← all games

7 Wonders Duel

#173346BGG ↗

2015 · 2-2 players · 30min · weight 2.23 · 108,493 ratings

v2 v3 fit 0.633

BGG raw

ID
173346
Name
7 Wonders Duel
Year
2015
Rank
16
Min players
2
Max players
2
Playing time
30
Min playtime
30
Max playtime
30
Avg weight
2.2271
Num weights
3413
Bayes avg
7.95318
Average
8.07782
Users rated
108493
Num owned
183581
Wanting
856
Wishing
12185
Num comments
14197
Fetched at
Sat Apr 25 2026 16:15:22 GMT+0000 (Coordinated Universal Time)
Mechanisms (15)
End Game BonusesIncomeMelding and SplayingModular BoardMulti-Use CardsOnce-Per-Game AbilitiesOpen DraftingScore-and-Reset GameSet CollectionSudden Death EndingTagsTech Trees / Tech TracksTrack MovementTug of WarVariable Set-up
Categories (5)
AncientCard GameCity BuildingCivilizationEconomic
Description (1739 chars)

In many ways 7 Wonders Duel resembles its parent game 7 Wonders. Over three ages, players acquire cards that provide resources or advance their military or scientific development in order to develop a civilization and complete wonders. What's different about 7 Wonders Duel is that, as the title suggests, the game is solely for two players. Players do not draft cards simultaneously from decks of cards, but from a display of face-down and face-up cards arranged at the start of a round. A player can take a card only if it's not covered by any others, so timing comes into play, as it can with bonus moves that allow the player to take a second card immediately. As in the original game, each acquired card can be built, discarded for coins, or used to construct a wonder. Each player also starts with four wonder cards, and the construction of a wonder provides its owner with a special ability. Only seven wonders can be built, though, so one player will end up short. Players can purchase resources at any time from the bank, or they can gain cards during the game that provide them with resources for future building; as they are acquired, the cost for those resources increases for the opponent, representing the owner's dominance in this area. You can win 7 Wonders Duel in one of three ways: each time you acquire a military card, you advance the military marker toward your opponent's capital (also giving you a bonus at certain positions). If you reach the opponent's capital, you win the game immediately. Or if you acquire six of seven different scientific symbols, you achieve scientific dominance and win immediately. If none of these situations occurs, then the player with the most points at the end of the game wins.

LLM v2 (wide)

Core verb
draft card from pyramid
Decision shape
combinatorial
Reward schedule
mixed:immediate+delayed
Aesthetics
["Challenge", "Discovery"]
Core loop pitch
Take a face-up card from a tiered pyramid (uncovering the row beneath); build it for resources, military, or science, or burn for coins/wonders.
Translation difficulty
Easy
Difficulty reason
Repos shipped a strong mobile app in 2019; the pyramid mechanic is naturally touch-first and 30-min sessions fit phone perfectly.
Direct digital port
7 Wonders Duel (iOS/Android, Repos Production)
Port kind
first_party
Closest loop translation
none yet
Primitive tags
["pyramid_card_reveal_drafting", "two_sudden_death_paths", "resource_cost_inflation", "wonder_picks_one_extra_turn", "science_set_collection"]
Confidence
0.9
Extracted at
Mon Apr 27 2026 11:40:03 GMT+0000 (Coordinated Universal Time)
Raw v2 JSON (940 chars)
{
  "game_id": 173346,
  "name": "7 Wonders Duel",
  "core_verb": "draft card from pyramid",
  "decision_shape": "combinatorial",
  "reward_schedule": "mixed:immediate+delayed",
  "aesthetics": [
    "Challenge",
    "Discovery"
  ],
  "core_loop_pitch": "Take a face-up card from a tiered pyramid (uncovering the row beneath); build it for resources, military, or science, or burn for coins/wonders.",
  "mobile_translation_difficulty": "Easy",
  "translation_difficulty_reason": "Repos shipped a strong mobile app in 2019; the pyramid mechanic is naturally touch-first and 30-min sessions fit phone perfectly.",
  "direct_digital_port": "7 Wonders Duel (iOS/Android, Repos Production)",
  "closest_loop_translation": "none yet",
  "primitive_tags": [
    "pyramid_card_reveal_drafting",
    "two_sudden_death_paths",
    "resource_cost_inflation",
    "wonder_picks_one_extra_turn",
    "science_set_collection"
  ],
  "confidence": 0.9
}

LLM v3 (deep)

