Last week I wrote about Turing Machine, a deduction puzzle that fascinated me with its cardboard computer but ultimately left me a little cold with the multiplayer experience. This week I’m talking about designers Fabien Gridel and Yoann Levet’s follow-up game, Kronologic: Paris 1920. It turns out the same designers have taken some of those clever ideas and turned them into something that feels much more like a game you’d actually want to sit down and play with other people.
Kronologic is still a logical deduction game, but the premise is wrapped in a small mystery scenario. In the puzzle I played, a detective has been poisoned, and the goal is to figure out exactly when, where, and by whom the crime occurred. The structure of the mystery revolves around an opera house made up of several rooms, a handful of characters constantly moving through those rooms, and a sequence of time steps representing the evening’s events. The key piece of logic driving the puzzle is that the detective was poisoned when he was alone in a room with someone else. Your job is to track where everyone was, when they were there, and eventually determine which character found themselves alone with the detective.
To uncover this information, the game uses another clever punch-board system reminiscent of Turing Machine, though thankfully, this system is much simpler to wrap your head around. When you want to investigate something, you pick a room, and then combine the room with either a character, and a time period, then place the corresponding punch cards together before flipping them over. The overlapping holes reveal information about that combination, giving you clues about where someone was or wasn’t during a specific moment. What’s particularly interesting is that when you ask a question in a multiplayer game, some of the information you uncover is private, while some of it must be shared openly with the table. Everyone gets a small piece of the puzzle, but you might receive a slightly more precise insight that helps you get ahead of the pack.
I really enjoyed playing Kronologic. The rules are straightforward, the setup is small, and the puzzles themselves manage to create that satisfying feeling of deduction throughout the entire experience. Because every character must move to a different room at each time step, the logic starts to unfold in interesting ways. Once you know where someone definitely was at one point, it constrains where they could possibly be later, and slowly the possibilities begin to collapse in on themselves. If you remove the impossible, whatever remains, however improbable, must be the truth. There’s a good sense of progression as you start to piece together the timeline of the evening, marking down your deductions and watching the puzzle reveal itself in front of you.
The moment when you think you’ve cracked the case, when you’re ready to declare exactly when and where the poisoning occurred and who was responsible, is genuinely exciting. That feeling mostly comes from the fact that you arrived at your conclusion before everyone else did, and so you can rub it in their faces. Sometimes that happens because you simply asked the right question at the right time. Maybe you just happened to pick the one character who actually mattered and asked about them early, which then gave you just enough information to unravel the rest of the puzzle. Meanwhile, your opponents might have spent their early turns investigating characters who ultimately had nothing to do with the crime. In that sense there is definitely some luck involved, particularly in the opening turns when you’re still feeling around in the dark for a useful thread to follow.
That early randomness can make some of the choices feel somewhat arbitrary at first. You’re often picking a character simply because you don’t like their vibe, because they seem as good a place to start as any. But I suppose that’s also thematically appropriate. A good mystery always has its share of red herrings, and part of the detective’s job is to follow the logic rather than their instincts. Someone might give off suspicious vibes, but that doesn’t necessarily mean they committed the crime.
What I appreciate most about Kronologic is how directly it addresses the issues I had with Turing Machine. The setup is dramatically faster; eighteen cards on the table and a sheet of paper to start scribbling your deductions on, then you’re ready to go. Kronologic also avoids that solitary puzzle feeling, because every question you ask leaks a little bit of information to your opponents. Even when someone else is taking their turn, you’re paying attention to what they’re investigating and what that reveals. That small layer of shared information creates just enough interaction to make the experience feel communal rather than parallel.
My biggest concern with Kronologic has to do with its long-term replayability. And maybe it’s a little rich for me to worry about long term replayability as I tend to play most games a mere half a dozen times before moving on to the next new shiny thing, but the difference between Kronologic and Turing Machine is stark. Turing Machine boasts over 7 millions puzzles, meaning you could do 5 puzzles a day for 3,835 tears before running out. Kronologic, on the other hand, comes with three main puzzles, and each puzzle has six variations, or difficulty levels. Once you’ve solved them, that’s kind of it. You know the answers. Unless you’re like me and manage to forget the details after a few months, those particular mysteries aren’t going to surprise you again.
