Hacker News Reader

Comment by iainmerrick | Hacker News Reader

Wordle is so amazingly well-designed. I’m starting to feel it’s comparable to Tetris in how deceptively simple but well-balanced it is. The choice of 5-letter words and 6 guesses gives just the right level of difficulty; these optimality results back up that intuition pretty well.

View on HN

The author also did a great job of building a game that scales. No server, everything is in a very small amount of HTML and JavaScript, easily cached. He can have the world play with no ads at very low cost. Of course, this means people can easily cheat, but where's the fun in that?

It also demonstrates how easy it is to ruin things by caring about monetizing them. A lot of what's great about it wouldn't really be possible if you cared about making money off it.

all the Wordle dev would have to add is a single link to a place to donate. He would make a fair amount given it's current popularity. But I agree with your point.

Without diminishing any accolades to the author, this is exactly how I would design the game - and the most sensible approach. It's the bulk of the web that is shit these days, and the low expectations and surprise I see expressed is a testament to how bad things have gotten.

The entire allowed word list is in the JavaScript?

Also, the entire list of solutions for every single day.

I think it would be possible to build an even faster, more robust solution with plain HTML, CSS and a Rails / PHP server on the backend. The less propagation of JavaScript, the better.

Wouldn't that mean a network request and full page reload for every single guess? Genuinely curious how that could be faster than it is.

The current implementation has easily survived a worldwide hug of death due its massive viral success. It’s plenty robust.

Your server-side PHP version would be slower for most users and might get expensive to run.

I'm curious how adding network requests and server dependencies to code makes it faster and more robust than code that works entirely offline.

Rails and PHP would be the last thing on my mind when scaling a project like this one. Worldle only has to serve a bunch of static files to clients (the exact same files to everyone). How would this be improved by having back and forth with a server to play the game?

It is also a fun little game to program/make a clone of

I made one simple web page with JS, HTML, and CSS, and now I can play it whenever I want (or show it off to my friends who introduced me to Wordle a few days ago).

Essentially Wordle is just Hangman with a neat twist.

> Essentially Wordle is just Hangman with a neat twist.

I'd say the closer comparison is to Mastermind[0].

[0]: https://en.wikipedia.org/wiki/Mastermind_(board_game)

Wordle is basically a clone of Lingo, so that'd be the closest comparison available.

https://en.wikipedia.org/wiki/Lingo_(American_game_show)

I was thinking about Mastermind too. I played that one as a kid, but Wordle is a better version of the same idea:

- Mastermind is based on abstract codes, Wordle is words. Words are fun!

- Wordle is much easier to explain. I could explain it to somebody who hadn’t heard of it and we could be playing on pen and paper inside two minutes.

I initially thought it's just like Mastermind, but with Mastermind you are not told which colours (corresponding to letters) are right: you are just told there is one right in the right place and two right in the wrong places, for example.

it reminds me of the puzzles that had a starting word, an ending word, and you could only change one character at a time to make a new word to finally achieve the ending word, and mixed in with the game "master mind"

I believe they are called “word ladders” [0]. Coincidentally, I was listening to Genesis’s song “Supper’s Ready”, yesterday. It includes 2 word ladders in its lyrics.

  Mum to mud to mad to Dad

and

  Dad to dam to dum to Mum

https://en.m.wikipedia.org/wiki/Word_ladder

that's the one!

I had the same exact idea. It turned into a fun weekend project[0] that allowed me to dust off my vanilla JS skills and dive deeper into Tailwind CSS and put my own spin on the concept by basically turning it back into an asymmetric multiplayer game.

I do commend Josh Wardle for implementing Wordle with native web components, I personally wouldn't have the patience.

[0] https://word.rodeo

Games with less complexity are generally easier to balance, so one shouldn't be too surprised.

I'd consider Tetris to be far more impressive design-wise in that it's a easy to learn but difficult to master, despite its simplicity. It's a mind-blowing experience to watch the best Tetris players in the world. The skill gap between a beginner Wordle player and the best in the world simply cannot be that great, by nature of the game's design. I guess a large part of the reason is because the player can't alter the game state: they can only gain information. It doesn't allow for complexity to emerge.

That's fair -- I definitely agree that Tetris is deeper, in the "minutes to learn, a lifetime to master" sense.

Partly that's state, as you say, but I think it's also significant that Tetris is a realtime arcade-style game. Experts are vastly faster than newbies, and the main difficulty increase mechanism is that the blocks drop faster.

Wordle might have some similarities there. People don't commonly share their times, but I bet even among competent players who routinely solve it in 4-5 moves every day, there are some expert players who can do it much faster -- just like some people can solve the daily crossword in a few minutes, while others work on it gradually over the whole day.

You could also distinguish between people who usually take 4-5 moves and those who can do it in 3-4 (e.g. optimal play with its average of ~3.4 moves). It's just not obvious immediately because there's a lot of variance; if a great player solves it in 3 you could write it off as a fluke, and if a poor player is genuinely lucky and gets it in 2 you could mistake that for skill.

Right, if time to solve were included in the shared results, you’d be able to pick up on skill variation a lot easier, and players would start to regard speed as something to improve at. I think it would make the game more interesting (not that it isn’t already).

Hmm, I feel like 6 guesses is fine for your first few games. But once you begin to think about pruning and letter frequency strategies, you'd probably never hit 6.

Luck takes over after a certain point and I don't know that the solver programs and averages really reflect this. If you're 3 guesses in and looking at _ _ACK, skill isn't necessarily the thing that's going to distinguish a 4 result from a 6.

Maybe it's just me that's very bad. I don't use any self-computed or computed word lists though.. :)

We can't share the wordle emoji thing here because hacker news strips something from it that removes the colors.

I don't feel like I'm good at this but I hit 6/6 only once - on my first day.

I'm in a group chat where we all share our Wordle results for the day. It seems everyone is at least as good as me.

2048 gave me a similar feeling.

2048 could be defeated with a simple repeated key sequence. Not so this. No comparison.

i find this much easier than 2048

It's also much less time-consuming.

You mean Threes, I hope!

Threes is the better game, IIRC there's only one rule difference but it makes all the difference. And I really feel for the small team that spent months designing Threes only for a clone of a clone to be the one that everyone remembers.

No, that's a completely different game, with different mechanics and a very different feeling. Threes feels more slow and methodical and 2048 feels more fluid and slidey.

Both are very good games, but only one of them became a world wide sensation.

I don’t recall 2048 being nearly as big as Wordle, or say Flappy Bird, but of course it’s fairly subjective.

I think it's too derivative of Fallout's computer hacking to garner that much praise.

Reply on news.ycombinator.com