By Jason Snell

April 28, 2022 12:05 PM PT

Knotwords offers crossword puzzles… without clues

One of my very favorite game developers, Zach Gage, is back with a new one. Knotwords, by Gage and Jack Schlesinger, is a crossword-puzzle style game with a twist: instead of filling the puzzle via clues, you have to fill various regions of the board with a limited selection of letters.

It’s a logic puzzle involving getting the right words to fit in all the right places. There’s also a second type of puzzle, Twist, which also limits the number of vowels you can use in each row and column, giving you clues—but also limiting your options. There are new puzzles every day in both normal and Twist formats, as well as monthly puzzlebooks that you can complete.

I’ve been playing Knotwords for more than a week, and it’s fantastic. It’s become a daily habit, and if you like crosswords and Wordle and similar types of puzzles, you’ll love it. Knotwords is free, but the Twist puzzles, access to 30 extra monthly puzzlebook puzzles, statistics, access to archived puzzles, and customization options are all locked behind a single in-app purchase. It’s $4.99 a year or $11.99 to unlock everything forever.

I was able to ask Gage a few questions about what inspired Knotwords and how he approaches building games like this. Here’s that interview:

How does Knotwords build on your previous games (or not)? You’ve done SpellTower, TypeShift, and Good Sudoku and I can see the influences of them here.

For years I’ve been making games in more traditional genres that a lot of game designers don’t do original work in. Like making new Solitaire variants, Chess rulesets, or Billiards games. A little joke I’ve had as I release these games is that I always feel like players are less tolerant of my twists and changes than I’d expect. I’m trying to make games for everyone, but I always encounter whole swathes of players who just want “standard chess” or “traditional solitaire” modes inside my games. Because of this feedback, I feel like every game I make gets slightly closer to its source material. SpellTower was wholly unique, but FlipFlop Solitaire is just one major rule change on spider solitaire, Really Bad Chess is just a different starting condition, Good Sudoku is basically just Sudoku with different tools!

That’s a long intro to this point: When I started out as a designer, I wanted to invent entirely new things that people would play, and they’d be dazzled by the novelty. And what I’ve realized after years of doing this is that duplicating something old is easy, but crucially, so is doing something brand new. The real magic trick of truly great game design is doing both — designing something that everyone is already deeply familiar with, even though it’s completely and utterly new.

It took a lot of hard-won lessons over the course of my career to even consider designing like this.

In a less abstract sense, there are a lot of direct lessons from Good Sudoku that came into play with Knotwords. The biggest being that when Jack and I set out to make Good Sudoku we both thought puzzle generation would be a snap. We thought the genius of Sudoku was the beautiful ruleset. But like all ideas, it turns out the genius of Sudoku is in the work—the puzzle construction. Realizing how important a great puzzle generator is to these kinds of games really took the beautiful ruleset down off the pedestal for me, and gave Jack and me the design insights we needed to return to a failed prototype we’d built a year before and turn it into the Knotwords you see today.

So much of what feels so great when you play a Knotwords puzzle is because of the generator that Jack built and the process we developed during Good Sudoku for collaborating on generator construction.

This is a game that feels like a crossword puzzle… but doesn’t have any clues. How did you consider the puzzle mechanics of crosswords as you were designing this game?

Yes absolutely. Knotwords began as a game about the process of crossword construction.

Something I noticed with crosswords is when you play them, the focus is very heavily on solving the clues. The board itself is almost always just functioning as support for the clue game you’re playing. But if you’re constructing crossword puzzles, the board is a big part — actually fitting the words in and making them work with your theme is a huge challenge. I thought there might be a game in that challenge when I came up with the original prototype.

Later when we were thinking about how to do hints in the game, clues in crosswords came up again, and we thought a more interesting hint mechanic than just showing wrong letters or giving players answers, would be to use censored definitions of the word you’re looking for as sort of an orthogonal hint.

The game is based on daily puzzles that build over a week. Any specific inspiration for this?

Thinking about and designing new meta-structures around small games is something I’ve focused on throughout my career — Daily puzzles, wagering modes, weekly challenges, and instant tournaments — so I knew dailies would be a big part of Knotwords right from the start. In Good Sudoku, puzzles get harder throughout the week, culminating on Sunday with a particularly difficult challenge.

In Knotwords we keep that structure, but we also take a bit of a clue from crosswords and have more of a stochastic difficulty throughout the week. Friday puzzles are less dense and more challenging, and Saturday puzzles, while being large, are filled mostly with smaller words than you would expect.

We also have two 30-puzzle “books” that are specific to the month and must be completed within the month. Jack is really into the “gem-road” layouts that you see in games like Candy Crush, and we wanted to have something for players who were interested in that sort of challenge.

The Twist puzzle feels very related to Sudoku and adds this tactical vowel element. What was the inspiration there?

I love adding little twists or alternate modes in my games. Usually the ground around great game mechanics is littered with other interesting ideas, and I think players get more joy from my games when I can widen their view of the play space.

The numbers in Twist puzzles are directly inspired by Picross. They tell you how many vowels exist within their row or column in the puzzle solution. When I was designing Twist mode, I wanted a mechanic that was lightweight (only one additional rule), and that would expand what was already interesting about the game (putting in letters gives you nearby answers). The cool thing about Twist numbers is that they extend the impact of your letter placements far beyond their neighbors, all the way to the other side of the puzzle.

Are there multiple solutions for the regular puzzles? As long as it fits, is it legal? And I assume the Twist puzzle has only one solution.

All puzzles in the game have one valid solution. Rarely, maybe one in a thousand puzzles will have an alternate solution, but usually these alternate solutions are just a single word being altered (like UNTIE and UNITE). Twist puzzles are specifically built so that they would have multiple valid solutions if the numbers weren’t taken into account, but only one solution that fits the numbers.

What went into selecting the word lists? What was the philosophy there? Part of the strategy of something like Wordle is guessing what words are likely to be solutions. I’ve been surprised by some of the words in Knotwords – it feels more Scrabble-like to me.

Wordle was able to hand-build their list because in Wordle, one word equals one puzzle. For us, it can take thousands of words to become one puzzle because those words have to fit together (not all words fit well) and because we don’t want to reuse the same words in puzzles that appear within the same month.

I took a lot of lessons I’ve learned making word games into the choices we made for Knotwords.

We use a number of different lists that let us know what kinds of words people are likely to know and supplement with the entire dictionary minus slurs, crude words, and archaic spellings. It’s an ongoing process, and we have some tools in place to allow us to remotely block things that we’ve missed.

Knotwords is available on the App Store, Google Play, and Steam.

If you appreciate articles like this one, support us by becoming a Six Colors subscriber. Subscribers get access to an exclusive podcast, members-only stories, and a special community.