That said, there are already two other Kronologic boxes available using the same system (Cuzco 1450 and Babylon 2500), which I’m glad to see. I also saw some PNP puzzles on Kronologic: Paris 1920‘s BGG page. I genuinely enjoy the physicality of the deduction here, and I’m curious to see how the other scenarios differ from the Paris 1920 setting I played.
In the end, while I still think Turing Machine is the more astonishing production, the cooler toy to hold in your hands, Kronologic is the better game. It captures that classic deductive thrill of piecing together a mystery while adding just enough shared information to keep everyone engaged in the same puzzle. If you have any fondness for old-school mystery games like Clue, Kronologic gives off very similar vibes, but in a way that feels fresher and more satisfying to unravel.
It’s a question I’ve been mulling over for years, and one that tends to pop into my head whenever I’m browsing an upcoming release and trying to get a sense of what people are thinking. I scroll past the preview images, maybe skim a few comments, and then my eyes drift over to the rating… only to see that bar graph with a giant foot, the 1 ratings outnumbering every other number by a large margin. Also, why the heck are there ratings on this game if it isn’t even out yet? These 1s aren’t low scores from disappointed players, they aren’t thoughtful critiques explaining why something didn’t land. These 1s feel more of a punishment than anything else. And I always find myself wondering: what is that number actually trying to say?
Because, at least in my mind, a rating is supposed to represent an experience. It’s meant to capture what it felt like to sit down, learn the rules, fumble through a first play, and how much joy someone had during their play. But when a number gets assigned before any of that has happened, it starts to mean something else entirely. It’s less about the game, and more about the drama surrounding the production, or perhaps one of the people involved.
Two recent examples of unreleased games with a large number of 1 ratings
The Things We’re Really Rating
In my experience, a lot of these early 1s don’t come out of nowhere. They’re reactions to decisions made long before a game ever reaches players. Recently, AI-generated art has been a flashpoint. People see something that feels off, or read a comment suggesting shortcuts were taken, and suddenly the rating becomes a place to push back. The most recent case of this was Concordia Special Edition by Awakened Realms. The cover looked a little Ai generated, people reacted, and Awakened Realms responded by saying “No AI art will be in the final game“
Sidebar: I’m surprised people continue to be surprised every time Awakened Realms uses AI images, considering Awakened Realms used AI art in their pre-production images many of their projects, including the special editions of Agricola, Puerto Rico, and more. They always publicly state that there will be “no AI art in the final products”, but it seems like every time they release a new product, there’s a new backlash over their continued use of AI promotional images.
No AI in my copy of Agricola
But honestly, I’m glad that people are willing to raise a stink over AI images. I don’t have the patience for it and I end up silently voting with my wallet instead of grandstanding on social media. But without vocal pushback, how is a company supposed to know what they’re doing is wrong? That said, I do dislike when those concerns get funnelled into a single number on Board Game Geek, especially in a context where a rating is for an entire game. It feels like it distorts the purpose of that number and platform.
What makes it even harder to untangle is when a game is getting slammed for multiple reasons. Some people give it a 1 for using AI art, others give it a 1 for being too expensive, packed with unnecessary deluxe components and premium materials. All of these concerns are valid, but is it worth dragging the entire production through the mud for it? Does Concordia designer Mac Gerdts get mud on his face by association because a publisher made the choice to use AI artwork for a promotional cover?
When Numbers Stop Meaning What We Think They Mean
The tricky part is that once ratings start being used this way, the meaning of the numbers begins to shift. A 1 no longer necessarily means “this game is terrible to play.” or “It’s utterly broken”, like in the recent case of RoboRover 2077. It now might just mean “I disagree with how this was made,” or “I don’t like what this represents,” or even “I’m frustrated with the publisher.”
And to be clear, those feelings aren’t inherently wrong. People engage with games for all kinds of reasons, and the hobby doesn’t exist in a vacuum. Themes matter. Production choices matter. The broader industry matters. But when all of those things get compressed into a single score, it becomes harder to extract useful information, especially for someone who just wants to know: is this a good game to play?
Sometimes just looking at a game cover or back of box picture won’t let you know if a game is for you or not
That’s ultimately why I look at ratings in the first place. Not as a verdict, but as a rough barometer. Sometimes I’ll be standing in my friendly local game store and I’ll pick up a box I hadn’t heard of. A quick search on BGG will sometimes tell me that a game might be a diamond in the rough, or, that a game isn’t really worth a second look. When that signal gets overwhelmed by reactions that aren’t rooted in gameplay, it becomes harder to trust what I’m seeing.
