A monster bears the number 37. How do you take it down?

Imagine a fight where your opponent has no health bar. It bears a number — say, 37 — emblazoned right in the middle of its body, and the only way to make it disappear is to reach that exact value with the cards in your hand. Not 36, not 38. Exactly 37. That's the principle behind Talystro, a roguelite deckbuilder — meaning a deck-building game where every run starts from scratch but leaves the player progressively stronger — developed by the small Norwegian studio Filiokus. Its free demo has been available on the game's Steam page since May 2026.
The idea is as simple to state as it is tricky to execute. And that's precisely where its appeal lies.

The Math Mouse versus the Necrodicer

The player takes on the role of a "Math Mouse" (the mathematician mouse, if you like), a small creature sent on a mission to thwart the plans of the Necrodicer — an antagonist whose ambition is to carry out Talystro's forbidden ritual. Around her, the enemies are literally numbers made flesh: according to Filiokus's official page, the monsters are "shaped as numbers," creatures whose very form is a number. The pitch borders on the absurd. And yet, once you get your hands on the game, the coherence becomes obvious.

Because everything else follows from this premise. If the enemy is a number, defeating it means solving it. Not hitting it hard, not stacking up damage over several turns. Solving it. In one stroke. The Steam page puts it unambiguously: "Match the exact value of your enemies to defeat them in one blow." Instant elimination, or nothing.

Cards as operations, dice as chaos

How do you build a 37? That's where the game system comes in. Each card in your deck doesn't represent a classic attack — it represents an arithmetic operation. An addition, a multiplication, a division, a modifier. Unfinished computations, with their variables and parameters, which the analysis site Watermelo describes as "unfinished calculations, variables and modifiers." You combine them in a precise order to arrive at the desired result.

But there's a surprise guest: dice. Every turn, they dynamically reshape the possibilities. A die can change the minimum or maximum value of an operation, alter the cost of a card, or trigger special effects. This is no longer just computation — it's computation under uncertainty. You have to anticipate probabilities, manage the order in which operations apply (yes, order of operations matters here), and adapt your plan in real time.
As the launch announcement carried by GamesPress sums it up: "Every encounter is a number puzzle where success depends on sequencing operations and reading probabilities." Every battle is a combinatorial optimisation problem dressed up as fantasy.

What the demo really contains

The demo isn't just a tutorial. It opens the doors to the Wildlands, the game's first act, with seven levels chained together and a final boss that tests every mechanic you've learned. Expect around 45 minutes for a full run — an ideal format for figuring out whether it hooks you or not.
Two difficulty modes are available from the start. Normal mode gives you time to grasp the system's logic. Expert mode, meanwhile, is a reminder that mathematics shows no mercy. Between attempts, new cards, artefacts and dice upgrades enrich your arsenal. The Steam page even mentions artefacts capable of creating particularly powerful combinations — what the developers call "game-breaking" builds. It remains to be seen whether these combinations are numerous and varied enough to sustain interest over the course of the full version, planned for later in 2026.
The structure openly draws on established references in the genre. Watermelo cites Slay the Spire and Balatro as points of comparison for the deckbuilding system and the logic of repeated runs. What sets Talystro apart is the replacement of the damage system with an exact-match system: where other games ask you to deal more damage than the enemy has hit points, here you have to hit the mark exactly. Not too much, not too little.

Maths without meaning to

The question naturally arises: is Talystro an educational tool disguised as a game, or a game that uses maths as thematic dressing? Filiokus's answer seems to be: neither, or rather both at once. The studio's official page positions the game as a roguelite deckbuilder centred on "the creative application of mathematics" — not on explicitly teaching it.
That's an important nuance. Talystro doesn't ask questions like "what is 6 × 7?" It puts you in a situation where, to survive, you must spontaneously work out how to combine your operations to reach a target value. Learning, if it happens, happens through immersion and necessity — not through drills. The player doesn't solve equations because they're told to. They solve them because otherwise, they lose.
This mechanism recalls what education researchers call problem-based learning: we retain better what we needed to compute to reach a concrete goal than what we memorised in the abstract. Talystro applies this principle without ever spelling it out — and that may be its greatest design achievement.
A player quoted in the GeekNPlay article sums up the experience with a line that sounds like a confession: "This is the first time I've had this much fun doing maths." This testimony proves nothing about the game's educational effectiveness — but it says something about its ability to disarm a well-known resistance.

A handcrafted visual identity

The game doesn't rely on its mechanics alone. The animations are hand-drawn, frame by frame, giving every movement an almost organic texture — a choice that deliberately contrasts with the cold abstraction of numbers. The soundtrack and full sound design accompany every successful operation, every dice roll, every elimination. The whole creates an aesthetic coherence you wouldn't necessarily expect from a game whose heart is arithmetic.

The demo is accessible with no time limit and no aggressive camera constraints, with an interface whose readability of values and operations has clearly been worked on. You can try it without pressure, make mistakes, start over. That, in short, is what every learning environment should be — whether academic or playful.

Concepts to take away

Defeating an enemy by solving it. In Talystro, the monsters are numbers. To eliminate them, you have to reach their exact value with your cards — not deal damage, not survive longer. Find the right equation. One shot, or nothing.The order of

operations really matters. Multiplying before adding doesn't give the same result as adding before multiplying. In Talystro, this basic mathematical rule can make the difference between victory and utter disaster.

Dice introduce randomness into the computation. It's not just pure computation: every turn, the dice change the available values. So you have to compute under uncertainty — a bit like estimating a tip when you don't know exactly what was ordered.

Learning without realising it is a real thing. Talystro doesn't ask you maths questions. It puts you in situations where you must do maths to move forward. The difference is huge — and it's exactly what educators call problem-based learning.

A Norwegian studio, a card game, and equations. Filiokus is a small independent studio. Talystro is their bet: to prove that mathematics can be the heart of a strategy game without anyone feeling like they're in class.

For maths enthusiasts

Talystro's central concept — reaching an exact target value from a set of operations and dice values — is formally close to a classic problem in theoretical computer science: the subset sum problem (SSP). In its general form, the problem is to determine whether some subset of a set of integers sums to a given target value. It is widely accepted that this problem is NP-complete — meaning that no efficient (polynomial-time) algorithm is known to solve it in the general case, although particular instances can be handled quickly.
Talystro isn't limited to sums: cards can represent multiplications, divisions, conditional modifiers, and the dice introduce a stochastic component. The question of whether a winning combination exists for a given game state is therefore even richer combinatorially. It's precisely this complexity that makes every turn non-trivial — and that explains why finding the "right" sequence of operations is a genuine reasoning problem, not a simple check.