A tool for exploring Wordle consonant-vowel patterns

I hope you’re having fun solving Wordle puzzles; I know I am. As I attempt the puzzles, I’m often thinking about how vowels and consonants appear in different patterns in five-letter words. I wanted to start exploring those patterns, so I wrote a Python program to probe the results of different patterns. If you want to jump right to it, it’s here.

To start, ask yourself this question: What pattern of consonants and vowels do you think is the most common? And what is second?

Let’s see if you’re right. To understand the result shown here from my Wordle Solution Explorer tool, note that & represents any consonant and + is any vowel. (I make the simplifying assumption that Y is a consonant at the start of a word and a vowel anywhere else, which is a decent approximation.) Here’s the list of frequencies for the consonant-vowel patterns:

Most common consonant-vowel patterns (& = consonant, + = vowel)
         &+&+& 403
	 &+&&+ 396
	 &&+&& 352
	 &&+&+ 212
	 &&++& 194
	 &++&& 137
	 &++&+ 121
	 +&&+& 99
	 &+&&& 82
	 +&+&+ 75
	 &+&++ 54
	 +&+&& 49
	 +&++& 28
	 +&&++ 23
	 +&&&+ 22
	 &&&+& 19
	 &+++& 15
	 ++&+& 10
	 ++&&+ 8
	 &&&++ 5
	 &++++ 4
	 ++&&& 3
	 ++&++ 2
	 +++&& 1
	 +&&&& 1

You may have guessed that the most common Wordle solutions start with a consonant and alternate between vowels and consonants (for example, PANEL and LOSER). That pattern accounts for 403 of the 2315 possible solutions, or 17%. But the pattern &+&&+ (for example, PARSE or BERRY) is almost as common, making up another 17% of solutions. The seven most common patterns start with a consonant, but three of those patterns end with a vowel. Can you identify the single word that starts with a vowel followed by four consonants? If this question makes you feel existential dread, you’re on the right track.

While there are 32 possible combinations of consonants and vowels (25), but only 25 of those appear in the Wordle solution set. None of the Wordle solutions have all vowels or all consonants, for example, and the pattern starting with four consonants and ending with a vowel is also missing.

Of course, once you’ve got a few letters nailed down, these frequencies change. As I shared in January, the most common letter in a specific position is E as the last letter of the word. If you correctly guess that the word ends in E, what pattern of the other four consonants and vowels is most likely? Here’s what you’ll find:

Most common consonant-vowel patterns for words ending in E (& = consonant, + = vowel)
         &&+&E 165
	 &+&&E 104
	 &++&E 57
	 +&+&E 54
	 &+&+E 19
	 +&&+E 9
	 +&&&E 7
	 ++&&E 3
	 &&&+E 3
	 &+++E 2
	 ++&+E 1

Although without knowing any letters, &+&&+ is the second-most common consonant-vowel pattern, that changes when you know the word ends in E. Now the word is more likely to start with two consonants (for example, THERE or CRATE). And again, only 11 of the 16 possible combinations exist in the set of solutions: there are no words starting with three vowels and ending in E, for example.

The second-most-common letter to end a word is Y. If you know a word ends in Y, here are the frequencies of the consonant-vowel patterns:

Most common consonant-vowel patterns for words ending in Y (& = consonant, + = vowel)
	 &+&&Y 221
	 &++&Y 53
	 &&+&Y 32
	 &+&+Y 18
	 +&+&Y 14
	 +&&+Y 11
	 +&&&Y 10
	 ++&&Y 2
	 &&&+Y 2
	 &+++Y 1

So the smart guesser, having put a Y in its place at the end, is guessing words that match &+&&Y, such as FERRY and BUSHY. Double letters are common in such words, but they create a problem, too; if you guess words with repeated letters, you’ll be probing a smaller number of letters in each turn.

The most common letter for Wordle solutions to start with is S. Here are the frequencies for those words:

Most common consonant-vowel patterns for words starting with S (& = consonant, + = vowel)
	 S&+&& 126
	 S&+&+ 78
	 S&++& 64
	 S+&+& 29
	 S+&&+ 28
	 S&&+& 15
	 S++&+ 12
	 S++&& 5
	 S&&++ 4
	 S+&++ 3
	 S+&&& 2

The trend that jumps out of this list is that S is highly likely to be followed by a consonant (287 out of 366 Wordle solutions, or 78%). So once you nail down that S, you’ll likely want to try words with two or three more consonants to figure out where they might fall in the word.

Helping without giving away the answer

To go a bit further helping people think about consonant and vowel combinations, I made it possible to enter any such combination and see the first 20 Wordle solutions that match it. (I didn’t want to show longer lists, that makes the game too easy.) If you’re stuck, this can often help you think of more word ideas.

For example, suppose you know the word starts with B and ends with E. You theorize the vowel-consonant pattern B&+&E. How many words match that? There are 13, from BLADE to BRUTE. But there are also 13 words that match B+&&E, from BADGE to BUTTE. Seeing these lists might help you decide which combinations of vowels and consonants to test.

One consequence of the way this tool works is that you can test any word as a possible solution. You can verify that YEARN is a possible Wordle solution, but YOURS is not (too similar to a plural, I guess). If you’re not sure whether a potential word is “common” enough to qualify as a Wordle solution, you can use my tool to vet your guess before entering it.

Wordle should be fun. I wanted to contribute in that spirit.

Wordle is just the right kind of puzzle for the pandemic-crazed world: bite-sized, brain-tickling, and more of a challenge to one’s self than a competition. I’ve been pleased to see that people generally don’t share the answer, to avoid spoiling everyone else’s fun. Now that the New York Times has bought Wordle, I’m hoping it will continue to be enjoyable.

It would have been easy enough to write a tool that generated all possible matches for any given set of Wordle results, but I’m sure such tools already exist (and I won’t be linking to them). More importantly, my objective was to help people learn more about the structure of words, not to make it easy to just cheat. Surprisingly, I’ve found that my long experience with Scrabble is not a help here, since Scrabble players are constantly looking for obscure words like QANAT, prefixes (like UN- or IN-), and suffixes (like -ES, -ED, and -ING), but most of those are intentionally missing in the Wordle solution list.

Put simply, I wanted to contribute to the fun, not spoil it. For me, creating this analysis tool was way to expand the ways to explore the relationships among letters, words, and language — and a way to challenge myself as a novice Python programmer to create something worthwhile. I’m sharing it with the world in the same spirit.

Don’t use the Wordle Solution Explorer to compete — where’s the fun in that? Use it to learn, and to get unstuck. Please enjoy it in the spirit in which I created it.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

5 Comments

  1. Miaou, which I never would expect to be used in Wordle, but which I have used often playing Words With Friends. For someone like yourself who possesses a large vocabulary to start with, I doubt your enjoyment of playing Wordle will be enhanced by defining the game via Python. I personally prefer to challenge my brain and not use computer support. I once spent time trying to see if I could figure out which group of three letters could create the most three-letter words. After deciding I had gotten the solution, I challenged folks I worked with at UConn to see if they could come up with the same solution. One person did, but he did it by downloading all the three letter words that existed and used a spreadsheet to create all possible variations of three letters, then matched the results against the download. I doubt his exercise exercised his brain as well as my process exercised mine. Anyway, to each his own, eh? Take care! Bruce

  2. Paul, if plurals are ok I’ll go with strengths. Otherwise I’ll keep looking for a Yiddish or German word that’s part of our language now, like schmaltz.