Umi Deck #772623 — Yu-Gi-Oh! OCG

Deck — Deck Power Evaluation

Final Contextual Deck Power: 30/100 (±10 · Confidence: Low)

• Primary strategy: Midrange
• Secondary strategies: Control
• Format: ocg

Legality

Main 57 / Extra 15 / Side 0 — status under the rules at evaluation time: Illegal. 2 violation(s).

Identity & game plan

• Win condition: Umi / Atlantis engine: Atlantis field plus Warchief, Levia-Dragon and Daedalus Link/Synchro extra-deck pressure (Inferred from card text; no combo data); Kairyu-Shin and Rage of Kairyu-Shin removal backed by Mulcharmy draw and Chengying/White Aura Whale boards (Inferred)
• Turn 1 (going first): Open Advent of Atlantis or Call of Atlantis with Fish Sonar / Siren Vortex to reach Atlantis, then Link into Daedalus monsters (Inferred; no recorded boards)
• Turn 2 (going second): Hand traps (Ash Blossom, Droll & Lock Bird, Maxx C, Solemn Judgment, Dominus Impulse, Solemn Warning) to interrupt the opponent's summons, then rebuild through Warchief and Mulcharmy draws (Inferred)
• Recovery: Warchief and Ib / Siren Vortex recursion from GY plus Pot of Sloth card advantage (Inferred from text)
• After siding: Not evaluable: Side Deck size is 0, so no post-side plan exists in the evidence

Why it scores here

The deck is Illegal in OCG: it runs one copy each of two Forbidden cards (Bahamut Shark and Triple Tactics Talent), so it receives a structural-only score capped low and not a normal Final Contextual score. Within that structural read, the Atlantis/Umi engine is reasonably accessible: 21 archetype-core copies give P(at least one core card in a 5-card opening hand) ≈ 0.91, and 17 searcher-text cards give ≈ 0.84 (Inferred hypergeometric estimates). Interaction is dense, with 7 hand traps and 19 negation-text cards, which supports the end-board and going-second reads. Consistency is held back by low exact archetype purity (5.6%, spread across Atlantis, Daedalus, Mulcharmy, Solemn and generic cards) and by the absence of any recorded boards, so floor and median output are unknown. The score reflects the system's structure (engine access, interaction, main deck size) and not the popularity of the Umi archetype, and no card scores were averaged.

Dimension scores

How the score was computed

Strategy profile: Midrange primary with Control secondary (not separately computed). Structural ≈ 53, Going First ≈ 55, Going Second ≈ 50, Meta = provisional, Tournament Reliability ≈ 45. Modifiers: banlist exposure (-12). Final formula: Structural 53 → 0.80·53 + 0.20·45 = 51.4; minus banlist-exposure modifier (−12) = 39.4; Illegal cap applied → 30 (theoretical, structural-only). Meta provisional.

Probability & combo summary

• Opening a starter: P(≥1) ≈ 0.91 · key two-card combo ≈ — · full-brick estimate ≈ — (all Inferred)
• Combo output: best PowerIndex —, median —, board-per-starter —
• Turn order: 0 going-first lines vs 0 going-second lines

Strengths

• Core engine access is high (archetype-core P ≈ 0.91 in opening 5)
• Dense interaction: 7 hand traps and 19 negation-text cards
• Extra Deck covers Link, Synchro and Xyz with 15 slots used
• Main deck size of 57 is inside the OCG 40–60 range

Weaknesses

• Illegal: two Forbidden cards are run (Bahamut Shark, Triple Tactics Talent)
• Low exact archetype purity (5.6%); engine spread across several sub-archetypes
• No side deck, so no post-side adaptability
• No combo, tournament or win-rate evidence to validate floor or median output

Critical choke points

• Removing both Forbidden cards is required before any normal legality rating
• Atlantis field and Warchief/Levia-Dragon routes are the main engine; loss of either slows the plan

Recommended changes

Evidence & confidence

• Verified: Illegal: Bahamut Shark and Triple Tactics Talent are Forbidden with 0 legal copies; Main 57 and Extra 15 are inside the OCG size rules; side is 0; Copy limits for Limited and Semi-Limited cards are respected
• Observed: Archetype usage: Atlantis-family and Umi core cards are present in the list; Usage stats show 10 of 16 core cards present (62.5%)
• Inferred: Engine access P ≈ 0.91 from the hypergeometric estimate; Interaction density from hand-trap and negation text counts; Going-first and going-second plans from card text
• Provisional: Meta Power (null) and matchup fit, pending a dated metagame sample
• Data gaps: No combos supplied; combo floor, median and ceiling are text-only inference No tournament link; archetype meta-share and pilot record are absent No win-rate telemetry for going-first, going-second, matchup, brick or OTK rates Side deck is empty, so post-side plans cannot be assessed Meta Power is provisional without a dated metagame sample