Umi Deck #772168 — Yu-Gi-Oh! OCG

56 Main, 15 Extra, incl. Atlantean Dragoons x3, First Penguin x3, Mulcharmy Purulia x3. Created Oct 21, 2026 by YGO Mod.

Deck Umi OCG

Created by: YGO Mod (20517634)

Created On: 10/21/2026

 
 


Official AI Deck Power Analysis
MidrangeConfidence: LowEvaluated Oct 21, 2026
30± 8/ 100
Power Score
Structural 52
Reliability 45
Going First 50
Going Second 56

Deck — Deck Power Evaluation

Final Contextual Deck Power: 30/100 (±8 · Confidence: Low)

Overall score Value
Structural Deck Power 52
Going First 50
Going Second 56
Meta Power provisional
Tournament Reliability 45
  • Primary strategy: Midrange
  • Secondary strategies: Toolbox
  • Format: ocg

Legality

Main 56 / Extra 15 / Side 0 — status under the rules at evaluation time: Illegal. 1 violation(s).

Identity & game plan

  • Win condition: Umi / Atlantis field-driven board of Atlantean, Mermail and Daedalus monsters that converts into large Link and Xyz boards (Poseidra Abyss 3000 ATK, Azure Ocean Dragon Lord 2900 ATK, Brionac 2300 ATK) (Inferred from card text); Stealth Kragen and Toadally Awesome/Number-style Xyz boards that remove or negate the opponent's monsters (Inferred from card text)
  • Turn 1 (going first): Open a WATER starter (Mulcharmy Purulia, First Penguin, Atlantean Dragoons) and search Atlantis, City of the Sea Dragon or a Daedalus/Atlantean enabler to make Umi live, then extend into a Link or Xyz with a Level 7 WATER base (Inferred)
  • Turn 2 (going second): Hold Ash Blossom, Maxx C, Droll & Lock Bird, Infinite Impermanence or Ghost Ogre as hand traps against the opponent's first summon or search, then rebuild Umi from the Deck (Inferred)
  • Recovery: Recycle the GY through Mulcharmy draw effects, Warchief and Levia-Dragoon revival effects, and Advent of Atlantis searches (Inferred from text)
  • After siding: Not assessable: the list runs no Side Deck, so no post-side plan exists in the evidence (Provisional)

Why it scores here

This list is Illegal because Bahamut Shark is Forbidden in OCG and is run once, so the score is a structural-only read capped low and is not a normal Final Contextual score. The structural read comes from the system: the 17-copy archetype core gives about an 85 percent chance of at least one core card in a 5-card opening hand (Inferred), and Umi/Atlantis enablers, searchers and 10 hand traps give the deck a real interaction layer. Offensive closing power is credible because several Level 7 WATER Link and Xyz bosses reach 2800 to 3000 ATK. The score is held down by the low exact-archetype purity (4.2 percent, with 68 off-archetype copies across Mulcharmy, PSY-Frame, Daedalus and Atlantean sub-archetypes), by the absence of any combo evidence, by the absence of a Side Deck, and by the 56-card main with many single-copy pieces. Popularity and usage statistics were not used to add points, and card scores were not averaged into the deck score.

Dimension scores

# Dimension Score Evidence
D1 Engine Accessibility & Consistency 58 P(>=1 of the 17-copy archetype core in a 5-card opening hand) = 0.85; core present 10 of 16 typical core cards; no combo evidence to confirm the starter converts (Inferred)
D2 Opening-Hand Quality & Brick Management 50 Mixed-engine main (Umi, Atlantean, Mulcharmy, PSY-Frame, Daedalus) with 3-copy dead-on-arrival pieces such as Gordian Slicer and 10 hand traps; brick rate estimated from composition only (Inferred)
D3 Combo Conversion & Ceiling 52 No parseable boards; ceiling read from card text only, where Umi-field Link and Xyz bosses can chain several summons (Inferred)
D4 Combo Floor & Median Output 45 combos_evidence contains zero boards, so the median and floor cannot be measured; scored from text and composition only (Inferred)
D5 Interruption Resilience & Route Redundancy 55 Ash Blossom x2, Infinite Impermanence x2, Dominus Impulse x2, Droll and Ghost Ogre give several lines of interruption, and Mulcharmy plus Advent of Atlantis give some recursion (Inferred from text)
D6 Resource Economy, Recovery & Grind 58 19 searcher-text cards plus Mulcharmy draw and Warchief/Levia-Dragoon recursion keep the resource economy healthy (Inferred from text)
D7 End-Board & Interaction Quality 60 21 negation-text and 18 removal-text cards plus Stealth Kragen and Toadally removal-style effects; layering is strong but the hand traps are cheap single copies (Inferred semantic read)
D8 Going-Second & Board-Breaking Power 58 Going-second hand-trap density (Ash, Maxx C, Droll, Impermanence, Dominus Impulse) challenges an established board, but the board breakers are mostly Xyz and Link bosses (Inferred)
D9 Offensive Pressure & Closing Power 62 Poseidra Abyss 3000, Azure Ocean Dragon Lord 2900, Poseidra 2800 and Brionac 2300 are strong closers; Bahamut Shark (2600) is illegal and removed from the calculation (Verified stat plus Inferred)
D10 Main Deck Efficiency & Flexibility 48 56 main cards with 32 distinct entries, 3-copy staples across unrelated archetypes, and 4.2 percent exact-name purity lower efficiency (Verified composition)
D11 Extra Deck Efficiency & Coverage 55 15 Extra Deck cards with 14 distinct entries, 6 Link and 8 Xyz; Bahamut Shark is illegal and several flex Links are single copies (Verified composition)
D12 Side Deck Adaptability 10 Side Deck is empty (0 cards), so no adaptability can be credited and no post-side data exists (Verified)
D13 Matchup Coverage & Format Fit 45 No tournament link, no dated metagame sample and no match record; Umi archetype share is unavailable, so this is Provisional (Provisional)
D14 Pilotability & Error Tolerance 35 Multiple sub-archetype interactions (Atlantis field name changes, Daedalus conditions, Mulcharmy draw triggers) make decision-heavy turns (Inferred)