And people weaponizing their 1 ratings can go a step too far. The brigading between fans of Gloomhaven and Brass: Birmingham didn’t just stay in comment threads, it spilled directly into ratings, with people boosting one and tanking the other in a kind of ongoing tug-of-war. At that point, the numbers stop reflecting experience altogether and start reflecting the zealotry of the fanbase.
Can You Even Fix This?
Whenever I feel dissatisfaction with a system, my brain always shifts to trying to figure out a solution, even when I am powerless to make changes. I know BGG does take action against review bombing, and it can be challenging sifting out the actions of bad actors vs the legitimist grievances. But beyond that, I can’t help but wonder if ratings should be weighted differently? Should people who log plays have more influence than those who don’t? Could the system identify and limit users who consistently “review bomb” games before release? Should there be a separate rating for ‘verified’ reviewers, like Rotten Tomatoes has for movies?
But the moment you start going down that road, you run into a different kind of problem. Not a technical one, but a philosophical one. Who gets to have a voice?
Because there are infinite edge cases that don’t fit neatly into these solutions. What if you’ve played one edition of a game and want to rate another? What if you have strong objections to a game’s theme or message? Should those perspectives be excluded entirely just because they’re not tied to logged plays?
There’s also the simple reality that any system designed to police behaviour will eventually be gamed. If ratings required comments, people would leave empty ones. If they required play logs, people would log plays they didn’t have. At a certain point, you’re not solving the problem, you’re just moving it around.
Maybe the System Isn’t the Only Issue
Another idea that comes up fairly often is whether the rating system itself is part of the problem. A single number is a blunt tool. It tries to capture too many things at once: gameplay, components, art, rules clarity, personal taste, and compress everything into a single data point.
Would it be worth breaking the rating system apart? Would a system where you rate different aspects separately: gameplay, components, art, rules, overall experience. A composite score could still exist, but it would be built from multiple perspectives rather than a single gut reaction. Maybe that would make it harder to use the system as a blunt instrument.
Or maybe it would just give people more places to express the same frustrations.
A simpler, more immediate change might be to restrict ratings before release. Let previews be previews. Let early impressions live in comments and reviews. And let ratings reflect actual time spent with the game as it’s intended to be played. It wouldn’t solve everything, but it might curb the knee-jerk reactions to pre-production decisions.
Sometimes I wonder just how many people are turned away from Bullet because of the anime artwork.
Or Maybe This Is Just Who We Are
There’s a part of me that keeps coming back to a less satisfying answer: maybe this isn’t a systems problem. In my previous job as a Systems Administrator, I used to tell managers all the time “IT are really bad managers.” It’s not about building a system resilient to abuse, but it’s about how people choose to engage with the systems.
Some people will always use ratings as a way to express frustration, or to push back against trends they don’t like, or to support the things they care about. Others will treat them as carefully considered reflections of their experiences. Hell, I read one account of someone who used the rating system as a reminder of how many times he played each game in his collection (so a game he played twice got a 2, etc)
No system can really account for every use case that the public will invent.
Where I Personally Land
For my part, I don’t include ratings in my reviews. And even when I do rate games on BGG, I try to be mindful of what that number represents. Not just how I felt in the moment, but how the game held up over time, how it played across multiple sessions, how much joy it brought me each time it hit the table.
I also tend to avoid the extremes. A 1 or a 10 should mean something, at least to me. They’re not just expressions of dislike or hype, they’re markers of something truly exceptional, whether good or bad. Most games I play fall somewhere between a 6 and a 9, and I’m perfectly comfortable with that. Any game that would hit those lower scores get weeded out before they even hit my table, so they don’t even get a chance to get a score.
But more than anything, I find myself relying less on the number on the BGG page or it’s placement in the overall top games list, and more on the opinions of people I trust. The written reviews, the YouTube Videos, and posts people share after actually playing the game. That’s where I find the real value.
Because at the end of the day, I don’t think the BGG top games list is a objectively correct measure of the quality of a game, but it does serve as a barometer for me. And the more people use the ratings to talk about everything, the less value the BGG ratings has for me.
