at a glance.

Letter frequency in the Smotix quotes

We counted every letter in the 95 quotes in the built-in Smotix library, 4,583 letters in all. The ranking closely follows general English, which is why frequency is such a reliable starting point.

Together, E, T, O, I, A and N make up 53.7% of all the letters. Q and X each make up 0.2%, and Z almost never appears. In practice this means the number that appears most often in a puzzle is very likely E or T, and the next four or five most common numbers almost always come from O, I, A, N and S.

How much the starter letters give you

Every Smotix codeword reveals some numbers at the start, always the most frequent ones in that puzzle. Because the most frequent letters cover so much of the text, even a few starters go a long way. We measured how much of each quote they reveal on average:

• Easy (3 starter letters): 40.5% of the quote's letters are filled in from the start.
• Medium (2 starter letters): 28.4%.
• Hard (1 starter letter): 15.3%.

This is why easy puzzles often feel half-solved from the first second, while hard ones need real work before the first word appears. It also explains the best first move at every level: fill in the starters everywhere, then look for the word that is closest to complete.

Step 1: Find the short words

Short words are the fastest way in, because so few English words fit each pattern. In the Smotix quotes, the most common words are THE (59 times), IS (55), TO (35), OF (22), YOU (21), NOT (19), ARE (18) and AND (18).

The only one-letter words are A and I. In our quotes A appears on its own 17 times and I 12 times. If a single number stands alone and you already know it is not A, it is almost certainly I.

Look next for a three-number word that appears more than once. If its last number is also one of the most frequent numbers in the puzzle, it is very likely THE. Solving THE gives you three letters, including two of the most common, across the whole grid.

Step 2: Use the vowels

Almost every English word contains a vowel. Once you know E and one or two other vowels, look at the unsolved numbers that sit in the middle of many words. Those are probably the remaining vowels, because consonants tend to cluster at the start and end of words.

A word with no known vowel yet is a strong hint: one of its unsolved numbers must be a vowel (or Y). If the word is short, try each vowel in turn and see which one produces a real word.

Step 3: Cross-check every placement

A letter you place appears everywhere its number appears. That is the codeword's built-in check. After each placement, read every word that contains it. If one of them becomes impossible, for example a word with no vowel or a run like QX, undo it straight away.

Keep track of which letters are certain and which are guesses. When you are wrong, you want to undo only the guesses, not minutes of good work.

Step 4: Use the 26-letter limit

Each letter is used by exactly one number. Late in a puzzle this does a lot of the work for you: when only a few letters are still unused, a gap in a word may have only one possible fill. The Smotix keyboard marks letters you have already used, so checking what is left takes a second.

If you get stuck with several unsolved numbers, list the unused letters and try the common ones first. By this stage the remaining letters are usually the rarer consonants such as B, V, K or W, and their positions are easy to test.

Putting it together: a worked routine

Practise this routine on easy puzzles until it feels automatic. On hard puzzles the steps are the same; you just spend longer on step 3.

• Fill in the starter letters everywhere they appear.
• Find one-letter words (A or I) and any repeated three-number word (likely THE).
• Count the numbers. Treat the most frequent unsolved one as E or T, the next few as O, I, A, N or S.
• Complete the word that is closest to finished, then cross-check every word it touches.
• When stuck, check which letters are unused and test them in the gaps.

About these numbers

Letter shares and common words come from the 95 quotes in the built-in Smotix library. The library also grows with quotes stored in our database, so live puzzles may differ slightly. Starter percentages are the average share of each quote's letters covered by its most frequent one, two or three letters, grouped by the quote's difficulty.