The Deck

Choose the archetype the field rewards, then the sixty cards that best express it. Ladder data and controlled testing informed both decisions.

54

slots fixed by the Dragapult skeleton

6

flex slots per build in the report’s construction

2

builds: Control and Meta-predict

Kaggle metagame

The field shifted repeatedly. A general model let us train across archetypes and choose after the metagame settled. The archived ratings below come from different checkpoints, not a controlled same-model comparison.

Archetype share across the competition period. The field shifted repeatedly before settling ahead of the deadline.
Each dot is a scored upload, colored by archetype; rings mark selected submissions. Peak ratings: Alakazam 901.7, Grimmsnarl 1,023.0, Dragapult 1,040.2. Historical peaks through August 16, from different checkpoints—not a controlled same-model comparison.

Why Dragapult

Dragapult rewards setup, disruption, and damage placement. Turn-level evaluation values the board it builds; multi-action search sequences competing lines; information-gate sampling handles Drakloak’s reveal-and-keep decisions.

Sample reveals, plan turns, and evaluate positions. Green boxes show estimated win probability.

Decklist construction

Our library combines ladder and over-the-board lists. Inclusion rates identify the skeleton and help predict an opponent’s list for search. Cards appearing in at least 95% of lists define the report’s 54-slot starting point.

Weighted card inclusion across 17 Dragapult ladder lists. Blue marks cards included in at least 95% of lists; gold marks less common choices.

Two builds from one skeleton

Control emphasized consistency with Dudunsparce. Meta-predict included Chi-Yu to answer Crustle, a difficult matchup we expected to resurge.

The skeleton fixed 54 of 60 slots. Control filled the remaining six for consistency, Meta-predict for the field we expected.

Testing individual choices

Every variant faces the same seeded deals with seats swapped and the same pilot throughout. This isolates the effect of a card or card combination. The sensitivity figure remains a placeholder until the paired-match receipts arrive.

Matched meta, pilot and seeds. Basic: highest-played cards. Jamming swap: Meta-predict with the more common Jamming Tower replacing Watchtower, our hypothesized high-value choice.

The submitted lists

Both archived 60-card lists are preserved in full.

Chi-Yu / Watchtowerthe meta-predict build

Pokémon17

4DrakloakDrakloakTWM-129
4DreepyDreepyTWM-128
3Dragapult exDragapult exTWM-130
2MunkidoriMunkidoriTWM-95
1BudewBudewPRE-4
1Chi-YuChi-YuTWM-39
1Fezandipiti exFezandipiti exSFA-38
1Meowth exMeowth exPOR-62

Trainer34

4Buddy-Buddy PoffinBuddy-Buddy PoffinTEF-144
4Crushing HammerCrushing HammerSVI-168
4Lillie's DeterminationLillie's DeterminationMEG-119
4Poké PadPoké PadPOR-81
4Ultra BallUltra BallSVI-196
3Boss’s OrdersBoss’s OrdersPAL-172
3CrispinCrispinSCR-133
3Night StretcherNight StretcherSFA-61
2Team Rocket's WatchtowerTeam Rocket's WatchtowerDRI-180
1Jamming TowerJamming TowerTWM-153
1JudgeJudgeSVI-176
1Unfair StampUnfair StampTWM-165

Energy9

4Basic {R} EnergyBasic {R} EnergySVE-2
3Basic {P} EnergyBasic {P} EnergySVE-5
2Basic {D} EnergyBasic {D} EnergySVE-7
Chi-Yu / Watchtower — 60 cards, 23 distinct — the meta-predict build, submitted 16 August 2026.
Dudunsparce / Risky Ruinsthe control build

Pokémon19

4DrakloakDrakloakTWM-129
4DreepyDreepyTWM-128
3Dragapult exDragapult exTWM-130
2BudewBudewPRE-4
2MunkidoriMunkidoriTWM-95
1DudunsparceDudunsparceTEF-129
1DunsparceDunsparceJTG-120
1Fezandipiti exFezandipiti exSFA-38
1Meowth exMeowth exPOR-62

Trainer32

4Buddy-Buddy PoffinBuddy-Buddy PoffinTEF-144
4Crushing HammerCrushing HammerSVI-168
4Lillie's DeterminationLillie's DeterminationMEG-119
4Poké PadPoké PadPOR-81
4Ultra BallUltra BallSVI-196
3Boss’s OrdersBoss’s OrdersPAL-172
3CrispinCrispinSCR-133
2Night StretcherNight StretcherSFA-61
2Risky RuinsRisky RuinsMEG-127
1JudgeJudgeSVI-176
1Unfair StampUnfair StampTWM-165

Energy9

4Basic {P} EnergyBasic {P} EnergySVE-5
3Basic {R} EnergyBasic {R} EnergySVE-2
2Basic {D} EnergyBasic {D} EnergySVE-7
Dudunsparce / Risky Ruins — 60 cards, 23 distinct — the control build, submitted 16 August 2026.

Explore the decklist library

This interactive builder ranks candidates from the archived source lists. It is a data-exploration tool, not a winrate predictor.

Every card is ranked by how much of the field plays it, and the ranking recomputes each time you add one. A slot is a card and a copy number, so the fourth Dreepy is ranked separately from the first, and picking a card that belongs to a package pulls the rest of the package up with it. The two sources are kept apart because they disagree: energy denial is near universal on the ladder and a minority choice in human tournament play.

The list

Most played next

Historical feature coverage and stadium analysis

These descriptive summaries use the archived lists and field sample. Inclusion and coverage are not causal estimates of a card’s contribution to wins.

0 20 40 60 ours · meta-predict 67 Dragapult Blaziken 67 Slowking 65 ours · control 64 Dragapult 64 7% of field Crustle 58 4% of field Hydrapple Meganium 57 4% of field Raging Bolt Ogerpon 55 Lopunny Dudunsparce 54 12% of field Cynthia's Garchomp 51 Festival Lead 50 Grimmsnarl Froslass 49 24% of field Teal Mask Ogerpon 49 3% of field Lucario Hariyama 48 4% of field Alakazam Dudunsparce 47 15% of field mechanical features touched, of 194
Mechanical features touched by each archetype’s list, of the model’s 194; ours in blue. Each archetype is shown at its widest recorded list.
Team Rocket’s Watchtower Colorless Pokémon in play lose their Abilities 60% share of the field Dudunsparce · also ours 29% Meowth ex · also ours 28% Fan Rotom 13% Mega Kangaskhan ex 8% Jamming Tower Pokémon Tools have no effect 38% share of the field Hero’s Cape 16% Air Balloon 15% Brave Bangle 4% Handheld Fan 3% Risky Ruins 2 counters on each Basic non-Darkness Pokémon benched such Pokémon per list, the field’s largest families 6 Grimmsnarl 9 Dragapult 9 Lopunny 8 Alakazam 11 Hydrapple 10 Lucario
How much of the field each of our Stadiums touches, over every ladder list from 6 to 12 August.

Results

The submission won 57.4% of 1,999 games and peaked at a rating of 1124, inside the top twenty. Meta-predict did better against Crustle; Control was solid across the field. See Performance for matchup results and the post-deadline comparison.