I’ve always had a soft spot for puzzle games. From Tetris to those logic puzzles you find in the Penny Press game books, to word games and Sudoku puzzles. I love the moment when I sit down in front of one, utterly clueless, then start teasing at the edges, working the system to slowly unravel the answer.
That’s the feeling I had the first time I encountered Turing Machine, designed by Fabien Gridel and Yoann Levet, with art by Sébastien Bizos and published by Scorpion Masqué in 2022. Turing Machine is a deduction puzzle for one to four players where the goal is to determine a secret three-number code. In theory, it sounds simple enough: Each number has a colour (blue, yellow, and purple), and each one is between one and five. You’re just trying to deduce the correct combination. But the way the game gets you there is what makes it fascinating.
In the centre of the table are a number of “verifiers,” small logical rules that help guide you toward the solution. Each verifier has a large confirmation card associated with it, and these confirmation cards combined with the punch board numbers act like a kind of cardboard computer. During a round, each player chooses a potential code consisting of three numbers between one and five, one number for blue, one for yellow, and one for purple. You take the punch cards corresponding to those numbers and stack them together, lining them up so that all the cut-out holes overlap. Once the cards are stacked, only a single square remains visible. You then take that assembled code and test it against one of the verifiers by placing the large confirmation card underneath. It will reveal either a check mark or an X, telling you whether your code satisfies the condition being tested or not. It’s a simple action mechanically, but the first time you a little green check mark, it’s a little startling, like watching a mechanical calculator click and clack to arrive at the right answer.
There are a lot of these logical verifiers in the box, forty-eight in total, but you only use four to six of them in any given puzzle. Each verifier tests a different logical condition, and collectively they provide all the information you need to narrow down the possible solutions. One example, a verifier might test the value of the yellow number in relation to three. In that case, the rule could be one of three possibilities: the yellow number is less than three, the yellow number is equal to three, or the yellow number is greater than three. If you test a code where yellow is one and the verifier returns a check mark, then you immediately know that the verifier rule must be “yellow is less than three.” It doesn’t tell you the exact number, you still don’t know whether yellow is one or two, but it eliminates several possibilities for the yellow number at once. The puzzle, then, becomes a process of gathering small pieces of information from multiple verifiers and slowly collating them until the three-number code reveals itself.
What makes Turing Machine particularly impressive is the sheer scale of what it’s capable of generating. The game’s website boasts over seven million possible puzzles, and you can go there at any time to generate a daily challenge or create puzzles of varying difficulty. The site will simply give you the verifiers you need and the corresponding answer cards, and from there you can assemble the puzzle on your table and start deducing. Easy puzzles might use four verifiers, while harder ones ramp up to five or six, each additional rule adding another layer of complexity to untangle or another step in the logic you’ll need to take to deduce the correct 3 numbers. It’s a clever system that keeps the game feeling fresh long after you’ve understood its basic structure.
I really have to say how much I appreciate the math and artistry that went into designing this thing. The fact that this little stack of punch cards and a set of cardboard verification strips can function like a logical computer boggles my mind a little. You lay these punch cards on top of each other, isolating a single square, and somehow that physical arrangement accurately reflects the numbers you’re testing against the rule. It’s the kind of design that feels elegant in a way that’s hard to fully articulate. There’s no denying how brilliance it is, and every time I hold an answer card up to my stack of punch cards, I find myself admiring how such a simple set of components can produce such a robust deduction system.
Where the experience starts to lose me, though, is in how the game handles multiplayer. On paper, Turing Machine supports up to four players, but the structure of the game makes that feel a little misleading. In each round, every player selects their own three-number code and can test it against up to three verifiers. After that, everyone gives either a thumbs up or a thumbs down. Thumbs down if you want to keep gathering information, thumbs up if you believe you’ve solved the code. If everyone gives a thumbs down, the game simply continues into another round where all players test new codes. When one or more players signal that they think they’ve solved it, they can attempt to validate their answer against the solution. If they’re correct, they win; if they’re wrong, they’re eliminated from the game. If multiple players solve it in the same round, the winner is whoever asked the fewest verification questions in total.
