The first playable duel

This is the current browser prototype: a two-player creature duel with a CPU occupying the other seat. The goal is a game that is easy to learn, lasts about 10–15 minutes, and supports different deck strategies. Duration, fun, and balance still need human play. The numbers below are test values, not a finished design.

Start here

Reduce the opposing hero from 20 health to zero. Each player has a 30-card deck and space for five units. On your turn, spend energy to play cards and use ready units to attack. End your turn when you are finished.

Your energy grows automatically: one on your first turn, two on your second, up to six. It refills at the start of each turn. You do not need resource cards or a separate action budget.

Units wait until your next turn before attacking. Attacking exhausts the unit. Each can attack once per turn, choosing an enemy unit or the enemy hero. If the enemy has a Guard, attacks must target a Guard first. Guard restricts only unit attacks, never tactics or Arrival effects.

When units fight, both deal their power as damage at the same time. Damage stays between turns. A unit is defeated when its damage reaches its health. Attacking a hero deals damage to that hero without retaliation.

Setup and turn details

1. Choose a starter deck. Both decks contain 30 cards, with two copies of each included design. Shuffle each deck and randomly choose who goes first.
2. Draw five cards. Each player may replace any number once: set them aside, draw replacements, then shuffle the set-aside cards back into the deck.
3. Begin the first player's turn. Increase maximum energy by one, capped at six, refill it, and ready that player's units. Draw one card; the first player skips the draw on the first turn of the match.
4. Play affordable cards and attack with ready units in any order. Each action resolves completely. There are no response windows during the opposing player's turn.
5. End the turn. Temporary bonuses expire as stated on the card. If your hand exceeds eight cards, choose which cards to discard until eight remain. Then the other player begins their turn.

A unit enters exhausted and takes one of your five board spaces. A full board prevents playing another unit. You cannot voluntarily discard a unit to make space. An Arrival effect resolves when its unit enters; optional Arrival text may be skipped. A unit can still enter if its Arrival has no legal target; that effect does nothing.

Playing a tactic requires its printed energy and legal targets. It resolves, then goes to discard. Check targets before spending energy. Playing a card and attacking do not have additional action-marker costs. Readiness limits attacks; available energy limits card plays.

Guards protect the entire defending side from attacks: both the hero and non-Guard units. Multiple Guards let the attacker choose which to attack. Exhausted Guards still protect and fight back. Exhaustion does not reduce power. Healing removes damage, never increasing a unit's printed health; hero healing cannot exceed 20.

If you must draw with an empty deck, take one hero damage for each missing card. Check hero defeat after the current action fully resolves, and after start-of-turn or end-of-turn bookkeeping. At zero or less hero health, you lose; if both are at zero or less at the same check, the match is a draw. The starting card pool has no health modifiers, generated units, or additional timing triggers. A player can concede. There is no site score, fixed six-round finish, or secondary victory track.

Why begin here

The user's requirements do not imply a territory game. Magic and Pokémon demonstrate that a duel is a broad structure within which creature development, resources, combat, and card relationships can differ greatly. Their actual rules are compared in the research. This prototype is a familiar baseline for discovering our own worthwhile interactions; it does not yet claim a distinctive finished identity.

The early design questions are concrete: is it satisfying to trade units, when is attacking the hero preferable, can a defensive deck recover without dragging the game out, and do combinations reward planning? A small shared rule set should support those decisions. More systems need to justify the learning cost they add.

Two starter decks let us compare pressure against resilient play. The complete browser card data is the implementation's source of truth. The earlier site-game card file belongs to an alternative and is not used by this duel.

What to record while playing

First user feedback, 3 October 2026: after playing a sample round, the user described the duel as a mechanically reasonable start and asked for more action variety. No duration, deck choice, or result was supplied. This is encouraging qualitative feedback, not a balance or timing measurement.

The next card experiments to discuss are a choice between two effects, a unit ability that competes with attacking, and developing a creature over several turns. These are candidates rather than implemented rules. Keep the current duel available as a baseline while evaluating which choices improve play.

• Time from the first turn to the result, separating learning games from familiar play.
• Decisions that felt meaningful, automatic, or confusing.
• Turns where one player had no plausible way back into the match.
• Cards that were consistently exciting or consistently unused.
• Whether you wanted another match, and what you would change in your deck.

CPU results can reveal broken rules or repetitive behavior. They do not establish human matchup balance. First-player advantage, the opening draw rule, the energy curve, board size, and the starter lists remain open to change. The initial set target is approximately 100 cards; the smaller playable pool should earn expansion through play.