How the score was computed

Strategy profile: Midrange. Structural ≈ 52, Going First ≈ 50, Going Second ≈ 56, Meta = provisional, Tournament Reliability ≈ 45. Modifiers: Illegal list: Forbidden card in deck (-20); Off-archetype engine spread (-1). Final formula: 0.80*Structural + 0.20*TournamentReliability + ContextModifiers, Meta provisional; Illegal cap applied via modifiers (0.80*52.15 + 0.20*45 = 50.7, minus 21 = 29.7, rounded to 30)

Probability & combo summary

  • Opening a starter: P(≥1) ≈ 0.85 · 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 archetype engine is reliably accessible in opening hands (about 85 percent for at least one core card)
  • Ten hand traps and about 21 negation-text cards give going-second and mid-turn interaction
  • Strong Link and Xyz closers (Poseidra Abyss 3000 ATK, Azure Ocean Dragon Lord 2900 ATK) with multiple recursion routes

Weaknesses

  • Illegal list: Bahamut Shark is Forbidden and run once
  • Low exact-archetype purity (4.2 percent) splits the engine across Mulcharmy, PSY-Frame, Daedalus and Atlantean sub-archetypes
  • No Side Deck (0 of 15 slots), so there is no post-side plan
  • No combo evidence, so the conversion and median output are unmeasured

Critical choke points

  • Bahamut Shark (Forbidden) must leave the list before the deck can be played in OCG at all
  • Umi must be on the field for many Umi-dependent effects to work, so a single Atlantis field or Umi enabler is a choke point
Priority Remove Add Reason
1 Bahamut Shark (440556) a second copy of Poseidra Abyss, the Atlantean Dragon Lord (already in the Extra Deck pool, legal at 3) Removes the only violation and makes the list legal; Poseidra Abyss is the strongest legal Xyz closer in the list
2 Gordian Slicer x3 or one of the single-copy off-archetype tech cards Side Deck cards (up to 15) drawn from the existing hand-trap and negation pool The list has no Side Deck, so any best-of-3 plan is absent; adding side cards gives a measurable post-side plan

Evidence & confidence

  • Verified: Legality: Illegal, Bahamut Shark Forbidden with 1 copy run; main 56 and extra 15 are in size range; Exact archetype purity 4.2 percent; 68 of 71 copies are off-archetype; Usage stats: 10 of 16 core cards present; Hypergeometric probabilities: core-group P(>=1 in hand5) = 0.8493; hand-trap P(>=1 in hand6) = 0.7115
  • Inferred: Engine accessibility is high from the 17-copy core, but conversion is unmeasured; Brick risk is elevated by 3-copy dead-on-arrival pieces; Interaction density (10 hand traps, 21 negation texts) gives a credible going-second plan; Mid-game recovery is supported by Mulcharmy draw and Warchief/Levia-Dragoon revival
  • Provisional: Meta power and matchup fit: no tournament link, no dated meta sample; Going-first versus going-second split: no win-rate telemetry
  • Data gaps: no combos supplied, so combo floor, median and ceiling are text-only inference no tournament link: archetype meta-share and per-pilot record are absent no win-rate telemetry (going-first/second, matchup, brick/starter rates) Meta Power is Provisional unless a dated metagame sample is supplied no Side Deck, so no post-side evidence

More OCG Info

Umi

Umi Breakdown

 
 

 

 

 

 

 

 
 

 

 
https://metaduelist.com/
Top