Mechanically it works fine, but the effect is that the game feels extremely solitary. Yes, technically you’re racing the other players to reach the answer first, but the reality is that everyone is just solving their own puzzle in parallel. What your opponents do doesn’t really give you any additional information, and there’s no meaningful way to interact with their deductions or build on their discoveries. You’re not debating theories, you’re not negotiating clues, and you’re not influencing each other’s decisions in any meaningful way. At the table it ends up feeling less like a shared experience and more like several people sitting beside each other working through separate logic puzzles. At that point it almost feels like you might as well be doing Sudoku puzzle side by side and just comparing how long it took for each of you to finish.
None of this changes the fact that I genuinely admire what Turing Machine accomplishes as a design. As a solo puzzle, it’s pretty cool. There’s something really calm and satisfying about sitting down with a cup of coffee in the morning and working through one of its logical challenges, kind of like my mom used to do with her Penny Press puzzle books. The system and production itself is clever enough that the act of solving the puzzle becomes its own reward. It’s satisfying to look at all the verifiers and puzzle out the most efficient way to whittle down the potential options. The only practical drawback to Turing Machine is the physical setup. Pulling out the correct confirmation cards from a large stack, arranging six verifiers, and then carefully putting everything back in order afterward can feel a little tedious, especially if you’re planning to play multiple puzzles in a row. It’s not a dealbreaker, but it does add a small amount of friction to what is otherwise a very clean experience.
If you happen to be a premium subscriber to Board Game Arena, Turing Machine is available there in digital form. In that environment the game shines a little brighter, simply because all of that setup and sorting disappears. The system runs smoothly, the puzzles generate instantly, and you can focus entirely on the logic of the challenge rather than the logistics of the components. But in that case the magical moment of assembling the punch cards and having it reveal the answers is lost.
For me personally, though, I don’t see Turing Machine returning to my physical table very often. As a multiplayer experience it doesn’t give me the sense of interaction or shared triumph that I’m usually looking for when I sit down to play with friends. And while I appreciate the elegance of the puzzle, and the brilliance of the production, I’m not particularly drawn to playing it solo either with its tedious set up. What I’m left with, then, is a strong admiration for the wit and craftsmanship behind the design. It’s an incredibly cool system, one that manages to simulate a logical computer using nothing more than punched-out cardboard and a handful of clever rules, and perhaps a game I’ll spin up in my web browser while I’m eating my lunch. Even if Turing Machine is not a game I’ll reach for regularly, I can’t help but marvel what the designers managed to achieve with some piles of card stock.
The longer I’ve been into board gaming, the more I’ve noticed a steady drift toward hybridization. Mechanisms get smashed together, so now a game isn’t just a deck-builder game or a worker placement game, but some intricate fusion of both, a la Lost Ruins of Arnak or Dune: Imperium. And while I genuinely enjoy seeing the interesting ways games meld mechanics, there’s something refreshing about a game that picks a single idea and simply executes it as cleanly and completely as possible. So with that in mind, let’s talk about 1997’s For Sale, designed by Stefan Dorra.
For Sale is basically 2 auction games in one. Your performance in the first auction game directly sets you up for the second one, which is the half of the game that’s actually worth points. In the first half of For Sale, players are bidding on properties represented by cards numbered from 1 to 30, with each number reflecting not just value but a property with personality. The 1 is a broken cardboard box, getting soggy in the street while the 30 is a literal space station. Everyone starts with the same pool of money, and each round a number of properties equal to the player count is revealed. From there, players take turns bidding to stay in the round, raising the amount of cash incrementally or dropping out entirely, at which point they take the lowest valued property still available. If the bidding has looped around the table, then the player who passed forfeits half their bid to the supply in return for the lowest property card available. Only the final remaining player in a round surrenders all their cash and takes the final and highest property for himself. It’s a system that’s easy to explain, but it quickly shows that it’s holding a surprising amount of tension once you’re in it.
That tension comes from the constant push and pull between risk and reward. If a spread of cards includes one terrible property and several excellent ones, the bidding naturally escalates as players try to avoid being the one stuck with the worst option. But the moment someone chooses to drops out, it often triggers a chain reaction, as everyone reassesses the value of staying in versus cutting their losses. That rule about keeping half your money if you bow out is so clever. It creates a question in the players heads, maybe they’re bidding a bit high with the expectation that someone will bid over them, and by the time the round comes back to them, the current lowest card will be gone. Not every bid is going to make it to the final result, but it creates a dance of wills. A game of chicken where players are constantly reevaluating how much they’re willing to risk and how much they’ll drop to take the lowest card at the table.
