ARAM Draft Analyzer

Find the team hiding inside your rerolls.

Compare the champions your lobby actually gives you, find the swap that matters, and cross the bridge with a plan.

Get the desktop alphaSee how it worksFree alpha · Windows & macOS · Coming soon
League of Legends
League ARAM lobby with Garen, Vi, Vladimir and Karma selected, Leona highlighted, and a ten-champion available poolLive ARAM lobby
ARAM Draft Analyzer
Champion PoolSelect a champion to keep them in every recommended team.14 available

Locked: Garen, Vi, Vladimir, Karma

GarenGarenLockedViViLockedVladimirVladimirLockedKarmaKarmaLockedLeonaLeonaYouDianaDianaKha'ZixKha'ZixMordekaiserMordekaiserEvelynnEvelynnGangplankGangplankVel'KozVel'KozRenektonRenektonZedZedNautilusNautilus

Start with the lobby League gives you.

CorePriority picks and strongest combination
Most irreplaceable pick
GangplankRelated evidence 3
Strongest core combination
Gangplank + GarenEnables teammatesHow this core contributesEnables teammates63%Works internally20%Receives support17%Related evidence 17

See what holds the team together.

Core stays open while the guide reveals how the strongest combination distributes its value across the team.

AlternativesBench picks and available swaps
GangplankPreferredRenektonAlternative
Why Gangplank edged this teamLineup-specific evidence for the model result.
KarmaWith KarmaRelated evidence Related evidenceWith Karma64 Gangplank · 75 Mordekaiser · Limited sampleGangplank43.8%HighervsMordekaiser41.3%
Best complete-team scoreThe model ranks the full five-champion lineup, not each pick in isolation.
GangplankPreferredNautilusAlternative
Why Gangplank edged this teamLineup-specific evidence for the model result.
KarmaWith KarmaRelated evidence Related evidenceWith Karma64 Gangplank · 75 Mordekaiser · Limited sampleGangplank43.8%HighervsMordekaiser41.3%
Best complete-team scoreThe model ranks the full five-champion lineup, not each pick in isolation.

Compare the picks that almost won.

Hover or focus to open one comparison in place, then hold on the evidence behind the model’s choice.

Stronger scores. Better results.

Observed win rate95% interval
Observed wins by team score percentileWin rates rise from 35.0% in the bottom score decile to 63.6% in the top decile. The Y-axis spans 30% to 70%. Shading shows 95% bootstrap intervals. Based on 7,080 held-out matches from the June 2026 structured-stat benchmark.30%40%50%60%70%Win rate0–10th score percentile: 35.0% wins10–20th score percentile: 43.4% wins20–30th score percentile: 43.4% wins30–40th score percentile: 47.7% wins40–50th score percentile: 49.1% wins50–60th score percentile: 51.7% wins60–70th score percentile: 51.1% wins70–80th score percentile: 55.1% wins80–90th score percentile: 60.0% wins90–100th score percentile: 63.6% wins0255075100Team score percentile · weakest to strongest

7,080 held-out matches · June 2026 benchmark · Source data

Desktop alpha

Installers are being prepared for the first public release.

Windows

Windows 10+

Installer coming soon

macOS

macOS 12+ · Apple silicon

Package coming soon

ARAM only · Experimental recommendations · Unsigned alpha builds