Review: Cat in the Box

Introduction & Overview

Specimens should be enclosed individually in a box-shaped container that prevents their contents from being observed from the outside. Except during the experiment, refrain from placing multiple specimens in the same space, and the subject’s fur color should not be observed. The furry subject should be a cat. When the fur color of a subject is observed by any method (such as visual inspection), the fur color, which should be unique, will trigger quantum convergence. Good luck during the experiment!

This is the premise behind Cat in the Box, designed by Muneyuki Yokouchi (Yokai Septet, Sweets Stacks). First published in 2020 by Ayatsurare Ningyoukan, it received a new edition in 2022 by Hobby Japan. The artwork is illustrated by Osamu Inoue (Click & Clack, Dungeon Busters).

It is not published in Spanish, so you will have to resort to importing it (the game is completely language-independent, except for the rulebook). It plays with 2 to 5 players, with a suggested minimum age of 13, and plays in about 20 to 40 minutes. The retail price is €29.95. For this review, a copy of the deluxe edition co-published by Hobby Japan and Bezier Games was used as a review copy.

Important: if you already know the game and/or are only interested in my opinion about it, you can go directly to the Opinion section. The Content and Mechanics sections are especially intended for those who do not know the game and prefer to get a general idea of how it works.

Components

Inside a two-piece cardboard box (lid and bottom), measuring 17.8 × 12.9 × 4 cm (a small rectangular box similar in size to Honshu or Fungi), we find the following components:

50 Cards (62.5 × 88 mm):
- 45 Cat Cards
- 4 Research Cards
- Starting Player Card
5 Player Boards
Research Board
60 Player Tokens (12 in each color)
4 Observation Tokens
Scorepad
Rulebook

How to Play

Cat in the Box is a seemingly traditional trick-taking game (numbered cards in four suits, players must follow suit, there is a trump suit, and the trick is won by the player who plays the highest card of the led suit or the highest trump card if unable to follow suit). The twist is that cards do not have a predetermined suit at the start; instead, players declare their suit as they play them. To track this, there is a central board featuring a space for each suit and value, where players place a marker on the matching slot to indicate that specific card has already been played. A player may choose to declare a void (not follow suit), losing the ability to play cards of that suit for the remainder of the hand. At the start of each hand, players bid on how many tricks they will win. The round can end if all cards are played (leaving no empty spaces on the board) or if a paradox occurs because a player cannot legally play a card (no available spaces match the remaining values and valid suits in their hand). The player who triggers a paradox loses one point per trick won and forfeits the bonus for matching their bid. The remaining players score one point per trick won, plus—if they hit their exact bid—one point for each token in their largest orthogonally connected group on the central board. The game lasts as many hands as there are players.

Key Concepts

Let’s begin with the Cards. The deck contains cards numbered from 1 to 9, with 5 copies of each value. The top and bottom borders show player count indicators dictating whether the card should be included in the deck. It is worth noting that all cards share the same neutral color, displaying small flashes of each suit color.

To determine the suit of a played card, players refer to the Research Board. This board holds cards displaying numerical values (with two setups available: matching values in columns or staggered rows). When playing a card, a player must mark on this board that the specific value for that suit has now been claimed. The top row corresponds to red, which acts as the trump suit and cannot be led until a player has previously broken suit.

Players use Research Tokens in their chosen player color to block off spaces corresponding to the played values in each suit. Additionally, players aim to connect their tokens orthogonally into large continuous groups to maximize their bonus score.

This bonus score depends on the bid placed at the start of the round, where each player predicts how many tricks they expect to win. For this purpose, each player has a Player Board with numbered spaces (from 1 to 3 for 4/5-player games, or 1, 3, and 4 for 2/3-player games). Furthermore, this board serves to track which suits a player is still allowed to play, as well as indicating the declared suit of a played card by placing it adjacent to the corresponding colored edge.

That is enough to get started.

Preparación de la Partida

Form the deck of cards, returning any cards not matching the player count back to the box.
Set up the research board by inserting the cards with either the aligned columns side (for 4/5 players) or the staggered rows side (for 2/3/4 players).
In a 4-player game, place observation tokens on the value 9 spaces. In a 2-player game, place them on the value 6 spaces.
Each player chooses a color and takes a player board along with 12 tokens.
Randomly select the starting player and hand them the starting player card.

We are ready to begin!

Game Flow

A game of Cat in the Box is played over as many hands as there are players in the game. Each hand consists of three phases.

Phase I: Setup

Proceed as follows:

Gather all cards, shuffle them, and deal them out equally among all players.
Each player reviews their hand and selects one card to discard (placed face down in a discard pile).
Players place 4 of their tokens onto the suit spaces of their player board.
Starting with the lead player and proceeding clockwise, each player predicts how many tricks they will win during the hand by placing a token on the corresponding number.

Phase II: Tricks

During this phase, players play an indeterminate number of tricks (up to a maximum of one fewer than the cards in hand at the start of this phase).

