Game Mechanics

---
name: game-mechanics
description: >
  The gamification toolbox with judgment: core loops, progression, rewards,
  social mechanics — what each one does, whom it serves, and its shadow side.
  Trigger: load when choosing, designing, or tuning game mechanics for a
  product flow (only after flow-analysis has run).
license: Apache-2.0
metadata:
  author: jrodrigopuca
  version: "1.1"
---

## When to Use

- Only after [flow-analysis](../flow-analysis/SKILL.md) has run — a flow that
  isn't understood and simplified yet is not ready for mechanics
- Evaluating an existing mechanic that isn't working ("nobody uses our badges")
- A stakeholder asks for "points" or "a leaderboard" and you need to redirect
  to the underlying motivation problem

## Critical Patterns

### 1. Start from the motivation gap, not from the mechanic

Diagnose WHICH motivation is missing at the weak point of the flow — and for
WHOM: judgment #5's player types (achievers, explorers, socializers,
competitors) turn the same gap into different mechanics, so name your users'
mix before reading the table. Then pick the mechanic that feeds the gap:

| Missing | The user feels | Mechanics that feed it |
|---------|---------------|------------------------|
| Competence | "I'm not getting better / I can't tell" | Progress bars, levels, streaks-as-mastery, skill trees, "you improved X" |
| Autonomy | "I'm following orders" | Meaningful choices, multiple paths, customization, self-set goals |
| Relatedness | "I'm alone here" | Teams, gifting, shared goals, communities, co-op challenges |
| Purpose | "Why does this matter?" | Narrative framing, impact meters ("you saved N hours"), collective milestones |

Purpose sits over the three the body names (autonomy, competence,
relatedness): it is what makes the other three worth having. A mechanic
chosen before the diagnosis is decoration. "Add points" is never the answer
to a question you haven't asked.

### 2. Design the core loop first — it must close in seconds

This is the IN-SESSION loop, the one measured in seconds: **action → immediate
feedback → reward / progress → invitation to the next action**. Before adding
meta-systems (levels, seasons, collections), verify the basic loop is tight:
does every user action get an immediate, satisfying response? Is the next
action obvious and one step away? A juicy loop with zero meta beats a dead
loop with five meta-layers. The other loop — trigger → action → reward →
investment, measured in days — is the habit loop and lives in
[engagement](../engagement/SKILL.md) #3; when the complaint is "they don't come
back", that is the one to open, and only after this one closes.

### 3. Progression: fast early, meaningful always

- First levels come fast (the first "level up" within minutes, not weeks) —
  early progress is what converts a visitor into a player.
- Use endowed progress: a bar that starts at 20% ("account created ✓") gets
  completed far more than one at 0% — never show a user an empty bar for work
  they've partially done.
- Progress must map to REAL capability or value ("you unlocked X", "your
  profile now does Y"), not to arbitrary numbers. XP that buys nothing reads
  as homework within a month.
- Endgame matters: what does level 50 do when there's no level 51? Mastery
  displays, mentoring roles, and prestige resets outlast content treadmills.

### 4. Rewards: schedule and meaning over size

- **Immediate beats big**: a small reward NOW shapes behavior more than a
  large one next month.
- **Earned beats given**: rewards tied to genuine accomplishment feed
  competence; participation trophies devalue the currency for everyone.
- **Surprise sparingly**: variable rewards are the most powerful and the most
  dangerous tool in the box — delight users with unexpected recognition, but
  building the CORE loop on slot-machine uncertainty crosses the ethics line
  ([judgment #6](../../AGENTS.md)).
- **Crowding out is judgment #3 in practice**: amplify intrinsic motivation
  with FEEDBACK; reserve extrinsic rewards for genuinely boring necessities.
- **If points buy anything, you have an economy**: name the faucets (how
  points enter), the sinks (what removes them) and what a point is redeemable
  for. XP that buys nothing reads as homework within a month; XP that buys too
  much inflates and the reward stops meaning anything. And where rewards
  involve chance or money, the ethics line has a legal edge — loot-box and
  minors rules vary by country, and "surprise sparingly" is not a defence.

### 5. Social mechanics: the strongest force, the sharpest edges

- Judgment #5's leaderboard rule, made operational: if used, cohort them
  (friends, league tiers, percentiles) so everyone has a winnable race —
  never one global table.
- Comparison mechanics punish beginners hardest, exactly the users you can
  least afford to lose. Compare users to their own past self first.
- Cooperation retains better than competition for most audiences: shared
  goals, teams, gifting. Competition is a spice, not the dish.

### 6. Every mechanic ships with its shadow acknowledged

Streaks build habit AND create anxiety — their forgiveness design lives in
[engagement](../engagement/SKILL.md) #4. Badges celebrate milestones AND
become clutter when they rain down.
Timers create urgency AND erode trust when fake. Collections drive completion
AND frustrate when items are gated by luck or money. Proposing a mechanic
without naming its shadow and its mitigation is selling, not analysis — the
same rule as an architect's tradeoff table.

## Resources

- Sibling skills: [flow-analysis](../flow-analysis/SKILL.md) (mandatory before
  this), [engagement](../engagement/SKILL.md) (onboarding and return loops
  these mechanics plug into)
- Tradeoff presentation format: [architect/tradeoffs](../../../architect/skills/tradeoffs/SKILL.md)
  — mechanics decisions are product decisions with reversal costs too