[huhtenberg]

Good idea, but the lack of progression makes the gameplay a bit tedious and not engaging. That is, you guess and it's binary - OK / NOT OK. Nothing changes.

Some sort of progress would be nice, even if it's just a list of past guesses. Perhaps do what Wordle does - show if guessed letters are in the right place or present in the answer.

With regards to lifelines - allow selecting which letter to reveal. Because in today's puzzle case it's obvious that it's --A, so "reveal the last letter" lifeline is basically useless.

[triword]

I agree that puzzles with progression are the best kind. This puzzle just isn't that kind of puzzle. I didn't want to make another wordle clone, there's a thousand of those and everyone's tired of that style being copied over and over. It isn't for everyone, and that's okay! A lot of people like NYT Connections a lot hate it. Same with Spelling Bee and same with crosswords.

As for the letter lifeline, I am curious how you decided that the third letter's identity is obvious. There are plenty of words that end in ISS, IZE, and ILL, not to mention OSS, OZE, and OLL (e.g. SCROLL, EXTOLL). There is also USS, UZE, and ULL, (e.g. GAUZE, TRUSS, BULL, although they don't fit.)

[ofcourseyoudo]

The problem is that without the progression, it leads very quickly to either brute force or giving up. The lifelines don't feel like earned progression (even if they function that way in terms of solving), they feel like a minor defeat.

[quuxplusone]

I agree with the grandparent comment: Puzzles with progression are the best kind, and this is exactly that kind of puzzle. The only problem is that you didn't implement the progression parts yet. But you still can!

When the player guesses three letters, e.g. if they guess BRA, they're implicitly asking whether the answer is BRASS/ZEBRA/?????. For each of those words, there are several ways their guess can be "wrong": (1) The guess doesn't fit this clue at all: none of BRALL, LBRAL, LLBRA is a word. (2) The guess fits this clue, but its placement is wrong: ZEBRA is a word, but we're actually expecting ___ZE, not ZE___. (3) The guess fits this clue, but it's not the expected word: BRASS is a word, but only the A is in the right place to match the underlying expected word. ZEBRA is a word, but none of the letters are a match.

You should surface some of that new information to the player. You could do that Wordle-style by just surfacing it automatically with no further interaction; or, since there are sometimes different kinds of new information, you could let the player interactively choose what kind of new information they'd like to see as a reward for their guess: See which letters are right? See which placements are right? And so on. And, since there are three clues in play at once, maybe you allow the player to get new information from only one clue per guess.

Concretely: The clues are SS/ZE/LL. The player guesses BRA. You fill in "BRASS" and "ZEBRA" in yellow and "BRALL" in red. Tools appear to the right of BRASS — "Check placement" and "Check letters" — and also to the right of ZEBRA. The player clicks "Check placement" on ZEBRA and gets a negative response (which is recorded visibly somehow). A new prompt appears. The player guesses PAR. You fill in "SPARS" in yellow, PARZE and PARLL in red. Tools appear to the right of SPARS. The player clicks "Check letters" and the A turns green, indicating that the first word has an A in the third position. (Maybe the second S also turns green, and the first S turns yellow — Wordle rules. But that feels redundant with "Check placement.") Alternatively, the A turns yellow, indicating that there's an A somewhere in the correct three-letter answer; and the P and R are grayed out on the keyboard. A new prompt appears... And so on.

This kind of progression would make the game much more engaging!

[triword]

I appreciate the well-thought-out comment. I'm reluctant to imitate Wordle's style of successive partial confirmation, which I feel is a tired pattern now. It feels like there's a trillion [*]dles, all of whom operate on that principle. It might have made a better game in a vacuum, but in the context of that deluge is it still better? I'm not so sure.

[tzs]

> A lot of people like NYT Connections a lot hate it

I was almost turned off of Connections by the poor instructions.

The instructions say that categories 'will always be more specific than "5-LETTER WORDS," "NAMES" or "VERBS"'.

Then one of the first puzzles I did had "PRONOUNS" as a category! That did not make a good first impression.

The interface also probably turns off some people who would otherwise enjoy it. It would be greatly improved if they replaced the "Shuffle" with letting you re-arrange the words by dragging them.

[GrinningFool]

> hen one of the first puzzles I did had "PRONOUNS" as a category! That did not make a good first impression.

I think there's only about 120 or so pronouns, so having this as a category isn't as unreasonable as it first seems.

[gnicholas]

Their conception of 'category' is ever-evolving. Recently it was "ingredients in minestrone", but of course that's just a random subset of "food". They also do things like "numbers minus a letter".

I still play Connections, and I do what was written up as "hard mode" in the NYT today: try to enter the matching quadruplets in reverser order, starting with purple.

[forbiddenvoid]

I don't necessarily think this game needs progression at all. Wordle is a game where you gain information with each attempt. Triword doesn't need any additional information at all. There's one solution (or should only be one solution), and it can be deduced from only the information provided.

If there is any kind of progression, I could imagine it just being successive hints:

- One random letter, but no information about placement - The correct placement of the random letter in the affix - The correct placement of the affix to the stems

But really, I think the simplicity of this as a pure logic puzzle is great all on its own.

[SantalBlush]

If there is only one solution, no progression, and each of the three words is meant to be commonly known, then what is the point in adding an input field to check whether I have the correct answer?

If I have my three letters, then I can easily see for myself whether or not my guess is correct just by looking at the screen; the input field does not help me at all.

[triword]

A question I asked myself when I first created it. Some people will be annoyed by that, some won't. It's fine if you're the former! but the fact is, without an input field, there's no way to verify that your solution is accepted and mark the puzzle as completed.