The lead player (starting player for the first trick) begins by placing a card from their hand adjacent to one side of their player board to declare its suit, then places one of their tokens on the matching value and suit on the research board. Naturally, a player cannot declare a suit if that exact value has already been claimed on the research board. Furthermore, leading with red (trump) is forbidden unless at least one token is already on the red row or the player has no other legal suits remaining.

All other players must play a card from their hand (following the same placement process), with two possibilities:

Follow Suit: The declared suit of the played card matches the suit led by the lead player.
Break Suit: The player declares they do not have cards of that suit by placing their card adjacent to a different color. Consequently, that player may no longer play cards of the led suit for the rest of the hand, immediately removing the corresponding suit token from their player board.

Once everyone has played a card, determine who won the trick. The winner collects the played cards and stacks them into a personal score pile.

The trick is won by whoever played the highest card of the led suit, unless one or more red (trump) cards were played, in which case the highest red card wins. The first trick where this happens will naturally occur because a player broke suit.

The winner of the trick becomes the lead player for the next trick.

This phase ends as soon as either of the following conditions is met:

All values have been played; that is, every slot on the research board is filled with tokens.
A paradox occurs. This happens when an active player cannot legally play any card from their hand because no empty spaces exist on the research board for their remaining card values within their available suits. When this happens, the trick-taking phase ends immediately and the current trick is discarded without a winner, though any tokens placed on the board during this final trick remain in place.

Phase III: Scoring

Each player scores:

1 point per trick won during the previous phase, unless they triggered the paradox, in which case they lose 1 point per trick won instead.
If the player correctly hit their exact trick bid and did not trigger a paradox, they score 1 point per token in their largest orthogonally connected group of tokens on the research board.

A new round then begins, passing the starting player card to the player sitting to the left of the current starting player.

Game End

The game ends after completing as many hands as there are players. The player with the most points wins. In the event of a tie, the player who scored the most points in the final round wins. If the tie persists, victory is shared.

Variants

2-Player Mode. In the setup phase, each player receives 10 cards. The remaining five cards form a draw pile from which 3 cards are revealed, placing a neutral marker on the research board starting from the green row and moving upward in the case of duplicate values. There is no trick bidding at 2 players; the connection bonus is awarded if a player wins 4 or fewer tricks.

Personal Opinion

Today we look at one of the games that could easily be considered a standout hit at the 2022 Essen Spiel fair. Not because it shattered sales records or introduced mechanically groundbreaking concepts (though it certainly has its quirks). While it was well-positioned on pre-fair hype lists, the real trigger for the frenzy was the news that the publishers had brought very few copies to the event. In fact, they capitalized on the massive demand by hiking the price substantially (nearly 50% above the pre-fair announcement).

Fortunately, thanks to Dam (La Mesa de Dam), I was able to secure a copy, making this long review possible. Let’s find out how Cat in the Box holds up on the table. At its core, we are dealing with a relatively traditional trick-taking game: a numbered deck of cards with a set of suits. The match consists of a predetermined number of hands. In each hand, players compete to win tricks.

The lead player determines the suit that everyone else must follow. If players cannot follow suit, they may play any other color, with a dominant trump suit available to compete for the trick even when voiding the led color. After the deal, players bid on how many tricks they expect to win during the current hand, earning extra points beyond the base value of winning tricks. Up to this point, everything feels familiar, echoing gems like the brilliant Skull King or Sluff Off!

The twist in this design becomes obvious the moment cards are dealt: they have no printed suit. A player might hold several identical black cards showing the exact same number. This is where the theme kicks in. To explain it, we need to touch lightly on quantum physics (don’t worry, we’ll keep it accessible and lighthearted without requiring an advanced degree).

In classical physics—the kind taught in high school—observing an object allows us to pinpoint its exact state. We know its position, and by calculating the forces acting upon it, we can project where it will end up after a given time. Yet shrink down to the atomic scale, and things change dramatically. We begin dealing with particles that cannot be easily measured without the act of observation altering the system itself.

This leads directly to Heisenberg’s famous uncertainty principle (the real physicist, not the one who became famous cooking blue crystal in New Mexico), which states that one cannot simultaneously know with precision both the position and the momentum of a particle. The more accurately one parameter is measured, the less certainty remains for the other.

Consequently, quantum mechanics introduces superposed states: since we cannot pinpoint the exact state and motion of a particle, it is treated as existing in multiple states simultaneously. At least until that particle is observed.

This premise naturally evokes the legendary Schrödinger’s Cat thought experiment, which posits a living cat sealed in an opaque box with a vial of poison tied to a quantum decay event. Until we observe the particle (open the box), the vial is simultaneously broken and intact, leaving the cat both alive and dead.

I went into this brief scientific detour because it highlights just how cleverly the game ties its theme into the mechanics—except instead of endangering cats with poison, we simply paint them. When the box opens, the cat turns out to be red, green, blue, or yellow. Until we open the box—that is, until we play the card—the cat exists in all four colors at once.

