All Docs

Design Principles

Research-backed game design principles for making better decisions about Gambit's mechanics, balance, and player experience. Each section covers the theory, names the sources, and ends with concrete takeaways for our game.

1. The Decision Is the Game

"A game is a series of interesting decisions." -- Sid Meier, GDC 2012

A decision is interesting when it has trade-offs (no option is strictly best), situational context (the right choice depends on the board state), and persistent impact (it matters for more than one turn). If a player always picks the same option, the choice is decoration, not design.

Sid Meier identified five qualities of a good decision:

Equally important: decisions must be ambiguous. If the optimal play is obvious, the decision is dead weight. Soren Johnson (Civilization IV) warned: "Given the opportunity, players will optimize the fun out of a game." If one path dominates, variety vanishes. The designer's job is to make the optimal path and the fun path the same thing.

Pacing matters. Complex decisions back-to-back overwhelm. Simple decisions at a slow pace bore. The best games alternate between high-stakes pivots and lighter mechanical turns.

Gambit takeaways

2. Strategic vs. Tactical Depth

Strategic decisions are long-term: deck construction, mulligan choices, the overarching game plan. Tactical decisions are short-term and reactive: which unit attacks, whether to play removal now or hold it.

The best games layer both together. A taxonomy from design literature:

Decision Type Definition Card Game Example
Trade-off Choose between options with distinct pros/cons Play a unit vs. energize it
Dilemma Every option has a downside -- pick the least bad Spend your last energy on removal or a blocker?
Sacrifice Voluntarily give up something valuable Black's Sacrifice keyword; energizing your best card for ramp
Risk-reward Higher payoffs carry higher failure costs Attacking into unknown set cards

Key insight: "The key to whether a choice is compelling usually lies not in what we risk or sacrifice, but in what we need to take into account to make that decision." The more factors a player weighs, the richer the decision.

First Order Optimal (FOO) strategies are beginner-friendly approaches that remain surprisingly competitive despite ignoring core mechanics. If the simplest possible strategy dominates, the game's depth is illusory. The fix: tie theme to mechanics so the obvious strategy is also the interesting one.

Gambit takeaways

3. Resource Tension & Opportunity Cost

Resource systems serve three functions: power regulation (expensive cards can't come out turn 1), game pacing (early-game and late-game feel different), and design space (costs enable diverse card designs).

The fundamental tension in competitive card games is tempo vs. card advantage. Tempo means spending resources efficiently to affect the board now. Card advantage means investing in future resources. You almost never gain both simultaneously, so every turn forces a choice.

Magic has three core resources: cards, life, and tempo. Getting ahead on all three is nearly impossible, so players constantly evaluate which axis they're winning and which they're paying.

Any-Card-as-Resource Systems

Gambit uses an any-card-as-resource system (energize any card from hand). Research on similar systems (Duel Masters, Flesh and Blood):

Benefit Cost
Eliminates mana screw/flood entirely Reduces game variance
Every card has dual utility (play vs. energize) Turns can become predictable (always have X energy on turn X)
Creates meaningful sacrifice decisions If every card is equally good as fuel, the decision is trivial

Critical design requirement: the sacrifice must feel meaningful. Flesh and Blood solves this with red/yellow/blue pitch tiers -- cards have differentiated resource values. If every card is equally good as energy, the energize decision is automatic and uninteresting.

Gambit takeaways

4. Information Asymmetry & Hidden Information

Hidden information creates depth when the player's guess space has shape -- they can narrow possibilities through deduction, reads, and game knowledge. It creates frustration when outcomes arrive without interpretable context.

Three questions determine whether hidden info adds depth or randomness:

  1. What does the player know right now?
  2. What are they committing to when they act?
  3. What new information arrives immediately after?

When these answers are clear, players form testable hypotheses and improve through experience. When blurry, they blame randomness.

Input vs. Output Randomness

Players tolerate input randomness far more than output randomness because it offers mitigation options rather than negating decisions.

The Bluffing Layer

Face-down cards (set zone, traps) create a bluffing layer where incomplete information forces the opponent to respect possibilities rather than react to certainties. This transforms the game from a calculation exercise into a psychological contest.

Critical requirement: face-down cards only work if there's a critical mass of viable set-zone options. If only 2-3 traps are competitively playable, the opponent always knows what the face-down card is and the bluff collapses.

Gambit takeaways

5. Interaction & Counterplay

Counterplay has two components: counter (the opponent can respond in a way that improves their position) and play (the response involves interesting decisions for both players). Good counterplay creates a game-within-a-game. Bad counterplay is just pressing a "negate" button.

The Interaction Spectrum

Both extremes fail:

Extreme Problem Example
Too little interaction (solitaire) Both players execute plans without engaging each other; competitive element evaporates Combo decks that win without attacking
Too much interaction (permission/lockdown) One player can't execute any plan; agency is removed Draw-go counterspell decks; complete hand disruption

Mark Rosewater: "I'm not super fond of permission or discard decks that don't let you play." The core problem: if everything in your hand costs 3-4 energy and your opponent bounced your energy sources, you sit and wait. The player literally cannot make decisions, which violates every principle of engaging design.

Benchmarks for Fair Counterplay

From the Law of Game Design:

  1. What can the opponent do in response?
  2. Do those actions improve their position?
  3. How much do responses improve it?
  4. Do responses create interesting, layered situations?

If the answer to #1 is "nothing," the mechanic is broken regardless of how balanced it is numerically.

Gambit takeaways

6. Removal & Disruption Balance

The Answers vs. Questions Framework

Threats are "questions" the opponent must answer. Removal spells are "answers." Critical insight: threats are proactive and answers are reactive, and proactive strategies have an inherent structural advantage because they function even against unexpected matchups, while reactive strategies only work with the right answer at the right time.