Once all the properties have been claimed, the game shifts into its second phase, and this is where For Sale reveals its second auction type. Any excess money you have is put aside, and the properties you bought are now what you’ll use to bid with. Just like the first half, a number of cards equal to the number of players is revealed, but this time the cards represent sale values. Instead of a bidding system that goes around the table, with the value slowly swelling, now it’s a simultaneous bind bid. Everyone puts one card face down and simultaneously reveal. The highest number property takes the highest value sale price, and so on down the line. Suddenly all the decisions you made in the first half come back to haunt you. Who thought it would be a good idea to have the 16, 17, and 18? Why is your highest card a 23? Your pragmatic nature has left you with a string of low value houses and a pocket full of change.
What I find particularly compelling here is how differently the two phases feel, despite being so tightly connected. The first is open, conversational, and reactive. You can smack talk your opponents and change your mind halfway through a bidding phase. The second half is quiet and psychological, all the tension is built around hidden information and the simultaneous reveal. You’re not just evaluating the raw value of your cards, you’re considering their value in relation to your opponents. If you can correctly read your opponents, and snake a high value sale for one of your weaker cards, you’ll be in a great position. Or, if you’re like me, you’ll constantly play a card that’s a single digit below your opponents, costing you 5 or 6 thousand dollars in final score.
For Sale is not a game I would ever claim to be particularly good at. Valuing properties, both in terms of how much to spend in the first half and when to deploy them in the second, is a skill that feels just out of reach for me. I can see the logic, I can follow the flow of the game, but there’s an intuition at play that I just haven’t quite developed.
There are some things that become more noticeable the more you play. Turn order, particularly in the first phase, can have a huge impact on how a round unfolds for a particular player. The player who wins an auction becomes the starting player for the next round, which creates a shifting dynamic where position can be either an advantage or a liability depending on the cards in play. Being the first to drop out of an auction will often trigger that cascade of passing players, but being stuck at the end of the turn order can mean facing a heavily inflated bid with little room to manoeuvre. You’re sometimes left choosing between overpaying for something mediocre or settling for the worst option available, neither of which feels particularly satisfying.
That interplay between luck and planning is always present. The distribution of cards, the order in which they appear, and your position relative to other players all shape the decisions you’re able to make. It’s not something you can fully control, and while the game is short enough that this randomness rarely overstays its welcome, it does mean that some rounds feel more dictated than directed. The key, as with many auction games, is learning how to navigate those moments, how to affect what you can and minimize your losses and capitalize on the opportunities your opponents let slip through their fingers.
Where For Sale really shines is in showing how much it can accomplish with so little. It’s fast, it’s easy to teach, and it consistently generates meaningful decisions. The dual-auction structure gives it a satisfying arc, turning what could have been a single-note experience into something with interesting texture and variation. Even when For Sale frustrates, it does so in a way that invites another play, another attempt to better understand its rhythms.
Sometimes, when you’ve been inundated with new and complex games, it feels refreshing to pull out a game from almost 30 years ago and revel in its simplicity. For Sale is a game that has stood the test of time, and sits among the greatest auction games out there. It gives you the same satisfying feelings from its auction mechanics that much larger and longer games struggle to provide. It’s the perfect game to keep in your bag and pull out anywhere you have a few friends and a few minutes to spare.
At this point, there are a lot of trick taking games in the world. It’s kind of comforting to pick up a new one, and already know most of the rules. “This is a trick taking game, but here’s the twist…” and then bam, you’re off to the races. Maybe you’re trying to avoid taking the prince suit in Rebel Princess, or there’s a whole flowchart of special characters that beat one another in Skull King. Either way, trick taking games can be counted on to be taught extremely quickly, which means you’ll go from opening the rulebook to actually playing the game in mere moments.
That familiarity is part of the genre’s appeal. Trick-taking games feel communal in a way few other mechanisms do. Everyone comes to the table with a shared vocabulary: follow suit, trump, void, lead. Because of that, designers can afford to get weird. They can bend expectations, twist assumptions, and trust that players will roll with it rather than get lost in the weeds. When a trick-taking game introduces a new hook, it often lands immediately because the foundation is already there.
