[Community Puzzle] Wordle

Coding Games and Programming Challenges to Code Better

Send your feedback or ask for help here!

Created by @Tanvir_Ahmed1,validated by @NinjaFoxYT,@Gthev and @rc95401.
If you have any issues, feel free to ping them.

Hi, this seems really interesting.
Did you choose the word in a specific database and want us to find those word using that database, or are we to assume that the words are just a string of n letters and don’t mean anything ?

Another question to @Tanvir_Ahmed1: If there’s a fixed list of words, how many words are there in the list?

Also, I don’t really get how it can be done with less than 2 try per word, is it like extreme cherry-picking ?

1 Like

In my opinion, it is up to the player to decide how they want to approach the game. That’s why players are allowed to make guess using unreal word although I’ve used real words for test cases.

I think there is a problem with Wordle validators. Apparently, it’s quite easy to guess them and to hardcode for them. For instance, look at the insane scores of top players. This issue makes it really unattractive for a number of players.
Would it be possible to change Wordle so that the word to guess is fully random ? By this, I mean it changes every time the test case is run. We will ask then CG to reset the leaderboard.

2 Likes

Just to be clear, my previous post is not directed against people hardcoding solutions. I only believe that it is more interesting to design an algorithm to solve a tough problem.

2 Likes

I think it would be more equitable if the list of possible words is shared in the statement (or a link to an external list).

1 Like

I support the existence of the problem. We should either use a public dictionary or completely random validators (this option is more interesting).

No kidding. 50 tests, and a total of 90 guesses to get them all correct?

I started out using a RNG which showed big swings on the same tests. I guess theoretically it’s possible to randomly get all 50 correct in one guess.

I’ve since added an initializer that wastes guesses at the beginning, but it’s more stable. 9 “guesses” for 5 words is just ridiculous.

The word to guess changes every time a validator runs, the major issue is that the list of possible words is super small, there are 12 validators for each word length (5, 6, 7 & 8) and only 12 or 13 possible words for each length.

To be fair to everybody and if it’s not changed here’s the list of possible words:

Summary
const WORDS = [
    5 => ["bluff", "apeak", "yucas", "vertu", "scaup", "green", "stoor", "elect", "zebra", "maill", "story", "hunch"],
    6 => ["jejune","impact","ticket","search","nozzle","cancer","closer","ladder","trophy","prison","tattoo","button", "winter"],
    7 => ["xeroses","acerola","vacancy","impalas","jumbuck","targing","network","pachisi","elution","wafture","packwax","saboted", "equinox"],
    8 => ["dackered","biolytic","medicine","minimize","landlady","spiculum","endoderm","schedule","calypter","quitrent","security","workshop"],
];
1 Like

60 words and 50 tests… no wonder.
Why are they not randomly generated strings?

For instance with 5 letters, there are no {d,j,q,w,x} so there are five letters you can pull from the pool.
Then {g,i,k,m,v,z} all appear once, and {m,i} are in the same word. Guess GIKVZ first and you’ve pretty much nailed it. If any of them appear, you know the word. If not, guess BUCOY. If any of them are type 3, you know the word (if u AND c are 3, it’s YUCAS, otherwise HUNCH), if c AND u are 2 it’s SCAUP… and so on…

12 or 13 possible words for each length is not enough, you can choose the first guess to be sure to have the good word at the second guess.

1 Like

Looks like the puzzle was updated, would have nice to let us know.
Edit: The author did a good job, it’s way better now, fully random for the validators.

1 Like

Looks like the words to guess are just completely random letters instead of real words :open_mouth: Not sure if the author is still testing or the update is finalised.

1 Like

Full random is better than what we had before. But I’m not totally convinced about it either. Real language has more depth than this, considering letter frequency distribution and synergy effects of which letters follow each other. That aspect is completely lost with random letter combinations.

2 Likes

Thank you all for your valuable feedback and patience. I am currently exploring various options to address the issue with the validators. At this point, having a large and diverse open list of real words seems promising. I will continue to work on it and update you all as soon as possible.
Your continued patience and feedback are greatly appreciated. Thank you

7 Likes

When considering a large open word list, I’m leaning towards having a single list with a large number of words for one particular length, rather than having medium-sized lists for each length. For instance, having 10,000 words for a specific length rather than 2,500 words for each length. By length I’m meaning word length. Also, I’m talking about changing Wordle to have words of single length instead of four different lengths.
What are your thoughts on this? Do you have any specific preferences regarding word length?

1 Like

There are some differences in a letter patterns in words of different length.
Part of the challenge was to effectively use these patters depending on word length.

1 Like

The single list for a given length is a certainly a good idea

1 Like