Applying this clever twist to a traditional trick-taking framework changes the dynamic entirely. For starters, assessing whether a hand is strong or weak is no longer straightforward, because cards only take on concrete values and suits as tricks unfold.

Higher-value cards naturally offer an edge when competing for tricks. However, there is never a guarantee you will be able to play them in the desired suit, either because opponents claimed those slots earlier on the board, or because declaring yourself void locked you out of that suit entirely.

In my view, this is the most innovative concept the game brings to the table. Players themselves decide when to follow or break suit, bearing the full strategic weight of that choice. Often, the decision hinges on how many slots remain available in a given color. If a suit row is nearly full and following suit offers no chance of winning, deliberately breaking suit to drop a trump card can be a great trade-off if taking the trick suits your plan.

Conversely, breaking suit early when many copies of a value remain unplayed can disrupt your opponents’ plans, stripping them of tricks they took for granted. In fact, play often turns cutthroat, such as deliberately throwing a trick to an opponent who did not want it, causing them to miss their exact bid.

Another design triumph that sets this title apart from standard trick-taking games is how it eliminates card counting. In traditional trick-takers, tracking played cards is essential to evaluate the strength of remaining hands. For example, once the highest card of a suit has been played, the next highest becomes the new powerhouse. In lengthy hands, remembering early plays can become a memory chore.

Because cards lack inherent suits, the central board tracks every play openly. This shared public information levels the playing field, letting players focus on tactical decisions rather than raw memorization.

It also cleanly resolves the issue of illegal reneging (failing to follow suit when holding a valid card). Spotting a renege usually requires an observant opponent with a sharp memory who recalls that a player claimed to be void earlier in the hand. Here, playing a color requires having an active token on that suit; the moment you break suit, that token is discarded, making illegal plays impossible.

Another brilliant touch is giving players a secondary spatial goal beyond merely taking tricks. Hitting your exact bid unlocks bonus points based on your largest connected group of tokens on the central board. As a result, winning just one trick with a sprawling cluster can yield far more points than sweeping four or five tricks without hitting your bid.

Finally, the end-of-round trigger provides a tense dynamic. While one might expect hands to end only when all cards are played, having an extra copy of each value means hands frequently collapse early when a player has no legal moves left on the board within their available suits.

Tying neatly back to the theme, this gridlock triggers a paradox: the player holds cats that cannot theoretically exist because all physical instances have already been observed. The player who causes the paradox not only forfeits their connection bonus, but also suffers a penalty equal to the number of tricks they took that hand.

This dynamic creates two distinct attack angles. You might target a player building a massive token territory before they lock in their bid, or push a player who hoarded tricks into causing a paradox for a massive point penalty. There is no single dominant strategy; greedily sweeping tricks carries heavy risk if the board locks up ahead of schedule.

As you can see, Cat in the Box remains rooted in classic trick-taking conventions, yet subtle shifts give it a distinct identity in an overcrowded genre. Recent trick-taking games that stood out usually did so by integrating foreign mechanisms, like Shamans (tocho-review here), or by tailoring themselves to unusual player counts, like Claim (tocho-review here) or The Fox in the Forest (tocho-review here).

Its primary drawback, shared with many trick-taking games, is scaling. Cat in the Box shines brightest at four or five players. At three, the experience thins out as the delightful chaos of higher counts fades. At two players, the formula simply falls flat.

Replayability is high, provided you enjoy trick-taking games. These games are famously addictive, encouraging immediate rematches where whole afternoons slip away. Furthermore, the fresh quantum hook will keep fans of the genre eager to explore its depth.

Regarding production quality, the results are mixed. Cards are the core component, and here they offer decent quality, featuring a good core weight, nice snap, and linen finish. That said, sleeving them is strongly advised. On the plus side, the translucent resin tokens look great on the table, though selecting light blue alongside dark blue, and red alongside pink, makes them harder to distinguish at a glance (thankfully mitigated by distinct engraved symbols). On the downside, the research board feels rather flimsy, using thin, poorly pressed cardboard despite the recessed slot design. The rulebook is clearly organized, easy to follow, and supported by helpful examples.

Visually, the game is delightful. Unique art for every card value portrays charming cats in whimsical, space-floating poses. The colored speckles over monochrome line art look wonderful, complemented by an eye-catching box cover featuring a rainbow-outlined cat peeking through the lid.

And let’s wrap this up. Cat in the Box stands out as one of the best surprises of 2022. Without reinventing the wheel mechanically, it delivers a fresh personality, addressing classic trick-taking pain points like card counting and reneging. The core idea of suitless cards declared during play is superb, completely reshaping how you evaluate each turn. Moreover, the quantum theme is brilliantly integrated. While it falters at lower player counts, it is an absolute marvel at its ideal numbers. For all these reasons, it earns a…