When Removal Feels Healthy vs. Oppressive

The Worst Feeling in Card Games

Helplessness. Removal must exist to prevent degenerate board states, but it should never be so efficient that deploying threats feels pointless. The principle: removal should cost more than the threat it answers, or be conditional enough that it sometimes misses.

Gambit takeaways

7. Trap & Reaction Design

Lessons from Yu-Gi-Oh

Yu-Gi-Oh's trap cards have three inherent design problems:

  1. Set-and-wait tax. Dead in hand when going second. You're already behind and can't use them to claw back.
  2. Stiff activation requirements. If the opponent doesn't trigger the condition, the card does nothing.
  3. Reactive cards steal outs. Every trap in your deck is a card that isn't helping you break an established board.

Yu-Gi-Oh's solution was hand traps -- monster cards discarded from hand for instant utility. These have no set-and-wait requirement, work going second, and can't be read from face-down tells.

The "Do Nothing and React" Problem

When reactive strategies dominate, games devolve into both players passing turns with resources held open. Design solutions:

The Backrow Removal Question

Yu-Gi-Oh's history with backrow removal is a cautionary tale:

Gambit takeaways

8. Card Advantage & Tempo

Draw-to-Hand-Size Systems

Gambit uses draw-to-hand-size (refill to 5 at end of turn). Research on this system:

Aspect Draw-1 (MTG) Draw-to-Hand-Size (Gambit)
Individual card value High -- each card is precious Lower -- you refill every turn
Card draw effects Powerful deckbuilding decision Selection (see more this turn), not accumulation
Feel-bad potential Drawing dead cards is devastating Softened -- you always have options next turn
Tempo vs. advantage Distinct strategic axes Tempo dominates more; card advantage is about quality, not quantity

Key insight: when card draw is too cheap, card advantage stops mattering and only tempo matters. When draw is too expensive, games are decided by who draws best. The sweet spot: investing in draw is a meaningful deckbuilding and in-game decision.

Virtual Card Advantage

Beyond raw card count, virtual card advantage changes the value of cards:

Virtual card advantage is typically more powerful than raw card count.

Gambit takeaways

9. Meaningful Deckbuilding

Parasitic vs. Modular Mechanics (Rosewater)

The ideal: linear mechanics where cards reward a theme but find synergy partners across the full card pool. Build-around cards should be slightly worse in a vacuum but significantly better in context, making deckbuilding an exploration.

Why Formats Get Solved

Formats become stale when: (a) card pools are too large and optimal lists are found, (b) one strategy lacks sufficient counterplay, or (c) staple cards are so efficient that every deck converges on the same core.

A healthy metagame has 4-8 popular archetypes with natural rock-paper-scissors relationships. Every strategy has counters, but no counter wins 100% of the time.

Gambit takeaways

10. The Fun of Losing

The real problem: "When players feel like they have no chance of winning, they become disengaged." The design challenge is keeping the losing player making meaningful decisions.

Feedback Loops

The design tension: positive loops create satisfying engine-building, but risk runaway leaders. Negative loops keep games competitive, but can punish skilled play.

Richard Garfield on Variance

Garfield intentionally designs games where luck allows "the less-skilled player to win from time to time." The skilled player wins ~75% instead of 99%, keeping casual and competitive players at the same table. This isn't a concession to casual play -- it's a feature that makes every game feel alive.

The Iron Rule

"If a player so much as suspects that they could have avoided a bad turn of events, you have killed the drama." Losses must feel like the result of meaningful choices, not arbitrary punishment. Losing because the opponent made better decisions should feel instructive. Losing because you drew badly three turns in a row feels frustrating.

Gambit takeaways

11. Lenticular Design

Cards that appear simple to beginners but reveal strategic depth to experienced players. Rosewater's six rules:

  1. Some complexity is invisible to inexperienced players
  2. Cards must have obvious surface value
  3. Experience is connected to how far ahead a player thinks
  4. Novices don't think about causality chains
  5. Players try to use cards as their text suggests
  6. Let players play the game they want to play

Example: a unit with Endure is "a tough creature" to a beginner. To an expert, it's "a unit I can attack recklessly with because it survives the first kill, and I can pair it with healing to reset the Endure trigger."

New World Order: Complexity by Rarity

Cap complexity at common. Let uncommons and rares tick up. The cards players encounter first should be simple, with depth gated behind progression. This applies to a card game's initial set: the core pool should be learnable, with expansion sets adding layers.

Gambit takeaways

12. Core Principles Summary

The principles below are ordered by design priority. When two principles conflict, the higher-ranked one wins.

  1. Make the fun part the winning strategy. If optimal play is boring, the game is broken. (Rosewater Lesson 13)
  2. Every decision should involve opportunity cost. If there's no trade-off, there's no decision. (Meier)
  3. Protect player agency above all. The worst feeling is helplessness -- even the losing player should have meaningful choices. (Garfield)
  4. Answers should cost more than questions. Removal must exist but should never be so efficient that threats feel pointless.
  5. Hidden information needs shape. Players should narrow possibilities through skill, not guess blindly. Bounded possibility spaces create deduction; unbounded ones create frustration.
  6. Restrictions breed creativity. Color identity, resource limits, and set zone caps force diverse, interesting play. (Rosewater Lesson 18)
  7. Reactive play must cost tempo. If "do nothing and react" dominates, the proactive player's experience is destroyed.
  8. Design for passion, not universal approval. Mechanics that some players love are better than mechanics everyone finds merely okay. (Rosewater Lesson 11)
  9. Input randomness good, output randomness dangerous. Card draw before decisions creates problems to solve. Random resolution after decisions invalidates strategy.
  10. Players will optimize the fun out. Design so that the optimal path is also the interesting one. (Johnson)

Sources