How Bracket City went from diner placemats to The Atlantic's 'digital puzzle empire'

Earlier this month, The Atlantic introduced its readers to Bracket City, a word puzzle that tasks players with solving nested clues — embedded in the titular brackets. I’m trying to think of whether the magazine made any other big news around that time. Nope. Just the Bracket City launch.

The puzzle is tough, especially on the first few tries. But Bracket City quickly entered my rotation of daily puzzles, alongside a handful of other word games also run by major publications. (You know the ones.)

I spoke with Ben Gross, the puzzle’s creator, and Caleb Madison, director of games at The Atlantic, about the game's origins, how it came to the attention of a 168-year old magazine, and the future of The Atlantic’s game offerings.

This interview has been lightly edited for length and clarity.

ReaderGrev: Tell me a little bit about the origins of Bracket City. How did it get started?

Ben Gross: The full story is that I co-own a diner in Brooklyn Heights called Montague Diner, and at the diner I co-host a puzzle night where we do sort of diner specific puzzles. I co-host it, actually, with Wyna Liu¹ who makes the Connections puzzles for The New York Times. We do site-specific puzzles for the diner that are sort of about the diner, and they live on a placemat, and people come, and it's really fun. The closest analog is like an escape room, but you don't have to get up from your booth to play.

So we're always having to come up with puzzles, and I came up with this format of these nested clues and brackets and we actually printed it on the placemat. It was a paper version, it was much simpler than a typical Bracket City is today, and it resolved to a single number for a combination of a lock on a pouch on the table. And then people liked it and they thought it was fun, so over Christmas break I resolved to turn it into an actual game. I spent a lot of time at my wife's family's house in Arkansas not talking to them and instead programming Bracket City. And I got it to a condition where it could be up publicly around the beginning of the year.

Caleb, can you tell me how Bracket City got onto The Atlantic's radar?

Caleb Madison: An editor at The Atlantic, who's actually a friend of Ben's, Juliet Lapidos, sent it to me as my wife and I were driving to pick up groceries. I solved it immediately and was like: turn the car around — because it was the best puzzle I had ever encountered. And so I told The Atlantic that, and the rest is history. The rest is inscribed in the Bracket City annals.

What was the process like of getting this in front of your bosses, management, the editors, and saying: Let's make this a part of The Atlantic?²

Caleb: Honestly, the puzzle kind of spoke for itself. When my bosses were in a room with 10 people, all of whom were addicted to the puzzle already, it became a pretty clear choice that this was a puzzle that had amazing potential to reach and retain a lot of puzzle solvers. So truly the work spoke for itself.

I don’t know the right word to use for this: Can you tell me about the terms of the sale(?) or transaction(?). What would you call this, and what can you tell me about the details?

Caleb: It's not an acquisition. It's a license, with a burning desire to buy. But the goal is that our collaboration can yield something greater than the sum of its parts and make Bracket City into the center of the digital puzzle empire that it's destined to be. I think it's an amazing fit with The Atlantic, and Ben and I have already had a lot of fun thinking about how the puzzle could grow. The goal is that we build this together.

Ben: From the literal terms, I don't think that we can say anything else. That’s just how it goes with these kinds of deals.

Now that this is in front of The Atlantic's audience, is there anything you’ve thought to change about the puzzle? 

Caleb: There's a slight thing that I've pushed for: Just to make it overall a little bit easier, because we're bringing a new audience in and we want them to have as little barrier to entry as possible in terms of learning the mechanics of the game so that we can really turn the heat up on them and make them sweat eventually.

Ben: Yeah, there are some really weird clue formats that I've experimented with. And Caleb was like: let's hang on to a couple of those till people know how this game works. Because one of the trickiest things about the game is that I think it's actually not an incredibly difficult puzzle, but it's an intimidating visual when you first load it. It's such a chaotic looking starting place. And so I think It's good to ease people into it a little bit. I'm really on board with that. … But there's been no heavy hand telling me to make it anything really other than what I want to be.

How does Bracket City fit into The Atlantic’s portfolio?

Caleb: I think it walks a great line between intellectual and fun. It has one foot in the fun room, and the other foot in the smart room, and that's my favorite threshold to straddle. I love that it resolves to a fun fact from today in history, because I also think no matter what your experience of solving the puzzle is, you learn a little bit of something, but it doesn't feel like homework. It feels like you're learning something that is more of a story than a fact. And I think that's really fun.

I can't say too much, but there are more games to come in The Atlantic in the year, and they all kind of fall within that framework of bringing a sense of play and joy and delight to The Atlantic's history of intellectual rigor.

When can people expect to see that?

Caleb: All I can say is that it is going to be soon, and the world will never be the same.

Ben, can you walk me through the creation of a puzzle? Where do you start?

Ben: So the first step is you start with the solution and you work backwards. I actually spend an inordinate amount of time picking the solutions. Like, sometimes it takes as long to pick the solution as it does to make the puzzle. I want something that's interesting and often something that you might not know, but it's kind of fun and not too dark or weird, but is a delightful thing to be revealed.

