Worlds 2026: open Great League robustness

Which meta matchups are decided by IVs? A 9-scenario robustness matrix, per-species cheat sheets (click a row label or a name in the meta table), and the candidate table for the Worlds 2026 open Great League meta. Format, usage corpus and sim provenance are at the foot of the page.

Robustness matrix

LickilickyQuagsire (Shadow)QuagsireForretress (Shadow)ForretressWigglytuffCorviknightEmpoleonAltariaFeraligatrStunfiskCorsola (Galarian)AzumarillMedichamTinkatonGuzzlordGourgeist (Super)TogekissAegislash (Shield)NinetalesJumpluffFearowKingdraSableye (Shadow)JellicentClodsireFurretAltaria (Shadow)GrumpigDiggersbyMantineThievul (NS+IW)Thievul (IW+PR)GreninjaAnnihilape
Lickilickymirror
Quagsire (Shadow)mirror
Quagsire
Forretress (Shadow)
Forretress
Wigglytuffmirror
Corviknightmirror
Empoleon
Altariamirror
Feraligatr
Stunfisk
Corsola (Galarian)
Azumarill
Medicham
Tinkaton
Guzzlord
Gourgeist (Super)
Togekiss
Aegislash (Shield)
Ninetales
Jumpluff
Fearow
Kingdra
Sableye (Shadow)
Jellicent
Clodsire
Furret
Altaria (Shadow)
Grumpig
Diggersby
Mantine
Thievul (NS+IW)mirror
Thievul (IW+PR)mirror
Greninja
Annihilape

Row = the focal Pokemon (its rank-1-SP spread), column = the opponent (its top-512 SP spreads). Each cell shows all 9 shield scenarios. IV-decided cells are drawn at full strength; a cell whose outcome is settled in every tested slice (either probe spread, either cohort, either bait mode) is dimmed back - 1027 of 1190 directions are IV-decided. Never aggregated to one number.

Click any cell that has a per-pair detail page to open it - a pair earns one by being IV-decided in EITHER direction, so a dimmed cell can still be a link. A cell with no detail page opens a short note instead, with links to that pair's cheat sheet row and to the equivalent battle on pvpoke.com at PvPoke's own default IVs and movesets (not our probe spread). Hover or keyboard-focus restores a dimmed cell to full strength. Diagonal cells have no colors -- the Tier-1/Tier-2 screen never baked mirror planes -- but where a full mirror deep dive exists the cell links it (marked "mirror"; also listed under Deep joint-IV analyses below).

W beats every cohort spread   ? IV-decided: the green width is the share of cohort spreads beaten (uniform over the tested cohort, not a matchmaking distribution; each side keeps a minimum 1px sliver, so near-solid means "flips only in a corner")   L beats no cohort spread (an exact 500-500 tie counts as not beaten)   W outlined = settled in the displayed slice but IV-contested in a corner of the tested space (max-attack probe, high-attack cohort, probe-spread flip, or full-grid evidence) -- the "IVs decide" flags name the corner. Every 3x3 grid uses PvPoke's battle-matrix layout: rows = own shields 0/1/2 top to bottom, columns = opponent shields left to right (scenario labels read own-opp, e.g. 2-1 = you keep two shields, they keep one). Bait modes apply to the FOCAL side only -- in the no-bait line the opponent still baits, so the two directions of a no-bait pair describe different battles and their sheets need not mirror. Hover (desktop) for exact spread counts and score margins; on phones the cheat sheets' W/L/? letters and IV-decided column carry the outcome. The hub matrix itself is a color-only overview -- its row links open the text-carrying cheat sheet. IV-decided pairs with baked Tier-2 grids link to a full per-pair detail page -- click the 3x3 grid itself (cheat sheets), the matrix cell (hub), or the details expander; any budget-deferred pairs are listed on the hub.

Accessibility note: the dimmed fills carry less contrast than the full-strength ones, and dimmed cells are exactly the settled ones, so nothing on this page depends on reading them. The hub matrix stays a color-only overview either way; the text alternative is the cheat sheet, reachable from each row label, from the dimmed-cell note, and from the meta table below.

Per-pair detail pages (555 baked)

Amber-screen check: of 73 sampled clean (not-flagged) pairs given the full 4096-spread treatment, 11 show some IV-dependence in the top-512 x top-512 block. Worst case, 31.1% of a focal's top-512 spreads have an outcome that depends on the opponent's IV roll in some scenario (largest mixed-cell share of a block: 0.82%). The amber flags are a screen, not a proof of settledness; detail pages select their scenarios from the full grids, not the screen.

Nothing is deferred: every IV-decided pair in the meta has a full-grid detail page.

CMP board: meta-wide charge-move-priority order with per-pair IV flip thresholds for the contested pairs.

IV explorer: enter your IVs and level, see the breakpoints you reach and the bulkpoints you hold vs all 35 entries.

Deep joint-IV analyses

Hand-picked high-usage IV-decided pairs get the full treatment: every spread of BOTH sides (4096 x 4096 x 9 shield scenarios per moveset/bait grid), closed-form breakpoints, denial tech for the other side, and a paste-your-collection checker. The cheat-sheet rows for these pairs link the same pages.