Cat in the Box: Deluxe Edition, designed by Muneyuki Yokouchi and published by Bezier Games, is a trick-taking game where none of the cards have a suit until they’re “observed,” or played. Every card is black and white, so players must declare the card’s suit when it’s played. That single idea alone feels clever, but the real trick is that there are five of each card value in the deck, but only four suits in the game. So you really need to hope that no one is going to play the red 4 when you were counting on your 4 to be red, as that might just force you into a nasty paradox.
That tension between possibility and inevitability is where Cat in the Box really lives. At the start of a hand, everything feels wide open. Your cards could be anything, heck, they are everything. But as suits get claimed and the shared board fills up, the future begins to harden in uncomfortable ways. What felt like flexibility suddenly becomes constraint. You’re not just playing your hand anymore, you’re trying to weave in-between your opponents cards, trying to sneak in one last play before the proverbial door slams shut.
A paradox occurs when none of the cards in a player’s hand can legally be played. Thankfully there’s a dual layer board that you put your own coloured token onto whoever you play, a card that tracks all the cards that have been played so far. Also, having tokens connect on that board are what earn you points at the end of a round. I find the board that tracks the cards that have been played to be the most helpful thing. Granted, this would be a very difficult game to play without it, but as someone who struggles to remember which cards have already been played, I really appreciate its existence, to the point where I wish every trick taking game I play would have one.
And that board is not just functional, but it actively shapes how you think and visualize the game. Instead of relying entirely on memory, you’re constantly scanning the board, reading the patterns, and watching where other players are committing themselves. The board turns the abstract concept of “what numbers and suits are left” into a tangible and spatial arena. You can see the risk accumulating, sometimes literally clustering on one half of the board.
The concept of declaring your suit is one that’s tough to wrap your brain around in theory, but once you have the cards in your hand, and you start playing, it’s surprisingly natural. The trick really comes in knowing when to call yourself void in a suit to play the red trump suit, and how to maintain your own strategic tempo going forward. Sometimes a gambit pays off, sometimes the other colours fill up much faster than you were expecting, and before you know it, the only cards you have left have to be blue, and you told everyone you were out of blue 3 turns ago.
Those moments are equal parts satisfying and horrifying. When a plan comes together, it feels brilliant. When it collapses, it’s usually because of a decision you made much earlier, when the consequences weren’t yet obvious. Cat in the Box is very good at making you feel responsible for your own downfall. In other games I’d blame the bad hand of cards I was dealt, but here, I have no one to blame but myself.
Players earn points in 2 ways. Firstly, you earn one point per trick that you’ve won. Easy, straightforward. Unless you caused the paradox, then it’s -1 point for every trick you won. Whoops! The other way to earn points is via token adjacency on the main board. At the start of each round, after looking at your cards, you need to bid on the number of tricks you think you’re going to win. If you’re successful in your bid, you earn one point for every token in the largest group of adjacently connected tokens. Earning that you get to score those bonus points from token adjacency scoring is a huge benefit, and properly maximizing those points can easily swing the game on its own. What I like here is how the bidding doesn’t feel bolted on. It integrates naturally with the spatial puzzle on the board and gives players a clear incentive to take risks.
Making your bid and scoring your adjacent tokens can be a real boon, but it can be really tricky to accomplish, as the round ends immediately when someone triggers the paradox. This can be supremely frustrating for the other players too. If you managed to collect your tokens all together, but someone causes a paradox one turn before you’re able to win the last trick needed to satisfy your bid. It creates an exciting moment of tension. Speaking of tension, each hand has a really great arc, as cards get played, the options available to you quickly diminish. When everyone is holding only two or three cards left, it feels like a standoff. Whose going to be the one to fail, is the person who goes right before you going to take the last 3 spot?
Cat in the Box is a fantastic subversion of the trick taking mechanism that gets players excited. It’s novel, interesting, and strategic, which each play leaving you thinking about how you could have done better. The production by Bezier games is no slouch either. The dual layered board keeps all the tokens in the right spots, the player tokens themselves are brightly coloured, translucent, and screen printed to showcase a different science-y thing, which just makes this production extra charming.
The novel subversion of the trick taking mechanism is the most interesting part of Cat in the Box, which means players who don’t have a lot of experience with trick taking games won’t appreciate the whimsy the game is presenting. It’s for this reason that I wouldn’t recommend breaking it out amongst trick taking newbies. But for the groups that have a few different trick taking games under their belts, then Cat in the Box is a delightfully fun surprise.