I have a tool that I made that lets me build the puzzles. I can highlight any word or part of a word and replace it with a bracketed clue, and that interface makes it — compared to the beginning when I was sort of doing it in a more manual way — a pleasure. I just sort of look for words, and now, increasingly, partials of words that are like surprising words that are hidden within other words, that I think, even if I don't know what the clue is, that I imagine that there's a fun universe of clues that I could write for it.

I always say the spirit of Bracket City is like a solver using a parent bracket or even more than one parent bracket in an extreme case, but like using the parent bracket to figure out and disambiguate a clue that they don't know the answer to. A lot of people play the game and don't do this yet. As people get more familiar with the format, this will become a more popular thing to do. But that's what I love: when I see people solve, and they know it's one or two things, or they're not sure what it is and then they look at when that bracket is solved and that clue is resolved, how it fits into the next clue, and they know the answer to that clue so they're able to sort of a reverse engineer the solution. I really try to facilitate that kind of experience through the way the clues fit together. Caleb points this out to me a lot when he's reviewing the puzzles. A hard clue with a parent bracket that doesn't give you anything is really not ideal.

I basically aim for like 16 to 20 clues, generally speaking. It sort of just has felt like a satisfying duration. And then I show it to Caleb and he says, this is terrible or he says it's good and then we do it.

Are there any solutions you've wanted to use and then thought: this is too out there, or we can't do this?

Ben: Not yet. A great source of solutions is the Wikipedia page for a day, like, every day has a Wikipedia page. You can go to, like, the Wiki page for April 14th. And what you'll find on most Wikipedia pages for a day is a lot of plane crashes and battles and death and slave rebellions. Like, I'm looking at it right now. The Four Dead in Five Seconds Gunfight occurs in El Paso.³ Muslims in the Ottoman Empire massacre Armenians. Bombay explosion.

Today, actually, is like the pushing the edge of darkness. [ed. note: This interview happened on April 13.] Today’s solution is oxygen tanks explode on Apollo 13. Caleb and I talked about this. It's kind of a happy story. Everybody survived, and it was an amazing moment of engineering ingenuity. So we probably won't push it too much darker than that. Probably.

Caleb: I feel like danger is okay, death is not okay. We like a little danger.

Ben: Yeah, we try to avoid any true tragedies.⁴

Caleb, after Ben files a puzzle to you, what is your role in getting it to the page?

Caleb: I have a kind of puzzles and games lab that I'm not– I don't think it would be wise for me to disclose. I run it through a series of high-tech tests to figure out how durable it is, and the Enjoyability Factor™️. There's several chemical concoctions that I've devised that are really useful in testing the fun levels of a puzzle.

But then after that, I usually just play it and try to be aware of my experience: where I have drag, where I speed up. I try to think about, like with every puzzle or game or experience, in terms of kind of a narrative arc, where I like when it begins, and it's easy to enter, and then it kind of gets a little bit hard in the second act, and then there's a race to finish that feels exhilarating and exciting.

I'm curious about where the city theming came from. Why is it Bracket City?

Ben: I think probably subconsciously, I'm a creature of cities. I grew up here in New York. I’ve done a lot of city work in my past lives and stuff. I just love cities in general.⁵

I also really love the idea of… once I had the game, I almost immediately knew I wanted to do ranks. [ed. note: When you finish the puzzle, you are ranked based on how many errors you made.] I was very drawn to the idea of really trying to show how the true machine works in Bracket City. Like, the mayor is not the top [rank], you know?

And then I think a final piece of the puzzle was when I built it, I was trying to decide what to call it and what the URL would be. I just was looking at URLs and I didn't even, I don't think I knew that .city was a top level domain. And then I saw like, oh, bracket dot city is available. That's pretty good.

Thanks for reading ReaderGrev! Consider sharing it with a friend, on Discord, Twitter, LinkedIn, or even a subreddit where folks might appreciate it. Word of mouth helps this newsletter grow!

If you have a tip, I can be reached on Bluesky or via email at mikhail (at) readergrev (dot) com.

1. I asked Ben about the apparent puzzle community around his diner (a thing I do not traditionally associate with diners). He said: “I opened the diner with some friends from high school, and I actually went to high school with Wyna also, and so we've been into puzzles and doing projects together for a long time. Once we had the diner, all of us, the group of us that own it together, now we have a fun place to do things we want to do and like: I'm into puzzles, so I’m going to do a puzzle night with Wyna.”

2. When I worked for The Washington Post’s video game vertical, Launcher, my editor floated the idea that maybe we might want to partner with or acquire Wordle — well before we had any inkling that The New York Times was considering the same exact thing. I don’t want to overstate things: I can’t say for a fact that this idea was ever seriously discussed internally. But I’m generally aware that getting the ball rolling on something like this in a big organization can take a lot of pushing.

3. Both Caleb and Ben agreed that this sounded like a cool solution, and I will add that it’s a cool name for a gunfight. The Wikipedia page also includes this fun phrase: "You big son of a bitch! You murdered me!" 👍️👍️ 

4. Speaking of dates: Ben advised readers to check out the April 1 puzzle. Caleb warned that upon completion, “your Spotify algorithm will never be the same.”

5. Caleb later acknowledged that he knew the real life location of Bracket City, but said he was “not at liberty to discuss the location.”