The meta (35 entries)

Movesets are the field-modal set when modal share >= 60%, else the PvPoke default; disagreements are shown as data. Moveset slot order shown is the order the sims ran under (PvPoke default order when the sets agree). Why each entry is in the list -- the badge system -- is at the foot of the page.

Entry (cheat sheet)recent usagerankmoveset
Lickilicky dive36.9%1Rollout / Body Slam + Shadow Ball (field-modal 97%, n=665)
Quagsire (Shadow)26.3%8Mud Shot / Aqua Tail + Stone Edge (field-modal 62%, n=473; differs from PvPoke default)
Quagsire dive7.5%9Mud Shot / Aqua Tail + Mud Bomb
Forretress (Shadow) dive17.5%16Bug Bite / Rock Tomb + Sand Tomb (field-modal 68%, n=315; differs from PvPoke default)
Forretress dive15.9%11Volt Switch / Sand Tomb + Rock Tomb (field-modal 81%, n=286)
Wigglytuff dive33.1%48Charm / Icy Wind + Swift (field-modal 94%, n=597)
Corviknight dive29.3%20Sand Attack / Air Cutter + Payback (field-modal 94%, n=528)
Empoleon dive25.4%4Metal Sound / Hydro Cannon + Drill Peck (field-modal 96%, n=457)
Altaria dive19.8%3Dragon Breath / Sky Attack + Flamethrower (field-modal 92%, n=357)
Feraligatr dive16.3%15Shadow Claw / Hydro Cannon + Ice Beam (field-modal 93%, n=294)
Stunfisk dive18.9%72Thunder Shock / Mud Bomb + Discharge (field-modal 96%, n=341)
Corsola (Galarian) dive17.2%17Astonish / Night Shade + Power Gem (field-modal 98%, n=310)
Azumarill dive14.2%24Bubble / Ice Beam + Play Rough (field-modal 87%, n=256)
Medicham dive13.6%34Psycho Cut / Ice Punch + Dynamic Punch (field-modal 95%, n=244)
Tinkaton dive13.4%2Fairy Wind / Gigaton Hammer + Bulldoze (field-modal 87%, n=241)
Guzzlord dive13.1%25Dragon Tail / Brutal Swing + Sludge Bomb (field-modal 94%, n=236)
Gourgeist (Super)12.9%87Incinerate / Shadow Ball + Seed Bomb (field-modal 100%, n=232)
Togekiss12.0%114Peck / Psyshock + Aura Sphere (field-modal 88%, n=216)
Aegislash (Shield) dive4.5%67Psycho Cut / Shadow Ball + Gyro Ball (field-modal 98%, n=81)
Ninetales dive1.9%12Ember / Weather Ball (Fire) + Energy Ball (field-modal 65%, n=34)
Jumpluff dive0.9%18Fairy Wind / Energy Ball + Acrobatics
Fearow dive1.9%21Peck / Drill Peck + Drill Run (field-modal 97%, n=35)
Kingdra dive0.7%23Dragon Breath / Surf + Swift
Sableye (Shadow)4.7%31Shadow Claw / Foul Play + Drain Punch (field-modal 82%, n=84)
Jellicent dive11.7%10Hex / Surf + Shadow Ball (field-modal 98%, n=211)
Clodsire dive9.3%19Poison Sting / Earthquake + Stone Edge (field-modal 78%, n=167)
Furret dive10.6%26Sucker Punch / Swift + Trailblaze (field-modal 72%, n=190)
Altaria (Shadow)4.6%6Dragon Breath / Sky Attack + Flamethrower (field-modal 100%, n=83)
Grumpig dive7.2%41Psywave / Dynamic Punch + Shadow Ball (field-modal 99%, n=129)
Diggersby dive8.0%44Mud Shot / Fire Punch + Scorching Sands (field-modal 80%, n=144)
Mantine0.6%55Wing Attack / Twister + Water Pulse
Thievul (NS+IW) dive0.1%43Sucker Punch / Night Slash + Icy Wind (Icy Wind injected: the pinned sim gamemaster predates the CD)
Thievul (IW+PR) dive0.1%43Sucker Punch / Icy Wind + Play Rough (Icy Wind injected: the pinned sim gamemaster predates the CD)
Greninja0.7%187Water Shuriken / Hydro Cannon + Night Slash
Annihilape5.0%151Low Kick / Close Combat + Rage Fist (field-modal 67%, n=90; differs from PvPoke default)

Candidates that stayed out

The usage top-55 that did not make the meta, plus the banned row. Collapsed-rank PLAYED* entries (see the badge section at the foot) are shown as data; we do not claim nerf vs model error.