Core verb (long)
You and one opponent alternate plucking cards from a shared pyramid of 20 cards per Age, where lower rows are face-up and upper rows face-down; only cards uncovered (not partially overlapped by another card) are takeable. The repeated action is the small ritual of scanning the tableau, weighing what your opponent will pick if you reveal it, then sliding a card out and either building it (paying resource costs from your tableau or coins), discarding it for 2+ coins, or burying it under one of your four wonder slots to construct a wonder. Reveals cascade: taking a card flips whatever was hiding under it, sometimes giving the opponent a free turn of perfect information.
Core loop (long)
Setup: deal three Age decks into pre-printed pyramid layouts (Age I: 4 rows, Age III: a chevron with bonus face-down rows). Each turn, the active player takes one accessible card and resolves it (build/discard/wonder); play continues until the Age's pyramid is empty, then military and progress tokens trigger, and the next Age deals. Three Ages, ~20 cards each. The game ends instantly via military supremacy (conflict pawn reaches enemy capital — at +9 toward you), scientific supremacy (collect 6 of 7 distinct science symbols), or, failing those, after Age III on victory points (civilian/commercial/wonder/guild scores summed). Wonders are limited to 7 total constructed across both players, so the seventh wonder built denies the eighth its slot — a famous 'one short' tension.
Decision space
Every pick is a four-axis decision: do I want this card's effect, do I want to deny it to my opponent (cards I can't afford are still draftable to discard), do I want to construct a wonder using it as fuel (some wonders give an extra turn or token-steal), and what does taking it reveal? The option space is small at any single decision (typically 2-6 accessible cards) but extremely deep because revealing a card hands the opponent free information. The hard part is threat assessment across three simultaneous victory paths: ignore military and you lose to a rush; ignore science and you lose to a 6-symbol sudden-death; ignore civilian VPs and you lose the points race. The pyramid layout means tempo (who moves next, who picks up double-turn tokens) is its own resource.
Skill expression
Top skills are opponent-threat reading (counting their science symbols and military shields every turn, calibrating when to switch from your engine into a denial-pick), tempo and sequencing (which card unlocks which downstream pick — and on which side of the table), and resource curve management (your buildings 'chain' so an early Stone Reserve pays for free for the rest of the game). Pattern recognition matters because progress tokens and guild scoring create combinatorial endgame value. Math is constant but small: shield counts, science pairs, coin arithmetic. Memory load is moderate — you must remember discarded cards because some wonders let you resurrect them.
Tactile dependency
low
Tactile reason
All information is open or hidden categorical card data; the pyramid's 'partially covered' state translates trivially to a digital UI (and Repos has shipped a polished first-party port via Asmodee Digital that does exactly this). No physical dexterity, no spatial-reasoning component beyond the abstract pyramid graph.
Promise
Build a Mediterranean civilization in 30 minutes against one rival — choose your path: army it down, out-science it, or out-build it on monuments and culture.
Setting
Historical, ancient world, Mediterranean civilization
Narrative
none — pasted-on theme. Mechanically it is a tight head-to-head card-draft engine; the Babylonian/Egyptian/Greek wonders are flavor wallpaper, evocative but non-narrative. The Pantheon and Agora expansions add slightly more thematic structure but it remains a pure strategy game.
Audience
gateway, hobbyist Eurogamer
Art direction
Miguel Coimbra's signature illustration: warm earth-and-gold palette, painterly architectural studies of ancient wonders, micro-detail medallions on every card. The aesthetic is romantic-orientalist 'old empires' — sand, marble, copper — recognizable across his 7 Wonders / Small World / Cyclades body of work. High production gloss; cards read crisply at 18 inches.
Meta-layer ideas
["Async PvP ladder with seasonal Wonder rotation: standard duel format, Glicko-style ranked ladder, each season introduces a new pool of wonders/progress tokens (already canonical with Pantheon/Agora expansions) \u2014 but a first-party port already exists, constraining independent translations to clearly differentiated takes", "Roguelite civilization gauntlet: solo PvE \u2014 face 7 escalating AI rivals each themed to a victory path (rusher, scientist, builder), draft persistent civic 'edicts' between matches that warp resource costs, season ends on a final boss duel", "Draft-and-build merge hybrid: take the pyramid-reveal mechanic but graft it onto a merge-2 city-builder where merging buildings triggers chain bonuses \u2014 distances enough from the canonical port to ship as its own product", "Daily duel puzzle: shared seed \u2014 same pyramid, same opponent AI plays \u2014 leaderboard by margin of victory; uses determinism of 7WD to create a real puzzle"]
Closest mobile genre
async PvP card battler / draft strategy
Live-service potential
medium
Confidence
0.9
Extracted at
Mon Apr 27 2026 10:41:58 GMT+0000 (Coordinated Universal Time)
Raw v3 JSON (5671 chars)
{
  "game_id": 173346,
  "name": "7 Wonders Duel",
  "mechanics": {