Candidaterecent usagerankwhy not in the meta
Cradily9.0%49runner-up: biggest old->recent faller
Florges6.8%51below cut
Talonflame4.9%35runner-up: stays out
Marowak (Shadow)4.8%69below cut
Charjabug4.6%61below cut
Malamar4.3%59below cut
Moltres (Galarian)3.9%52runner-up: stays out
Clefable3.8%54below cut
Umbreon3.5%33below cut
Drapion (Shadow)3.4%62below cut
Ninetales (Alolan)3.2%64below cut
Feraligatr (Shadow)2.9%22below cut
Primeape2.9%185below cut
Lapras2.7%28runner-up: stays out, 0% of top cuts
Gastrodon2.6%63below cut
Annihilape (Shadow)2.2%164below cut
Shelgon2.2%89below cut
Ninetales (Shadow)2.0%14below cut
Talonflame (Shadow)1.9%56below cut
Steelix (Shadow)1.8%60below cut
Scizor (Shadow)1.8%310below cut
Stunfisk (Galarian)1.8%71below cut
Sableye1.7%47below cut
Blastoise1.6%53below cut
Sealeo1.5%39below cut
Malamar (Shadow)1.4%76below cut
MimikyuBANNED0.0%5Banned at Worlds: Play! Pokemon banned list, and Niantic states it is not eligible for the Competitors Cup (it only functions under the new turn system). PvPoke's open-GL #1, shown explicitly rather than silently omitted. Corpus sanity check: 0 Mimikyu in 21,719 roster entries across 36 captured events.

Why each entry is included

Badge rules (mechanical): PLAYED = top-25 recent usage AND top-30 current rank; PLAYED* = top-usage but current rank sank below 30; MODEL = current top-30 rank with no meaningful tournament footprint; FORCED = editorial include. A "(rule: X)" annotation means the literal badge and the mechanical rule disagree -- live PvPoke ranks move, and an editorial badge the rule no longer awards is shown, never hidden. The entry list itself is a human decision (recorded in scripts/worlds_meta.py), not rule-derived: the badges label the chosen list mechanically, and some current top-30-rank species stayed out (see the candidates table above).

Entrybadgenote
LickilickyPLAYED
Quagsire (Shadow)PLAYED
QuagsirePLAYED
Forretress (Shadow)PLAYED
ForretressPLAYED
WigglytuffPLAYED*
CorviknightPLAYED
EmpoleonPLAYED
AltariaPLAYED
FeraligatrPLAYED
StunfiskPLAYED*
Corsola (Galarian)PLAYED
AzumarillPLAYED
MedichamPLAYED*
TinkatonPLAYED
GuzzlordPLAYED
Gourgeist (Super)PLAYED*
TogekissPLAYED*
Aegislash (Shield)FORCEDeditorial include; reason on its cheat sheet
NinetalesMODEL
JumpluffMODEL
FearowMODEL
KingdraMODEL
Sableye (Shadow)MODEL (rule: no badge)
JellicentPLAYED
ClodsirePLAYED
FurretPLAYED
Altaria (Shadow)PLAYED (rule: MODEL)
GrumpigPLAYED*
DiggersbyPLAYED*
MantineFORCEDeditorial include; reason on its cheat sheet
Thievul (NS+IW)FORCEDeditorial include; reason on its cheat sheet
Thievul (IW+PR)FORCEDeditorial include; reason on its cheat sheet
GreninjaFORCEDeditorial include; reason on its cheat sheet
AnnihilapeFORCEDeditorial include; reason on its cheat sheet

Deliberately not built: team composition, lead/switch/closer roles, energy management lines, or "best 6" advice. This analysis is IV-robustness arithmetic on 1v1 sims, not necessarily expert-level competitive advice. What IS here: who beats whom across the opponent's plausible IV spreads, per shield scenario, both bait modes.

Format and provenance

Format: open GL 1500 + Play! banned list, legacy (old system; Worlds-confirmed). Usage: 1801 recent teams across 16 open-GL events (Dracoviz corpus, events since 2026-03; the EUIC/LAIC/NAIC Internationals are excluded as limited-meta events). Shares are team-level and per-variant (Shadow counted separately from the base form). The June 2026 rebalance went live in-game 2026-06-02 (1 p.m. PDT, with the GBL Season 27 switchover), so all recent corpus events predate it EXCEPT the final one - Turin, 2026-06-06/07, played post-rebalance. Model-side ranks are fully post-rebalance. Sim: legacy engine, planes baked 2026-08-10/2026-08-18/2026-08-26, engine 5839391a7596, gamemaster 8f1d6cca5c0f, producer c1395dfa10b9. Cohorts: opponent top-512 by stat product, plus best-SP-per-attack-IV band (labeled separately, never pooled). Focal probe spreads: rank-1 SP and max-attack within top-512. Community Day move injected for Thievul (NS+IW) (Icy Wind), Thievul (IW+PR) (Icy Wind): the pinned sim gamemaster predates that Community Day, so the move was admitted into the species pool deliberately; upstream PvPoke already lists it. Full provenance on the affected entries' cheat sheets.

Back to all dives

This project is built on PvPoke, created by Empoleon_Dynamite and released under the MIT license. The battle simulator here is a Python port of PvPoke's open-source battle logic, and all game data (the gamemaster, move stats, type chart, and meta rankings) comes from PvPoke. This project would not exist without it.