    "core_verb_long": "You and one opponent alternate plucking cards from a shared pyramid of 20 cards per Age, where lower rows are face-up and upper rows face-down; only cards uncovered (not partially overlapped by another card) are takeable. The repeated action is the small ritual of scanning the tableau, weighing what your opponent will pick if you reveal it, then sliding a card out and either building it (paying resource costs from your tableau or coins), discarding it for 2+ coins, or burying it under one of your four wonder slots to construct a wonder. Reveals cascade: taking a card flips whatever was hiding under it, sometimes giving the opponent a free turn of perfect information.",
    "core_loop_long": "Setup: deal three Age decks into pre-printed pyramid layouts (Age I: 4 rows, Age III: a chevron with bonus face-down rows). Each turn, the active player takes one accessible card and resolves it (build/discard/wonder); play continues until the Age's pyramid is empty, then military and progress tokens trigger, and the next Age deals. Three Ages, ~20 cards each. The game ends instantly via military supremacy (conflict pawn reaches enemy capital — at +9 toward you), scientific supremacy (collect 6 of 7 distinct science symbols), or, failing those, after Age III on victory points (civilian/commercial/wonder/guild scores summed). Wonders are limited to 7 total constructed across both players, so the seventh wonder built denies the eighth its slot — a famous 'one short' tension.",
    "decision_space": "Every pick is a four-axis decision: do I want this card's effect, do I want to deny it to my opponent (cards I can't afford are still draftable to discard), do I want to construct a wonder using it as fuel (some wonders give an extra turn or token-steal), and what does taking it reveal? The option space is small at any single decision (typically 2-6 accessible cards) but extremely deep because revealing a card hands the opponent free information. The hard part is threat assessment across three simultaneous victory paths: ignore military and you lose to a rush; ignore science and you lose to a 6-symbol sudden-death; ignore civilian VPs and you lose the points race. The pyramid layout means tempo (who moves next, who picks up double-turn tokens) is its own resource.",
    "skill_expression": "Top skills are opponent-threat reading (counting their science symbols and military shields every turn, calibrating when to switch from your engine into a denial-pick), tempo and sequencing (which card unlocks which downstream pick — and on which side of the table), and resource curve management (your buildings 'chain' so an early Stone Reserve pays for free for the rest of the game). Pattern recognition matters because progress tokens and guild scoring create combinatorial endgame value. Math is constant but small: shield counts, science pairs, coin arithmetic. Memory load is moderate — you must remember discarded cards because some wonders let you resurrect them.",
    "tactile_dependency": "low",
    "tactile_dependency_reason": "All information is open or hidden categorical card data; the pyramid's 'partially covered' state translates trivially to a digital UI (and Repos has shipped a polished first-party port via Asmodee Digital that does exactly this). No physical dexterity, no spatial-reasoning component beyond the abstract pyramid graph."
  },
  "theme": {
    "promise": "Build a Mediterranean civilization in 30 minutes against one rival — choose your path: army it down, out-science it, or out-build it on monuments and culture.",
    "setting": "Historical, ancient world, Mediterranean civilization",
    "narrative": "none — pasted-on theme. Mechanically it is a tight head-to-head card-draft engine; the Babylonian/Egyptian/Greek wonders are flavor wallpaper, evocative but non-narrative. The Pantheon and Agora expansions add slightly more thematic structure but it remains a pure strategy game.",
    "audience": "gateway, hobbyist Eurogamer",
    "art_direction": "Miguel Coimbra's signature illustration: warm earth-and-gold palette, painterly architectural studies of ancient wonders, micro-detail medallions on every card. The aesthetic is romantic-orientalist 'old empires' — sand, marble, copper — recognizable across his 7 Wonders / Small World / Cyclades body of work. High production gloss; cards read crisply at 18 inches."
  },
  "translation": {
    "digital_meta_layer_ideas": [
      "Async PvP ladder with seasonal Wonder rotation: standard duel format, Glicko-style ranked ladder, each season introduces a new pool of wonders/progress tokens (already canonical with Pantheon/Agora expansions) — but a first-party port already exists, constraining independent translations to clearly differentiated takes",
      "Roguelite civilization gauntlet: solo PvE — face 7 escalating AI rivals each themed to a victory path (rusher, scientist, builder), draft persistent civic 'edicts' between matches that warp resource costs, season ends on a final boss duel",
      "Draft-and-build merge hybrid: take the pyramid-reveal mechanic but graft it onto a merge-2 city-builder where merging buildings triggers chain bonuses — distances enough from the canonical port to ship as its own product",
      "Daily duel puzzle: shared seed — same pyramid, same opponent AI plays — leaderboard by margin of victory; uses determinism of 7WD to create a real puzzle"
    ],
    "closest_mobile_genre": "async PvP card battler / draft strategy",
    "live_service_potential": "medium"
  },
  "confidence": 0.9,
  "extraction_version": "v3"
}